Hangehis in the Property Management Tech Industry

Hangeh

hangeh

Reading this, you may think that I want to discuss cities. Well, it’s far from that— I am sorry to disappoint. Instead, the would be my focus. You may ask why, and I would be back with an answer in no time. Let's get right into it. Ethereum blockchain If you are an avid follower of the happenings in the crypto ecosystem for a while, you would agree that the Proof-of-Stake (PoS) project has been at the center of attention for quite a number of reasons. From the consensus switch to creating a 2023 roadmap, Ethereum seems to have proven why detest its courage. But where is this all heading? Bitcoin maximalists Before Shanghai, Drop-off at Sepolia Ethereum’s most significant event in 2022 was the where it exited the Proof-of-Work (PoW) mechanism to PoS. While the transition came with a lot of controversies, the second most valuable blockchain seems to have settled into its new home. Merge Meanwhile, the Ethereum development team is faced with new challenges in the new year. Prior to the time of writing, its Mainnet upgrade for validators to withdraw their already-staked Ether [stETH] was billed for the first quarter. For context, the validators are a group who have staked at least 32 ETH. These set of people also have the responsibility of verifying transactions on the blockchain and guaranteeing Ethereum’s security. However, there have been some delays in the event tagged the “Shanghai Upgrade”. But hold-ups did not happen without progress as Ethereum also recorded success with a Testnet stage. Recently, the was declared successful. Sepolia Testnet This stage helps Ethereum application developers to test their smart contracts without spending real EHT. Hence, the declaration implied that validators were almost set for the Mainnet. But this would not happen without the final Testnet process. And what is this stage? Georli Is Next Before the final meetup with the Shanghai upgrade, the developers would test the process at Georli. This stage confirms authority by testing the blockchain’s applications before the final Mainnet. So, once this is established as victorious on March 14 there would be no problem with the Shanghai upgrade. It is noteworthy to mention that the Shanghai upgrade is a hard fork of Ethereum. It is also the first significant upgrade after the 2022 Merge, aimed at implementing the EIP-4895. Finally… Beyond the fast-approaching upgrade, Ethereum’s roadmap has in it the Surge, Verge, Purge, and Splurge. Some of these could happen before the year runs to a close. However, like the Merge, there is no confirmation that the event would have a positive effect on the ETH price. While protocols under the blockchain might, nothing is guaranteed. So, consider this article as informative with no connection to financial advice. Cheers

What Next After Ethereum's Shanghai Upgrade? Hangzhou?

Alibaba linked up a city’s worth of security cameras with AI that sees problems before they happen What do security cameras and “a group of blind people trying to describe an elephant” have in common? According to Alibaba’s Distinguished Engineer Dr. Hua Xiansheng, this ancient Indian parable in which people grope for a sense of the whole animal is an apt metaphor for what you get when smart machines can only process what is right in front of them. Far from a lament of technology’s shortcomings, this “forest for the trees” problem is how Dr. Hua recently framed a new Alibaba AI venture, known as “City Brain”, for audiences at The Computing Conference in Nanjing. Built to make smarter use of aging camera infrastructures, City Brain coordinates data across every camera in an urban security system to enhance the smart functionality of each node and deliver actionable insights, from traffic forecasting to accident response. And with one of City Brain’s three component products now deployed throughout Hangzhou city, the proverbial blind elephant-survey is already giving way to vision-enhanced digital traffic administration for a population of almost 10 million. Crisis Response and Prevention City Brain’s first component, TianYao, couples 24-hour pedestrian and vehicle traffic monitoring with algorithms that recognize anomalies and notify authorities of incidents like traffic congestion and vehicle collisions. It also processes image feeds over time to identify sites with high incident frequency, helping human administrators locate problems and focus solutions for maximum impact. Adding to its sophistication, TianYao then offers immediate feedback on the effectiveness of those solutions. In practice, Tianyao has already proven an effective replacement for traffic patrol officers in Hangzhou even under low-light conditions, freeing up human labor to address more urgent tasks and emergencies. Image Recognition With City Brain’s second TianYing component, Alibaba introduces a technology more effective at pedestrian recognition than human agents provided with the same visual feed. TianYing draws on Alibaba’s two decades of image recognition research, specifically focusing on the concept of an index object (such as a person of interest) and visual cues that can describe and confirm that object in a camera scan. With 96% recognition success over the human average of 94%, it offers a powerful tool for quickly and accurately identifying missing persons or tracking hit-and-run fugitives, for example, freeing human agents to focus on pursuit and apprehension. Paving the Way for Smarter Traffic Management City Brain’s last component TianJi, still in development, is a predictive analysis tool for vehicle and pedestrian traffic flows. With early testing in Suzhou city confirming it can accurately predict traffic patterns in time spans ranging from minutes to an hour, TianJi offers the hope of better traffic management and could soon enable early warnings to help individual commuters plan their travels. As one of the first Four Open Innovation Platforms for AI chosen by China’s Ministry of Science and Technology, the majority of City Brain is currently open for public contributions to development. With early implementation standing up to the demands of real urban settings, Alibaba anticipates City Brain and other AI systems will prove an indispensable urban resource comparable to utilities and physical infrastructure in the near future. Know more about Alibaba’s CIty Brain: https://hackernoon.com/city-brain-cloud-computing-and-data-technology-help-our-city-think-deeply-382e72f7f7f Alibaba Tech First hand, detailed, and in-depth information about Alibaba’s latest technology → Facebook: . Twitter: . “Alibaba Tech” “AlibabaTech”

City Brain: Cameras Get a New Perspective in Hangzhou

AutoStory: Generating Diverse Storytelling Images with Minimal Human Effort: Experiments

Authors:
(1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn);
(2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn);
(3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn);
(4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn);
(5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com);
(6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Table of Links Abstract and Introduction
Related Work
Our Method
Experiments
Conclusion and References
Appendix 3 OUR METHOD The goal of our method is to generate diverse storytelling images of high quality and with minimal human effort. Considering the complexity of scenes in storytelling images, our general idea is to combine the comprehension and planning capabilities of LLMs [et al 2023; OpenAI 2023] and the generation ability of the large-scale text-to-image models [Rombach et al. 2022]. The pipeline is shown in Fig. 2, which can be divided into a condition preparation stage in (a) and a conditional image generation stage in (b). Specifically, we first utilize LLMs to convert the textual descriptions of stories into layouts of the storytelling images, as detailed in Sec. 3.1. To improve the quality of generated story images, we propose a simple yet effective method to transform sparse bounding boxes into dense control signals like sketches or keypoints, without introducing manual labor (detailed in Sec. 3.2). Subsequently, we generate story images with a reasonable scene arrangement based on the layout, as detailed in Sec. 3.3. Finally, we propose a method to eliminate the requirement for users to collect training data for each character, enabling the generation of identity-consistent story images from only texts (detailed in Sec. 3.4). Since our approach only fine-tunes the pre-trained text-to-image image diffusion model on a few images, we can easily leverage existing models on civiati[1] for storytelling in arbitrary characters, scenes, and even styles. 3.1 Story to Layout Generation Layout Generation. After dividing the story into panel descriptions, we leverage LLMs to extract the scene layout from each panel description, as shown in the following equation: 3.2 Dense Condition Generation Motivation. Although using sparse bounding boxes as a control signal can improve the generation of subjects and obtain more reasonable scene layouts, we find that it cannot consistently produce high-quality generation results. There are cases where the images do not exactly match the scene layout or the generated images are of low quality, as detailed in the experiments in Sec. 4.4. We believe that this is mainly due to the limited information provided by the bounding boxes. The model faces difficulties in generating a large amount of content all at once, with limited guidance. For this reason, we propose to improve the final story image generation by introducing dense sketch or keypoint guidances. To this end, we devise a dense condition generation module based on the layout generated in the previous section, as shown in Fig. 2(d). Note that the process of composing dense conditions is fully automatic and does not require any manual interaction. 3.3 Controllable Storytelling Image Generation Large-scale pre-trained text-to-image models, such as Stable Diffusion [Rombach et al. 2022], are capable of generating images from text. However, limited by the language comprehension ability of the text encoder in the model, and the incorrect association between text and image regions in the generation process, the directly generated images often suffer from a series of problems such as object missing, attribution confusion, etc [Chefer et al. 2023]. To tackle this, we introduce additional control signals to improve the quality of image generation. 3.4 Eliminating Character-wise Data Collection Identity Consistency. We propose a training-free consistency modeling method to meet the requirement of identity consistency, as shown in Fig. 3 (d). Specifically, we treat multiple images of a single character as different frames in a video and generate them simultaneously using a pre-trained diffusion model. In this process, the self-attention in the generative model is expanded to other “video frames" [Wang et al. 2023; Wu et al. 2022] to strengthen the dependencies among images, thus obtaining identity-consistent generation results. Concretely, in self-attention, we let the latent features in each frame attend to the features in the first and previous frames to build the dependency. The process can be represented as: Subsequently, we extract sketches for non-human characters and keypoints for human characters from these images. Finally, we use T2I-Adapter to inject the control guidance into the latent feature of corresponding frames in the generation process. In addition, in order to further ensure the quality of the generated data, we use CLIP score to filter the generated data, and select the images that are consistent with the text descriptions as the training data for customized generation. Discussion. In this section, we combine the proposed training-free identity-consistency modeling method with the viewpoint conditioned image translation model to achieve both identity consistency and diversity in character generation. A simpler approach is to directly use the multi-view images from the view-point conditioned image translation model as training data for customization. However, we found that the directly generated results often suffer from distortions or large differences in the color and texture of the images from different viewpoints (see Sec. 4.4 for details). For this reason, we need to leverage the above consistency modeling approach to obtain both texture- and structure-consistent images for each character. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. [1] https://civitai.com Authors: (1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn); (2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn); (3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn); (4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn); (5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com); (6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Authors: Authors: (1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn); (2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn); (3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn); (4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn); (5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com); (6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Table of Links Abstract and Introduction Related Work Our Method Experiments Conclusion and References Appendix Abstract and Introduction Abstract and Introduction Related Work Related Work Our Method Our Method Experiments Experiments Conclusion and References Conclusion and References Appendix Appendix 3 OUR METHOD The goal of our method is to generate diverse storytelling images of high quality and with minimal human effort. Considering the complexity of scenes in storytelling images, our general idea is to combine the comprehension and planning capabilities of LLMs [et al 2023; OpenAI 2023] and the generation ability of the large-scale text-to-image models [Rombach et al. 2022]. The pipeline is shown in Fig. 2, which can be divided into a condition preparation stage in (a) and a conditional image generation stage in (b). Specifically, we first utilize LLMs to convert the textual descriptions of stories into layouts of the storytelling images, as detailed in Sec. 3.1. To improve the quality of generated story images, we propose a simple yet effective method to transform sparse bounding boxes into dense control signals like sketches or keypoints, without introducing manual labor (detailed in Sec. 3.2). Subsequently, we generate story images with a reasonable scene arrangement based on the layout, as detailed in Sec. 3.3. Finally, we propose a method to eliminate the requirement for users to collect training data for each character, enabling the generation of identity-consistent story images from only texts (detailed in Sec. 3.4). Since our approach only fine-tunes the pre-trained text-to-image image diffusion model on a few images, we can easily leverage existing models on civiati[1] for storytelling in arbitrary characters, scenes, and even styles. 3.1 Story to Layout Generation Layout Generation. After dividing the story into panel descriptions, we leverage LLMs to extract the scene layout from each panel description, as shown in the following equation: Layout Generation. 3.2 Dense Condition Generation Motivation . Although using sparse bounding boxes as a control signal can improve the generation of subjects and obtain more reasonable scene layouts, we find that it cannot consistently produce high-quality generation results. There are cases where the images do not exactly match the scene layout or the generated images are of low quality, as detailed in the experiments in Sec. 4.4. Motivation We believe that this is mainly due to the limited information provided by the bounding boxes. The model faces difficulties in generating a large amount of content all at once, with limited guidance. For this reason, we propose to improve the final story image generation by introducing dense sketch or keypoint guidances. To this end, we devise a dense condition generation module based on the layout generated in the previous section, as shown in Fig. 2(d). Note that the process of composing dense conditions is fully automatic and does not require any manual interaction. 3.3 Controllable Storytelling Image Generation Large-scale pre-trained text-to-image models, such as Stable Diffusion [Rombach et al. 2022], are capable of generating images from text. However, limited by the language comprehension ability of the text encoder in the model, and the incorrect association between text and image regions in the generation process, the directly generated images often suffer from a series of problems such as object missing, attribution confusion, etc [Chefer et al. 2023]. To tackle this, we introduce additional control signals to improve the quality of image generation. 3.4 Eliminating Character-wise Data Collection Identity Consistency. We propose a training-free consistency modeling method to meet the requirement of identity consistency, as shown in Fig. 3 (d). Specifically, we treat multiple images of a single character as different frames in a video and generate them simultaneously using a pre-trained diffusion model. In this process, the self-attention in the generative model is expanded to other “video frames" [Wang et al. 2023; Wu et al. 2022] to strengthen the dependencies among images, thus obtaining identity-consistent generation results. Concretely, in self-attention, we let the latent features in each frame attend to the features in the first and previous frames to build the dependency. The process can be represented as: Identity Consistency. Subsequently, we extract sketches for non-human characters and keypoints for human characters from these images. Finally, we use T2I-Adapter to inject the control guidance into the latent feature of corresponding frames in the generation process. In addition, in order to further ensure the quality of the generated data, we use CLIP score to filter the generated data, and select the images that are consistent with the text descriptions as the training data for customized generation. Discussion . In this section, we combine the proposed training-free identity-consistency modeling method with the viewpoint conditioned image translation model to achieve both identity consistency and diversity in character generation. A simpler approach is to directly use the multi-view images from the view-point conditioned image translation model as training data for customization. However, we found that the directly generated results often suffer from distortions or large differences in the color and texture of the images from different viewpoints (see Sec. 4.4 for details). For this reason, we need to leverage the above consistency modeling approach to obtain both texture- and structure-consistent images for each character. Discussion This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv [1] https://civitai.com

AutoStory: Generating Diverse Storytelling Images with Minimal Human Effort: Our Method

Authors:
(1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn);
(2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn);
(3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn);
(4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn);
(5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com);
(6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Table of Links Abstract and Introduction
Related Work
Our Method
Experiments
Conclusion and References
Appendix 2 RELATED WORK 2.1 Story Visualization Story visualization aims to generate a series of visually consistent images from a story described in text. Limited by the generative capacity of the model, many story visualization approaches [Chen et al. 2022; Li 2022; Li et al. 2019; Maharana and Bansal 2021; Maharana et al. 2021, 2022; Pan et al. 2022; Rahman et al. 2022; Song et al. 2020] seek to largely simplify the task such that it becomes tractable, by considering specific characters, scenes, and image styles in a particular dataset. Early story visualization methods are mostly built upon GANs [Goodfellow et al. 2020]. For example, StoryGAN [Li et al. 2019] pioneers the story visualization task by proposing a GAN-based framework that considers both the full story and the current sentence for coherent image generation. CP-CSV [Song et al. 2020], DuCo-StoryGAN [Maharana et al. 2021], and VLC-StoryGAN [Li et al. 2019] follow the GAN-based framework, while improving the consistency of storytelling via better character-preserving or text understanding. Difference from these works, VP-CSV [Chen et al. 2022] leverages VQ-VAE a transformer-based language model for story visualization. StoryDALL-E [Maharana et al. 2022] leverages the pre-trained DALL-E [Ramesh et al. 2021] for better story visualization and proposes a novel task named story continuation that supports story visualization with a given initial image. AR-LDM [Pan et al. 2022] proposes a diffusion model-based method that generates story images in an autoregressive manner. While progress has been made, these methods rely on storyspecific training on datasets like PororoSV [Li et al. 2019] and FlintstonesSV [Maharana and Bansal 2021], making it difficult to generalize these methods to varying characters and scenes. The development of large-scale pre-trained text-to-image generative models [Ramesh et al. 2022, 2021; Rombach et al. 2022; Saharia et al. 2022] opens up new opportunities for generalizable story visualization. Several attempts have been made to generate storytelling images with diverse characters [Gong et al. 2023; Jeong et al. 2023; Liu et al. 2023c]. Jeong et al. [Jeong et al. 2023] utilized textual inversion [Gal et al. 2022] to swap human identities in story images, thus generalizing the characters in story visualization. However, the identity is not well preserved, and the method is limited to a single human character in storytelling. Intelligent Grimm [Liu et al. 2023c] proposes the task of open-ended visual storytelling. They collect a dataset of children’s storybooks and train an autoregressive generative model for story visualization. The limitation is clear: they focus on the storytelling of the children’s storybook style, and it needs to re-train the model to generalize to other styles, contents, etc., which is not scalable. Probably the most similar work to ours is TaleCraft [Gong et al. 2023], which also proposes a systematic pipeline for story visualization. Note that, they require user-provided sketches for each character in each story image to obtain visually pleasing generations, which can be laborious to obtain. Moreover, all existing methods rely on multiple user-provided images for each character to obtain identity-coherent story visualizations. In contrast, our method allows for generating diverse and coherent story visualization results with only text descriptions as inputs. 2.2 Controllable Image Generation The scaling of text-image paired data [Schuhmann et al. 2022], computational resources, and model size have enabled unprecedented text-to-image (T2I) generation results [Ramesh et al. 2022, 2021; Rombach et al. 2022; Saharia et al. 2022]. Large-scale pre-trained text-to-image models, such as Stable Diffusion [Rombach et al. 2022], are capable of generating images from text, i.e., 𝐼 = DM (𝑝), where DM(·) is the pre-trained diffusion model and 𝑝 is the text prompt that describe the image 𝐼. In this process, the text information is passed into the image’s latent representation through cross-attention layers in the model. The attention [Vaswani et al. 2017] operation can be written as However, limited by the language understanding capability of the text encoder and poor text-to-image content association [Chefer et al. 2023], T2I models, like Stable Diffusion [Rombach et al. 2022], can perform poorly in the generation of multiple characters and complex scenes [Chefer et al. 2023]. To alleviate this drawback, some approaches introduce explicit spatial guidance in T2I generative models. For example, ControlNet [Zhang and Agrawala 2023] uses zero convolution layers and a trainable copy of the original model weights, introducing reliable control in diffusion models. T2I-Adapter [Mou et al. 2023] achieves control ability by proposing the adapter that extracts guidance feature and adds it to the feature from the corresponding UNet encoder. GLIGEN [Li et al. 2023] injects a gated self-attention block into the UNet, enabling the model to make good use of the grounding inputs. Inspired by the ability of large language models (LLMs) [et al 2023; OpenAI 2023] being able to understand and plan, recent works [Feng et al. 2023; Lian et al. 2023] employ LLM for layout generation. Specifically, LayoutGPT [Feng et al. 2023] achieves plausible results in 2D image layouts and even 3D indoor scene synthesis by applying in-context learning on LLMs. LLM-grounded Diffusion [Lian et al. 2023] proposes a two-stage process based on the LLM-generated layout and local prompts. Specifically, it first generates the local objects within each bounding box based on the corresponding local prompt, and then re-generates the final result based on the inversed latent of local objects. While effective, LLMgrounded Diffusion requires careful hyper-parameter tuning for the trade-off between structural guidance and inter-object relationship modeling. Moreover, it is difficult for the users to control the detailed structure of the generated objects. In contrast, we use the intuitive sketch or keypoint to guide the final image generation. Thus, we can not only achieve high-quality story image generation, but also allow interactive story visualization by simply tuning the generated sketch or keypoint conditions. 2.3 Customized Image Generation Story visualization requires that the identities of characters and scenes in a story remain consistent across different images. Customized image generation can meet this requirement to a large extent. Early methods [Gal et al. 2022; Ruiz et al. 2022] focus on the customized generation of a single object. For example, DreamBooth [Ruiz et al. 2022] fine tunes the pre-trained T2I diffusion model under a class-specific prior-preservation loss. Textual Inversion [Gal et al. 2022] enables customized generation by inverting subject image content into text embeddings. Unlike these approaches, Custom Diffusion [Kumari et al. 2022] further achieves multi-subject customization by combining the multiple customization weights through closed-form constrained optimization. Cones [Liu et al. 2023a] finds that a small cluster of concept neurons in the diffusion model corresponds to a single subject, and thus achieves customized generation of multiple objects by combining these concept neurons. Cones2 [Liu et al. 2023f] further achieves more effective multi-object customization by combining text embedding of different concepts with simple layout control. Differently, Mix-of-Show [Gu et al. 2023] proposes gradient fusion to effectively combine multiple customized LoRA [Hu et al. 2022] weights and performs multi-object customization with the aid of the T2I-Adapter’s dense controls. While significant progress has been made, existing methods perform poorly on one-shot customization. The training data for subject-driven generation has to be identity-consistent and diverse. As a result, existing story visualization methods require multiple user-provided images for each character. To tackle this issue, we propose a training-free consistency modeling method, and leverage the 3D prior in 3D-aware generative models [Liu et al. 2023d,b] to obtain multi-view consistent character images for customized generation, thus eliminating the reliance on human labor to collect or draw character images. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. Authors: (1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn); (2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn); (3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn); (4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn); (5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com); (6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Authors: Authors: (1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn); (2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn); (3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn); (4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn); (5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com); (6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Table of Links Abstract and Introduction Related Work Our Method Experiments Conclusion and References Appendix Abstract and Introduction Abstract and Introduction Related Work Related Work Our Method Our Method Experiments Experiments Conclusion and References Conclusion and References Appendix Appendix 2 RELATED WORK 2.1 Story Visualization Story visualization aims to generate a series of visually consistent images from a story described in text. Limited by the generative capacity of the model, many story visualization approaches [Chen et al. 2022; Li 2022; Li et al. 2019; Maharana and Bansal 2021; Maharana et al. 2021, 2022; Pan et al. 2022; Rahman et al. 2022; Song et al. 2020] seek to largely simplify the task such that it becomes tractable, by considering specific characters, scenes, and image styles in a particular dataset. Early story visualization methods are mostly built upon GANs [Goodfellow et al. 2020]. For example, StoryGAN [Li et al. 2019] pioneers the story visualization task by proposing a GAN-based framework that considers both the full story and the current sentence for coherent image generation. CP-CSV [Song et al. 2020], DuCo-StoryGAN [Maharana et al. 2021], and VLC-StoryGAN [Li et al. 2019] follow the GAN-based framework, while improving the consistency of storytelling via better character-preserving or text understanding. Difference from these works, VP-CSV [Chen et al. 2022] leverages VQ-VAE a transformer-based language model for story visualization. StoryDALL-E [Maharana et al. 2022] leverages the pre-trained DALL-E [Ramesh et al. 2021] for better story visualization and proposes a novel task named story continuation that supports story visualization with a given initial image. AR-LDM [Pan et al. 2022] proposes a diffusion model-based method that generates story images in an autoregressive manner. While progress has been made, these methods rely on storyspecific training on datasets like PororoSV [Li et al. 2019] and FlintstonesSV [Maharana and Bansal 2021], making it difficult to generalize these methods to varying characters and scenes. The development of large-scale pre-trained text-to-image generative models [Ramesh et al. 2022, 2021; Rombach et al. 2022; Saharia et al. 2022] opens up new opportunities for generalizable story visualization. Several attempts have been made to generate storytelling images with diverse characters [Gong et al. 2023; Jeong et al. 2023; Liu et al. 2023c]. Jeong et al. [Jeong et al. 2023] utilized textual inversion [Gal et al. 2022] to swap human identities in story images, thus generalizing the characters in story visualization. However, the identity is not well preserved, and the method is limited to a single human character in storytelling. Intelligent Grimm [Liu et al. 2023c] proposes the task of open-ended visual storytelling. They collect a dataset of children’s storybooks and train an autoregressive generative model for story visualization. The limitation is clear: they focus on the storytelling of the children’s storybook style, and it needs to re-train the model to generalize to other styles, contents, etc., which is not scalable. Probably the most similar work to ours is TaleCraft [Gong et al. 2023], which also proposes a systematic pipeline for story visualization. Note that, they require user-provided sketches for each character in each story image to obtain visually pleasing generations, which can be laborious to obtain. Moreover, all existing methods rely on multiple user-provided images for each character to obtain identity-coherent story visualizations. In contrast, our method allows for generating diverse and coherent story visualization results with only text descriptions as inputs. 2.2 Controllable Image Generation The scaling of text-image paired data [Schuhmann et al. 2022], computational resources, and model size have enabled unprecedented text-to-image (T2I) generation results [Ramesh et al. 2022, 2021; Rombach et al. 2022; Saharia et al. 2022]. Large-scale pre-trained text-to-image models, such as Stable Diffusion [Rombach et al. 2022], are capable of generating images from text, i.e., 𝐼 = DM (𝑝), where DM(·) is the pre-trained diffusion model and 𝑝 is the text prompt that describe the image 𝐼. In this process, the text information is passed into the image’s latent representation through cross-attention layers in the model. The attention [Vaswani et al. 2017] operation can be written as However, limited by the language understanding capability of the text encoder and poor text-to-image content association [Chefer et al. 2023], T2I models, like Stable Diffusion [Rombach et al. 2022], can perform poorly in the generation of multiple characters and complex scenes [Chefer et al. 2023]. To alleviate this drawback, some approaches introduce explicit spatial guidance in T2I generative models. For example, ControlNet [Zhang and Agrawala 2023] uses zero convolution layers and a trainable copy of the original model weights, introducing reliable control in diffusion models. T2I-Adapter [Mou et al. 2023] achieves control ability by proposing the adapter that extracts guidance feature and adds it to the feature from the corresponding UNet encoder. GLIGEN [Li et al. 2023] injects a gated self-attention block into the UNet, enabling the model to make good use of the grounding inputs. Inspired by the ability of large language models (LLMs) [et al 2023; OpenAI 2023] being able to understand and plan, recent works [Feng et al. 2023; Lian et al. 2023] employ LLM for layout generation. Specifically, LayoutGPT [Feng et al. 2023] achieves plausible results in 2D image layouts and even 3D indoor scene synthesis by applying in-context learning on LLMs. LLM-grounded Diffusion [Lian et al. 2023] proposes a two-stage process based on the LLM-generated layout and local prompts. Specifically, it first generates the local objects within each bounding box based on the corresponding local prompt, and then re-generates the final result based on the inversed latent of local objects. While effective, LLMgrounded Diffusion requires careful hyper-parameter tuning for the trade-off between structural guidance and inter-object relationship modeling. Moreover, it is difficult for the users to control the detailed structure of the generated objects. In contrast, we use the intuitive sketch or keypoint to guide the final image generation. Thus, we can not only achieve high-quality story image generation, but also allow interactive story visualization by simply tuning the generated sketch or keypoint conditions. 2.3 Customized Image Generation Story visualization requires that the identities of characters and scenes in a story remain consistent across different images. Customized image generation can meet this requirement to a large extent. Early methods [Gal et al. 2022; Ruiz et al. 2022] focus on the customized generation of a single object. For example, DreamBooth [Ruiz et al. 2022] fine tunes the pre-trained T2I diffusion model under a class-specific prior-preservation loss. Textual Inversion [Gal et al. 2022] enables customized generation by inverting subject image content into text embeddings. Unlike these approaches, Custom Diffusion [Kumari et al. 2022] further achieves multi-subject customization by combining the multiple customization weights through closed-form constrained optimization. Cones [Liu et al. 2023a] finds that a small cluster of concept neurons in the diffusion model corresponds to a single subject, and thus achieves customized generation of multiple objects by combining these concept neurons. Cones2 [Liu et al. 2023f] further achieves more effective multi-object customization by combining text embedding of different concepts with simple layout control. Differently, Mix-of-Show [Gu et al. 2023] proposes gradient fusion to effectively combine multiple customized LoRA [Hu et al. 2022] weights and performs multi-object customization with the aid of the T2I-Adapter’s dense controls. While significant progress has been made, existing methods perform poorly on one-shot customization. The training data for subject-driven generation has to be identity-consistent and diverse. As a result, existing story visualization methods require multiple user-provided images for each character. To tackle this issue, we propose a training-free consistency modeling method, and leverage the 3D prior in 3D-aware generative models [Liu et al. 2023d,b] to obtain multi-view consistent character images for customized generation, thus eliminating the reliance on human labor to collect or draw character images. one-shot This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv

AutoStory: Generating Diverse Storytelling Images with Minimal Human Effort: Related Work

AutoStory: Generating Diverse Storytelling Images with Minimal Effort: Conclusion and References

AutoStory: Generating Diverse Storytelling Images with Minimal Human Effort: Abstract and Intro

Authors:
(1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn);
(2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn);
(3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn);
(4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn);
(5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com);
(6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Table of Links Abstract and Introduction
Related Work
Our Method
Experiments
Conclusion and References
Appendix A APPENDIX A.1 More Implementation Details Detailed Prompts for the LLM. As described in Sec. 3 in the main text, we utilize LLMs to accomplish the story and layout generation. Specifically, we leverage the LLM for (1) generating the story, (2) dividing the story into panels, and (3) generating prompts and layout from the panels. In implementation, we further split the third step into two sub-steps, where we first convert the text of each panel into prompts suitable for generating the image, and then parse the prompts into the layout and local prompts. The detailed prompts and sampled LLM outputs are shown in Fig. 11. More Details on the Main Results. Fig. 4 shows the story image generation results of our method with varying characters, storylines, and image styles. Here, we present the character images used to train the customized model for each story. The story visualization results in the left two columns in Fig. 4 are obtained with the user-supplied character images. The corresponding characters are shown in Fig. 9. Differently, the story visualization results in the right two columns are obtained with only the story texts as inputs, and the characters are automatically generated by our method. The generated images for each character are shown in Fig. 10. It can be seen that the animal and human characters generated by our method are of high quality and consistent identities. The images of a single animal character show high diversity, with the orientation of the bird and the cat changing constantly from left to right. The human characters, however, are slightly less diverse, with a lower degree of variance in facial orientation. We believe that this is mainly due to the fact that the diffusion model is trained primarily on humans with frontal faces, making it difficult to generate side-facing images. Nonetheless, the character image data generated by our method can be effectively used for training customization models in story visualization, without introducing overfitting. It is worth mentioning that even though the character Tom in our generated data wears suits, we can generate images with Tom wearing a T-shirt after we specify that the character wears a T-shirt in the local prompt, as shown in the story visualization in Fig. 4 (d). Moreover, the characteristics of Tom are well-maintained, such as the shape of his face and the white hair. This indicates that the customized model trained with our generated data learns the character’s identity without overfitting. A.2 Intermediate Results Visualization To better understand our approach, in this section, we visualize the intermediate process of generating a single story image, as shown in Fig. 12 and Fig. 13. We first generate singlecharacter images based on the Local prompts generated by LLM, as shown in (a) and (b). The perception models, including Grounding-SAM, PidiNet, and HRNet, are then utilized to obtain the keypoints of human characters, or sketches of nonhuman characters, as shown in (c) and (d). Subsequently, the LLM-generated layout is utilized to compose the keypoints or sketches of individual subjects into a dense condition for generating story images, as shown in (d). Finally based on the dense conditions, prompts, and layout, we generate the story image as shown in (f). A.3 More Story Visualization Results In Fig. 14 and Fig. 15, we showcase more story visualization results of our method. As can be seen, our AutoStory can produce high-quality, text-aligned, and identity-consistent story images, even when generating long stories. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. Authors: (1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn); (2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn); (3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn); (4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn); (5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com); (6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Authors: Authors: (1) Wen Wang, Zhejiang University, Hangzhou, China and Equal Contribution (wwenxyz@zju.edu.cn); (2) Canyu Zhao, Zhejiang University, Hangzhou, China and Equal Contribution (volcverse@zju.edu.cn); (3) Hao Chen, Zhejiang University, Hangzhou, China (haochen.cad@zju.edu.cn); (4) Zhekai Chen, Zhejiang University, Hangzhou, China (chenzhekai@zju.edu.cn); (5) Kecheng Zheng, Zhejiang University, Hangzhou, China (zkechengzk@gmail.com); (6) Chunhua Shen, Zhejiang University, Hangzhou, China (chunhuashen@zju.edu.cn). Table of Links Abstract and Introduction Related Work Our Method Experiments Conclusion and References Appendix Abstract and Introduction Abstract and Introduction Related Work Related Work Our Method Our Method Experiments Experiments Conclusion and References Conclusion and References Appendix Appendix A APPENDIX A.1 More Implementation Details Detailed Prompts for the LLM. As described in Sec. 3 in the main text, we utilize LLMs to accomplish the story and layout generation. Specifically, we leverage the LLM for (1) generating the story, (2) dividing the story into panels, and (3) generating prompts and layout from the panels. In implementation, we further split the third step into two sub-steps, where we first convert the text of each panel into prompts suitable for generating the image, and then parse the prompts into the layout and local prompts. The detailed prompts and sampled LLM outputs are shown in Fig. 11. Detailed Prompts for the LLM. More Details on the Main Results. Fig. 4 shows the story image generation results of our method with varying characters, storylines, and image styles. Here, we present the character images used to train the customized model for each story. The story visualization results in the left two columns in Fig. 4 are obtained with the user-supplied character images. The corresponding characters are shown in Fig. 9. Differently, the story visualization results in the right two columns are obtained with only the story texts as inputs, and the characters are automatically generated by our method. The generated images for each character are shown in Fig. 10. It can be seen that the animal and human characters generated by our method are of high quality and consistent identities. The images of a single animal character show high diversity, with the orientation of the bird and the cat changing constantly from left to right. The human characters, however, are slightly less diverse, with a lower degree of variance in facial orientation. We believe that this is mainly due to the fact that the diffusion model is trained primarily on humans with frontal faces, making it difficult to generate side-facing images. Nonetheless, the character image data generated by our method can be effectively used for training customization models in story visualization, without introducing overfitting. It is worth mentioning that even though the character Tom in our generated data wears suits, we can generate images with Tom wearing a T-shirt after we specify that the character wears a T-shirt in the local prompt, as shown in the story visualization in Fig. 4 (d). Moreover, the characteristics of Tom are well-maintained, such as the shape of his face and the white hair. This indicates that the customized model trained with our generated data learns the character’s identity without overfitting. More Details on the Main Results. A.2 Intermediate Results Visualization To better understand our approach, in this section, we visualize the intermediate process of generating a single story image, as shown in Fig. 12 and Fig. 13. We first generate singlecharacter images based on the Local prompts generated by LLM, as shown in (a) and (b). The perception models, including Grounding-SAM, PidiNet, and HRNet, are then utilized to obtain the keypoints of human characters, or sketches of nonhuman characters, as shown in (c) and (d). Subsequently, the LLM-generated layout is utilized to compose the keypoints or sketches of individual subjects into a dense condition for generating story images, as shown in (d). Finally based on the dense conditions, prompts, and layout, we generate the story image as shown in (f). A.3 More Story Visualization Results In Fig. 14 and Fig. 15, we showcase more story visualization results of our method. As can be seen, our AutoStory can produce high-quality, text-aligned, and identity-consistent story images, even when generating long stories. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv

AutoStory: Generating Diverse Storytelling Images with Minimal Human Effort: Appendix

Discard Manual Scheduling With DolphinScheduler 3.1.x Cluster Deployment

Authors: (1) Sha-Bo Lin, Center for Intelligent Decision-Making and Machine Learning, School of Management, Xi’an Jiaotong University; (2) Xingping Sun, Department of Mathematics, Missouri State University; (3) Di Wang, §Center for Intelligent Decision-Making and Machine Learning, School of Management, Xi’an Jiaotong University. Content Overview Abstract & Introduction Uncertainty Relation of Kernel Interpolation on Spheres Distributed Kernel Interpolation Operator Differences via Quadrature Rules Proofs Numerical Verifications References Abstract For radial basis function (RBF) kernel interpolation of scattered data, Schaback [30] in 1995 proved that the attainable approximation error and the condition number of the underlying interpolation matrix cannot be made small simultaneously. He referred to this finding as an “uncertainty relation”, an undesirable consequence of which is that RBF kernel interpolation is susceptible to noisy data. In this paper, we propose and study a distributed interpolation method to manage and quantify the uncertainty brought on by interpolating noisy spherical data of non-negligible magnitude. We also present numerical simulation results showing that our method is practical and robust in terms of handling noisy data from challenging computing environments. Kernel interpolation, distributed uncertainty mitigation, scattered spherical data Key words. Uncertainty Relation of Kernel Interpolation on Spheres 3. Distributed Kernel Interpolation. Corollary 2.2 shows that kernel interpolation performs poorly while confronting noisy data of non-negligible magnitude. To overcome this major drawback, we propose and study in this section a distributed kernel interpolation (DKI) method, which is motivated by the “distributed learning” in the literature [37, 19]. Figuratively speaking, this is a divide-and-conquer strategy for uncertainty quantification. To elaborate, we describe the method in three steps. 4. Operator Differences via Quadrature Rules. In this section, we first briefly exposit an integral-operator approach initiated in [8] and then derive tight upper bounds for differences of operators of our interest, obtaining a certain type of Sobolev sampling inequalities [12] as a byproduct. Highlights of the section include Proposition 4.5) and Lemma 4.8. 6. Numerical Verifications Four simulations are carried out in this section to verify the excellent performance of DKI. The first one shows that DKI succeeds in circumventing the uncertainty of kernel interpolation. The second one exhibits the role of m in DKI. The third one focuses on the role of the division strategy in DKI. The last one compares DKI with several popular spherical data fitting schemes including the distributed filtered hyperinterpolation (DFH) [21], sketching with s ∗ -designs [20], and distributed kernel ridge regression (DKRR) [8]. In this simulation, we show the role of the parameter m in DKI. We generate 10014 training samples (with 141-designs as inputs). The number of divisions, m, ranges from {5, 10, · · · , 200}. Figure 6.2 shows the relation between RMSE of DKI and the number of local machines under different levels of Gaussian noise, provided that the total number of training samples is given. From Figure 6.2, we can conclude the following assertions: 1) For training samples with higher levels of noise, the testing RMSE generally decreases at first and then increases slowly as the number of local machines increases. Moderate values of m are more conductive to good approximation property for DKI. The reason is that too small m does not successfully address the uncertainty issue in kernel interpolation; too large m increases the fitting error, resulting in slightly worse generalization performance. 2) The optimal number m with the lowest RMSE grows with increasing Gaussian noise. This verifies the equation (3.3) of Theorem 3.2, in which the approximation error is primarily concerned with the sample error for large noise (i.e., large M) and can be reduced using a large m. Simulation 2: REFERENCES [1] R. Bhatia, Matrix Analysis, vol. 169, Springer Science & Business Media, 2013. [2] J. S. Brauchart and K. Hesse, Numerical integration over spheres of arbitrary dimension, Constructive Approximation, 25 (2007), pp. 41–71. [3] G. Brown and F. Dai, Approximation of smooth functions on compact two-point homogeneous spaces, Journal of Functional Analysis, 220 (2005), pp. 401–423. [4] A. Chernih, I. H. Sloan, and R. S. Womersley, Wendland functions with increasing smoothness converge to a gaussian, Advances in Computational Mathematics, 40 (2014), pp. 185– 200. [5] F. Dai, Multivariate polynomial inequalities with respect to doubling weights and a∞ weights, Journal of Functional Analysis, 235 (2006), pp. 137–170. [6] F. Dai, On generalized hyperinterpolation on the sphere, Proceedings of the American Mathematical Society, 134 (2006), pp. 2931–2941. [7] H. W. Engl, M. Hanke, and A. Neubauer, Regularization of Inverse Problems, vol. 375, Springer Science & Business Media, 1996. [8] H. Feng, S.-B. Lin, and D.-X. Zhou, Radial basis function approximation with distributively stored data on spheres, arXiv:2112.02499, (2021). [9] T. Hangelbroek, F. J. Narcowich, X. Sun, and J. D. Ward, Kernel approximation on manifolds ii: The l∞ norm of the l2 projector, SIAM Journal on Mathematical Analysis, 43 (2011), pp. 662–684. [10] T. Hangelbroek, F. J. Narcowich, and J. D. Ward, Kernel approximation on manifolds i: bounding the lebesgue constant, SIAM Journal on Mathematical Analysis, 42 (2010), pp. 1732–1760. [11] K. Hesse, I. H. Sloan, and R. S. Womersley, Radial basis function approximation of noisy scattered data on the sphere, Numerische Mathematik, 137 (2017), pp. 579–605. [12] K. Hesse, I. H. Sloan, and R. S. Womersley, Local rbf-based penalized least-squares approximation on the sphere with noisy scattered data, Journal of Computational and Applied Mathematics, 382 (2021), p. 113061. [13] S. Hubbert, Q. T. Le Gia, and T. M. Morton ˆ , Spherical radial basis functions, theory and applications, Springer, 2015. [14] M. A. King, R. J. Bingham, P. Moore, P. L. Whitehouse, M. J. Bentley, and G. A. Milne, Lower satellite-gravimetry estimates of antarctic sea-level contribution, Nature, 491 (2012), pp. 586–589. [15] Q. T. Le Gia, F. J. Narcowich, J. D. Ward, and H. Wendland, Continuous and discrete least-squares approximation by radial basis functions on spheres, Journal of Approximation Theory, 143 (2006), pp. 124–133. [16] P. Leopardi, A partition of the unit sphere into regions of equal area and small diameter, Electronic Transactions on Numerical Analysis, 25 (2006), pp. 309–327. [17] J. Levesley, Z. Luo, and X. Sun, Norm estimates of interpolation matrices and their inverses associated with strictly positive definite functions, Proceedings of the American Mathematical Society, (1999), pp. 2127–2134. [18] S.-B. Lin, X. Chang, and X. Sun, Kernel interpolation of high dimensional scattered data, arXiv preprint arXiv:2009.01514, (2020). [19] S.-B. Lin, X. Guo, and D.-X. Zhou, Distributed learning with regularized least squares, The Journal of Machine Learning Research, 18 (2017), pp. 3202–3232. [20] S.-B. Lin, D. Wang, and D.-X. Zhou, Sketching with spherical designs on spheres, SIAM Journal on Scientific Computation, in press (2023). [21] S.-B. Lin, Y. G. Wang, and D.-X. Zhou, Distributed filtered hyperinterpolation for noisy data on the sphere, SIAM Journal on Numerical Analysis, 59 (2021), pp. 634–659. [22] J. D. McEwen and Y. Wiaux, A novel sampling theorem on the sphere, IEEE Transactions on Signal Processing, 59 (2011), pp. 5876–5887. [23] H. Mhaskar, F. Narcowich, and J. Ward, Spherical marcinkiewicz-zygmund inequalities and positive quadrature, Mathematics of Computation, 70 (2001), pp. 1113–1130. [24] H. N. Mhaskar, Weighted quadrature formulas and approximation by zonal function networks on the sphere, Journal of Complexity, 22 (2006), pp. 348–370. [25] C. Muller ¨ , Spherical Harmonics, vol. 17, Springer, 1966. [26] F. J. Narcowich, N. Sivakumar, and J. D. Ward, Stability results for scattered-data interpolation on euclidean spheres, Advances in Computational Mathematics, 8 (1998), pp. 137– 163. [27] F. J. Narcowich, X. Sun, J. D. Ward, and H. Wendland, Direct and inverse sobolev error estimates for scattered data interpolation via spherical basis functions, Foundations of Computational Mathematics, 7 (2007), pp. 369–390. [28] F. J. Narcowich and J. D. Ward, Scattered data interpolation on spheres: error estimates and locally supported basis functions, SIAM Journal on Mathematical Analysis, 33 (2002), pp. 1393–1410. [29] A. Rudi, R. Camoriano, and L. Rosasco, Less is more: Nystr¨om computational regularization., in NIPS, 2015, pp. 1657–1665. [30] R. Schaback, Error estimates and condition numbers for radial basis function interpolation, Advances in Computational Mathematics, 3 (1995), pp. 251–264. [31] S. Smale and D.-X. Zhou, Shannon sampling and function reconstruction from point values, Bulletin of the American Mathematical Society, 41 (2004), pp. 279–305. [32] S. Smale and D.-X. Zhou, Shannon sampling ii: Connections to learning theory, Applied and Computational Harmonic Analysis, 19 (2005), pp. 285–302. [33] Y.-T. Tsai and Z.-C. Shih, All-frequency precomputed radiance transfer using spherical radial basis functions and clustered tensor approximation, ACM Transactions on Graphics (TOG), 25 (2006), pp. 967–976. [34] H. Wendland, Scattered data approximation, vol. 17, Cambridge university press, 2004. [35] M. A. Wieczorek and R. J. Phillips, Potential anomalies on a sphere: Applications to the thickness of the lunar crust, Journal of Geophysical Research: Planets, 103 (1998), pp. 1715–1724. [36] R. S. Womersley, Efficient spherical designs with good geometric properties, in Contemporary computational mathematics-A celebration of the 80th birthday of Ian Sloan, Springer, 2018, pp. 1243–1285. [37] Y. Zhang, J. Duchi, and M. Wainwright, Divide and conquer kernel ridge regression: A distributed algorithm with minimax optimal rates, The Journal of Machine Learning Research, 16 (2015), pp. 3299–3340. In this Appendix, we present the detailed implementation of the select-and-judge (SAJ) strategy. Our aim is to derive a series of subsets of similar Cardinality with separation radius not smaller than a given tolerance c0. There are two stages for SAJ. SM1. Appendix A: Select-and-Judge Strategy for Data Division. This paper is under CC0 1.0 DEED license. available on arxiv

Distributed Uncertainty Quantification of Kernel Interpolation on Spheres

গোলকের উপর কার্নেল ইন্টারপোলেশনের বিতরণ করা অনিশ্চয়তার পরিমাণ

Verteilte Unsicherheitsquantifizierung der Kernelinterpolation auf Kugeln

Cuantificación de la incertidumbre distribuida de la interpolación del kernel en esferas

Quantification distribuée de l'incertitude de l'interpolation du noyau sur les sphères

क्षेत्रों पर कर्नेल इंटरपोलेशन की वितरित अनिश्चितता मात्रा

球上のカーネル補間の分散不確実性の定量化

구에 대한 커널 보간법의 분산 불확실성 정량화

Quantificação de incerteza distribuída de interpolação de kernel em esferas

Количественная оценка распределенной неопределенности ядра интерполяции на сферах

Küreler Üzerindeki Çekirdek Enterpolasyonunun Dağıtılmış Belirsizliği Ölçümü

Định lượng độ không đảm bảo phân phối của phép nội suy hạt nhân trên hình cầu

球体核插值的分布式不确定性量化

This paper is under CC 4.0 license. available on arxiv : Authors (1) Linda Pigureddu, University of Turin, Italy, 332079@edu.unito.it; (2) Cristina Gena, Dept. of Computer Science, University of Turin, Italy, cristina.gena@unito.it. Table of Links Introduction The Project Memes Conclusion and References CONCLUSION The analysis of the interactions and dialogues that took place during the therapeutic laboratory for autonomies endorsed with the inclusion of the Pepper robot provided an interesting perspective on the communication dynamics of children in the lab. The results obtained showed the effectiveness of the robot in promoting autonomy and functional acquisition, demonstrating that the therapeutic method based on assistive robotics is a valuable resource to support the rehabilitation needs of communication and social skills of autistic children. The analysis highlights new possibilities for their engagement and active participation as a co-designer, as we already experienced in the past with neuro-typical children [19], representing an innovative opportunity to promote the development and wellness of autistic people and opening to new perspectives for therapeutic intervention more aware of the needs of autistic minds. In conclusion, this paper highlights the need to provide the robot with the ability of adapt to the children’s peculiarity and features and sharing her/his vocabulary, especially by recognizing and using memes during interactions, to inspire greater trust in children and allowing the use of common slang, already in use with classmates, allowing them to consider Pepper a peer and effectively insert it in the role of mediator. In addition, it would allow children to use a simplified type of communication during meetings to compensate for the typical communication deficits associated with autism spectrum disorders. In the future we will work in this directions (co-design, user’s adaptations, using and recognizing memes during children-robot communication) having the robot able to communicate and express adapting to user tastes and preferences and and try again to field a real-world evaluation that takes into account the effectiveness of different levels of user adaptation [23]. REFERENCE [1] M. Biondi e M. Maj, DSM-5: diagnostic and statistical manual of mental disorders: text revision, 5. ed. Milano: Raffaello Cortina. [2] E. Hollander, R. Hagerman, e C. Ferretti, Textbook of Autism Spectrum Disorders, Second Edition. American Psychiatric Association Publishing, 2022. [3] ISS - Istituto superiore di sanità, «Osservatorio Nazionale Autismo», OssNA. https://osservatorionazionaleautismo.iss.it [4] P. Pennisi et al., «Autism and social robotics: A systematic review», Autism Research, vol. 9, fasc. 2, pp. 165–183, 2016, doi: 10.1002/aur.1527. [5] S. Shamsuddin, H. Yussof, S. Mohamed, e F. A. Hanapiah, «Design and Ethical Concerns in Robotic Adjunct Therapy Protocols for Children with Autism», Procedia Computer Science, vol. 42, pp. 9–16, gen. 2014, doi: 10.1016/j.procs.2014.11.027. [6] K. Dautenhahn e I. Werry, «Towards Interactive Robots in Autism Therapy: Background, Motivation and Challenges», Pragmatics & Cognition, vol. 12, pp. 1–35, giu. 2004, doi: 10.1075/pc.12.1.03dau. [7] C. Li, Q. Jia, e Y. Feng, «Human-Robot Interactoin Design for Robot-Assisted Intervention for Children with Autism Based on E-S Theory», in 2016 8th International Conference on Intelligent Human-Machine Systems and Cybernetics (IHMSC), Hangzhou, China: IEEE, ago. 2016, pp. 320–324. doi: 10.1109/IHMSC.2016.103. [8] A. Peca et al., «Robot Enhanced Therapy for Children with Autism Disorders: Measuring Ethical Acceptability», IEEE Technol. Soc. Mag., vol. 35, fasc. 2, pp. 54–66, giu. 2016, doi: 10.1109/MTS.2016.2554701. [9] B. Robins, K. Dautenhahn, e J. Dubowski, «Does appearance matter in the interaction of children with autism with a humanoid robot?», IS, vol. 7, fasc. 3, pp. 479–512, nov. 2006, doi:10.1075/is.7.3.16rob. [10] E. S. Kim, «Robots for social skills therapy in autism: evidence and designs toward clinical utility», 2013, doi: 10.13140/2.1.3038.2083. [11] «Autismo», Neuraxpharm Italia. https://www.neuraxpharm.com/it/disordini/autismo [12] L. Hull, K. V. Petrides, e W. Mandy, «The Female Autism Phenotype and Camouflaging: a Narrative Review», Rev J Autism Dev Disord, vol. 7, fasc. 4, pp. 306–317, dic. 2020, doi: 10.1007/s40489-020-00197-9. [13] D. Cohen et al., «NAO, a humanoid robot as a therapeutic mediator for young people with autism». [14] Cristina Gena, Claudio Mattutino, Andrea Maieli, Elisabetta Miraglio, Giulia Ricciardiello, Rossana Damiano, Alessandro Mazzei: Autistic Children's Mental Model of an Humanoid Robot. UMAP (Adjunct Publication) 2021: 128-129 [15] Rossana Damiano, Cristina Gena, Andrea Maieli, Claudio Mattutino, Alessandro Mazzei, Elisabetta Miraglio, Giulia Ricciardiello: UX Personas for Defining Robot's Character and Personality 213-216. IUI Workshops 2022: 213-216. [16] Cristina Gena, Claudio Mattutino, Stefania Brighenti, Andrea Meirone, Francesco Petriglia, Loredana Mazzotta, Federica Liscio, Matteo Nazzario, Valeria Ricci, Camilla Quarato, Cesare Pecone, Giuseppe Piccinni: et al., Sugar, Salt & Pepper - Humanoid robotics for autism, in Joint Proceedings of the ACM IUI 2021 Workshops co-located with 26th ACM Conference on Intelligent User Interfaces (ACM IUI 2021), College Station, United States, April 13-17, 2021, D. Glowacka e V. R. Krishnamurthy, A c. di, in CEUR Workshop Proceedings, vol. 2903. CEUR-WS.org, 2021. [17] Cristina Gena, Rossana Damiano, Claudio Mattutino, Alessandro Mazzei, Andrea Meirone, Loredana Mazzotta, Matteo Nazzario, Valeria Ricci, Stefania Brighenti, Federica Liscio, Francesco Petriglia: Preliminary results of a therapeutic lab for promoting autonomies in autistic children. arXiv, doi: 10.48550/arXiv.2305.02982. [18] B. C. Carbo, «The Use of Social Stories with Individuals with Autism Spectrum Disorders», University of Delaware ProQuest Dissertations Publishing, Newark, DE, 2005. [Online]. Disponibile su: https://udspace.udel.edu/server/api/core/bitstreams/cd5e8850-e57c-41d2-a7b4ac5e25a4b659/content [19] V. Cietto, C. Gena, I. Lombardi, C. Mattutino and C. Vaudano, "Co-designing with kids an educational robot," 2018 IEEE Workshop on Advanced Robotics and its Social Impacts (ARSO), Genova, Italy, 2018, pp. 139-140, doi: 10.1109/ARSO.2018.8625810. [20] A. Mousa, «Sherry Turkle, Reclaiming Conversation. The Power of Talk in a Digital Age», Questions de communication, fasc. 34, Art. fasc. 34, dic. 2018, doi:10.4000/questionsdecommunication.16891. [21] H.-M. Chiang e M. Carter, «Spontaneity of Communication in Individuals with Autism», J Autism Dev Disord, vol. 38, fasc. 4, pp. 693–705, apr. 2008, doi: 10.1007/s10803-007-0436-7. [22] Gena, C., Cena, F., Vernero, F., Grillo, P. The evaluation of a social adaptive website for cultural events. User Model User-Adap Inter 23, 89–137 (2013) https://doi.org/10.1007/s11257-012-9129-9 [23] Gena, Cristina, and Ilaria Torre. "The importance of adaptivity to provide onboard services: A preliminary evaluation of an adaptive tourist information service onboard vehicles." Applied Artificial Intelligence 18.6 (2004): 549-580.

The Power of Memes: The Pepper Robot as a Communicative Aid for Autistic Children - Conclusion

306 Stories To Learn About Artificial Intelligence

Where do they live exactly? What happens to them when they get lost? Will they ever die, like all software before them? How?? These and more questions, in this short story. Economics Satoshi Nakamoto set into motion the creation, and eventual destruction, of all 21 million bitcoins in existence. But like entropy creates wonderful things, such as the evolution of complex life forms, so too does Bitcoin and its halvings, with the evolution of cooperative technologies like the lightning network, multi-sig, conjoin, zero-knowledge proofs, and even the trail of 30,000 crypto projects tagging behind can be thought of as Bitcoin’s children. Eating its leftovers, and whatever it is too proud to eat. Especially important is how Bitcoin facilitates the speedy collapse of ponzi-scheming bank-like businesses like FTX. If Bitcoin had not existed, the FED would have given SBF a nice fat loan to save his a**. Alas, if you lose Bitcoins and try to buy them back with a loan, the price shoots up fast to expose you and your cronies. Bitcoin is thus the best moderator of the government’s appetite for expanding credit. Something which if not reigned in, helps fuel not just economic collapse but World Wars. Yes, the World Wars were financed by credit expansion. No more. There just isn’t enough room for credit expansion like that, for governments to misuse. And it could get smaller. But let us dial back to the beginning. Back to the days when CDOs wrecked the world’s financial markets. Origin Once upon a time, location unknown, it was 2008. There was probably noise outside from people shouting for the heads of the Wall Street moguls who had mercilessly vaporized their life savings. Sitting in his room and most comforted by the fact that he was a nobody, unfamous and unrecognized, was the man who would change the world of digital finance. The man who would put strange new ideas in the minds of kids dumping their Physics textbooks in the toilet, to go learn how to be Wolves of Wall Street. Dollar signs in their eyes, scientific reality be damned. This man was called __. That’s right, I do not know his name. Thankfully, he has a name he wants me to call him. “ ” Satoshi Nakamoto. The legend. The man whose pseudo-name will be used for the smallest unit of the Bitcoin cryptocurrency. Today is the day he puts it all together. It weighs virtually nothing at all. And if it were a book, it would hardly beat the Oxford English dictionary. It is the code of Bitcoin, his baby. An alien creature in the software world. With elements immune to copying the more users get a copy of the code. Elements that ossify, like energetic magma turning into precious rocks. All other software has elements that are easier to copy, paste, modify, and hack, the more users get their hands on it. Makes sense right? I mean, more users, more developer intelligence to hack the ossified parts? Wrong. These elements are the ones they will all call bitcoins, or satoshis. It is time to breed them on his laptop. See how they behave. He presses the RETURN key. Deep inside his laptop, a series of quantum mechanical events are set into motion to produce these bitcoins. It is an intricate dance, and if the clock is not synchronized just right, if the CPUs are not well-dusted, bye-bye genius idea. So, he has reviewed the code obsessively like a maniac prisoner on death row, reading and re-reading the constitution to save his ass from hanging. 10 minutes later, the first 50 BTCs are born. He punches the air. ‘Whoop’. He loves code. God bless all coders on Earth. If everyone on earth were a coder, nobody would be cheated by big finance, propagandized by big news, or surveilled by big politics. Quickly, he puts his favorite newspaper heading on it. Timestamps it for all the world to see, for as long as his creation lives. “The Times 03/Jan/2009/ Chancellor on Brink of second bailout for banks” Satoshi probably runs an O/S alike Kali Linux (which will be invented in 2013), with feelers straightened out for the slightest whiff of eavesdropping in the air. His laptop has been air-gapped these last many years, but now, he has to connect to the jungle that is the internet. Like any jungle, there is food, and there are predators. There is also a ton of mediocrity and below all that, even more sh*t. Tomorrow, he will unleash it, let it go native. Bitcoin belongs in this jungle, not his laptop. It will become a predator like no other. Would you look at that, some predator is sniffing at it already. Aargh, doesn’t smell like a Social Security number. It moves on. Vanishes more like. Haha. Satoshi laughs. He loves seeing these bots in action, zapping to and fro the internet highway. While the average person sees nothing, with his Kali-Linux-like feelers, he catches the equivalent of shooting stars, trying to hit his Windows. They all vaporize before landing. If he wanted, he would be working for GF - Google & Facebook. Or, he’d have his own multimillion-dollar company. Or maybe, steal all of next year’s meals, and live in this basement for another year. Like some of these top-notch predators. Nah. This is better. This is the game. “Satoshi??” The wife is calling. “Satoshi” “Dear.” “Dinner is ready.” He stands up. Stretches his muscular body out real nice. He is actually hungry, starving. Divisibility Is Not Debasement He loves this internet place. Been a week since he started chatting with nice, super smart chaps who speak code, logic, and mathematics, and the reception he is getting is all good. Nobody is getting jealous that he made his own bitcoins in his bedroom. To the thousands now. Not yet. He hopes it doesn’t happen before a critical mass of support is reached. Turns out he actually forgot something. The guys remind him that 21 million bitcoins are not enough for 7 billion people. Strangely enough, it is data from the CIA that will help determine how many Bitcoin satoshis there should be. The CIA helps create Bitcoin! Without knowing, I should add. Here is the to the story. link The conversation below is not on any forum, but it might have happened: Satoshi: “So you want me to divide these bitcoins?” Hal Finney: “Yeah” Ray Dilinger: “Indeed” Satoshi: “Hope you know this isn’t like Jesus feeding the 5000. And I will still make the subdivisions immortal to copying and counted as divisions of the full BTC” Interestingly, a lot of people in the future, 15 years later, still do not get this part. They still think that because BTC can be divided into satoshis, it can be debased. It is like thinking if you take a $100 note to the bank and are returned 10,000 pennies, you are somehow poorer. Your money has been debased? Noo. You’re actually richer because all that metal is more valuable than a single paper note. This is a classic cognitive bias called ‘Anchoring’ – Jumping to conclusions too early. Let’s jump into the future. Which is the present. It is the future. On the 14th of January, 2024, tweep @SamanthaLaDuc got this infinite money idea and yapped about it all over the place. Samantha felt that because each bitcoin can be divided into 100 million smaller portions called satoshis or “sats’, Bitcoiners cannot claim scarcity or lack of dilution … She ends by saying she did MATH. How does it happen? How does the system know these satoshis are the only satoshis in the Bitcoin blockchain? Where are these 21 million bitcoins, or if you prefer, 2.1 quadrillion satoshis? And can we make more of them? There Can Only be 2.1 Quadrillion The first thing to tell us we may be at the limits of how many divisions Bitcoin can suffer is the IEEE-754 standard for 64-bit computers. As Bitzuma.com in the previously linked story writes: To avoid turning Bitcoin into an accounting minefield, the number of total currency units must be kept below the IEEE rounding threshold - now and a hundred years in the future. The IEEE rounding threshold is 2^53 which is approximately 9 x 10^18. 2.1 quadrillion satoshis is approximately 2 x 10^15, so if we were to divide our satoshis some more, we could only divide into 4500 smaller units before the ALU (Arithmetic and Logic Unit) inside most computer CPUs starts getting lost in math. Like LaDuc. Any larger and the Bitcoin blockchain is going to become very bad at keeping records of the number of bitcoins in wallets, bitcoin divisions during halvings, etc. We wouldn’t want that. So if these satoshis are pushing memory counters to the limit, then they are pushing our computers to the limit. Who would have thought infinitely making Bitcoin subdivisions is impossible? Well, now we know. On Blackholes and Quantum-Resistance Like people forget their car keys and house keys, people forget their private keys (or seed phrases) and once this happens, they can no longer sign transactions, and their Bitcoin satoshis or sats, are lost to cyberspace in the world that is the Bitcoin blockchain. They are not really lost though, these are misplaced, and there is hope that when technologies such as quantum computing mature, they will be retrieved. To briefly expound, this promise of quantum computing rests on the theoretical conclusions of an algorithm called Shor’s factoring algorithm. Shor’s factoring algorithm allows a quantum computer with a to crack open a Bitcoin wallet address without the need for a private key. couple of million logical qubits The power of this algorithm provides logarithmic speedup, so it is really fast if we can get our hands on those logical qubits. Like you press the button, and only 8 hours later, the wallet is cracked open. No sweat. An alternative quantum algorithm is called Grover’s search algorithm, and this one provides a quadratic speedup. So, it is not as fruitful as Shor’s. This one would work by super-brute-force searching for all possible prime factors to eventually land on the right ones that form the correct private key. This is why today all sorts of quantum-resistant cryptographic schemes are being hatched. But I find this sort of thing a little ridiculous. Like building bunkers before splitting the atom, we are currently building digital bunkers before splitting the cryptographic atom. But why it is happening is because it is big government that is most scared of a quantum hacking system. A nuke empowers a government to threaten another with destruction, so that is why Oppenheimer and his buddies got a private village and many billions of dollars to build this weapon. A quantum hacking system, on the other hand, targets individually owned encrypted addresses, so the chance of it backfiring on its creators is a big one. For we are all individuals. Better to NOT create it, hence big-time pioneering work in quantum resistance. Bitcoiners are quite safe, and they really shouldn’t worry. A quantum hacking system will be deployed first on centralized systems with big secrets locked behind RSA and SHA-2 encryption before the hackers think of stealing your couple of Bitcoins. Moreover, the energy budget of running the hack on a one million logical qubit system, which will be something like a one billion physical qubit system, will be prohibitive. We are talking about a computer bigger than a plane hanger, with all parts cooled to nearly absolute zero. The return on that investment is only of seismic proportions if we are hacking something that should be hacked. Endnotes I have yet to mention black holes. Satoshis that get misplaced or inaccessible within the blockchain can be recovered, and are like items lost in the vast expanse of space. Except this is cyberspace. Hacking systems are like a quick way to search for these items without using eyes e.g., if you have a big powerful magnet, you can attract a bunch of keys lost in outer space. However, satoshis can be burned. Destroyed forever. These satoshis are said to have entered a black hole and technically, are as gone as data sent into the recycle bin on your PC and recycled, is gone. Crushed and turned into hawking radiation if it is a black hole, or heat if it is the recycle bin. Basically, all memory circuits are reset to 0. Releasing all the energy stored in various parts that were holding onto a 1. As the Bitzuma article I shared shows, a lot of satoshis have been burned already. And people keep burning some by sending them into the Genesis block, as a way of paying homage to Satoshi. The Genesis block is a black hole for none of the sats that go in there will ever return. When Bitcoin finally runs out of worldwide utility, perhaps because mining blocks is a net negative expense and too many satoshis are lost anyway, people will discard their hard drives containing the Genesis block. They will send it to the recycle bin, essentially finalizing the demise of all those satoshis in there. Like the entropic decay of the universe, Bitcoin satoshis are decaying into waste heat. But no worries, we can start it all over again. Perhaps better, for we shall have quantum networks then, hence, the ability to create truly synchronized blockchain operations using entanglement.

21M Bitcoins/2.1 Quadrillion Satoshis

Authors: (1) Nowshin Nawar Arony; (2) Ze Shi Li; (3) Bowen Xu; (4) Daniela Damian. Table of Links Abstract & Introduction Motivation Related Work Methodology A Taxonomy of Inclusiveness Inclusiveness Concerns in Different Types of Apps Inclusiveness Across Different Sources of User Feedback Automated Identification of Inclusiveness User Feedback Discussion Conclusion & References 10 CONCLUSION Our study follows a socio-technical grounded theory approach to gain a deeper understanding of inclusiveness related user feedback from end users. Across manual analysis of over 23K user feedback posts from Reddit, Twitter, and Google Play Store regarding 50 popular for profit apps, we build a taxonomy of inclusiveness. Our taxonomy has six main categories, including fairness, technology, privacy, demography, usability, and other human values. The classifier that we train on our data shows that we can automatically identify inclusiveness related feedback among general user feedback. Our results indicate to practitioners that these online sources contain a rich trove of inclusiveness feedback that organizations should consider to build more inclusive software products for diverse end users. We also present our labelled dataset that researchers can use to refine tooling to better support practitioners. REFERENCES [1] A. Savidis and C. Stephanidis, “Inclusive development: Software engineering requirements for universally accessible interactions,” Interacting with Computers, vol. 18, no. 1, pp. 71–116, 2006. [2] J. Shepherd, “23 essential twitter statistics you need to know in 2023,” https://thesocialshepherd.com/blog/twitterstatistics#: :text=TwitterMay 2023. [3] T. Gaurdian, “Twitter apologises for ’racist’ image-cropping algorithm,” https://www.theguardian.com/technology/2020/sep/21/twitterapologises-for-racist-image-cropping-algorithm, 2020. [4] Z. S. Li, M. Sihag, N. N. Arony, J. B. Junior, T. Phan, N. Ernst, and D. Damian, “Narratives: the unforeseen influencer of privacy concerns,” in 2022 IEEE 30th International Requirements Engineering Conference (RE). IEEE, 2022, pp. 127–139. [5] M. Shahin, M. Zahedi, H. Khalajzadeh, and A. R. Nasab, “A study of gender discussions in mobile apps,” arXiv preprint arXiv:2303.09808, 2023. [6] A. Alshayban, I. Ahmed, and S. Malek, “Accessibility issues in android apps: state of affairs, sentiments, and ways forward,” in Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering, 2020, pp. 1323–1334. [7] H. Khalajzadeh, M. Shahin, H. O. Obie, P. Agrawal, and J. Grundy, “Supporting developers in addressing human-centric issues in mobile apps,” IEEE Transactions on Software Engineering, 2022. [8] J. Tizard, T. Rietz, X. Liu, and K. Blincoe, “Voice of the users: an extended study of software feedback engagement,” Requirements Engineering, vol. 27, no. 3, pp. 293–315, 2022. [9] R. Hoda, “Socio-technical grounded theory for software engineering,” IEEE Transactions on Software Engineering, vol. 48, no. 10, pp. 3808–3832, 2021. [10] K. B. M. , “Designing inclusive software in Windows - Windows apps,” https://learn.microsoft.com/enus/windows/apps/design/accessibility/designing-inclusivesoftware, may 13 2022. [11] M. Pattison and A. W. Stedmon, “Inclusive design and human factors: Designing mobile phones for older users.” PsychNology J., vol. 4, no. 3, pp. 267–284, 2006. [12] D. Metaxa-Kakavouli, K. Wang, J. A. Landay, and J. Hancock, “Gender-inclusive design: Sense of belonging and bias in web interfaces,” in Proceedings of the 2018 CHI Conference on human factors in computing systems, 2018, pp. 1–6. [13] M. Burnett, S. Stumpf, J. Macbeth, S. Makri, L. Beckwith, I. Kwan, A. Peters, and W. Jernigan, “Gendermag: A method for evaluating software’s gender inclusiveness,” Interacting with Computers, vol. 28, no. 6, pp. 760–787, 2016. [14] I. Nunes, A. Moreira, and J. Araujo, “Conceptual modeling of gender-inclusive requirements,” in International Conference on Conceptual Modeling. Springer, 2021, pp. 395–409. [15] ——, “Gire: Gender-inclusive requirements engineering,” Data & Knowledge Engineering, vol. 143, p. 102108, 2023. [16] M. Vorvoreanu, L. Zhang, Y.-H. Huang, C. Hilderbrand, Z. SteineHanson, and M. Burnett, “From gender biases to gender-inclusive design: An empirical investigation,” in Proceedings of the 2019 CHI Conference on human factors in computing systems, 2019, pp. 1–14. [17] J. C. Grundy, “Impact of end user human aspects on software engineering.” in ENASE, 2021, pp. 9–20. [18] J. Grundy, I. Mueller, A. Madugalla, H. Khalajzadeh, H. O. Obie, J. McIntosh, and T. Kanij, “Addressing the influence of end user human aspects on software engineering,” in International Conference on Evaluation of Novel Approaches to Software Engineering. Springer, 2021, pp. 241–264. [19] S. Donetto, P. Pierri, V. Tsianakas, and G. Robert, “Experiencebased co-design and healthcare improvement: realizing participatory design in the public sector,” The Design Journal, vol. 18, no. 2, pp. 227–248, 2015. [20] E. C. Groen, J. Doerr, and S. Adam, “Towards crowd-based requirements engineering a research preview,” in Requirements Engineering: Foundation for Software Quality: 21st International Working Conference, REFSQ 2015, Essen, Germany, March 23-26, 2015. Proceedings 21. Springer, 2015, pp. 247–253. [21] D. Pagano and B. Bruegge, “User involvement in software evolution practice: A case study,” in 2013 35th International Conference on Software Engineering (ICSE). IEEE, 2013, pp. 953–962. [22] E. C. Groen, N. Seyff, R. Ali, F. Dalpiaz, J. Doerr, E. Guzman, M. Hosseini, J. Marco, M. Oriol, A. Perini et al., “The crowd in requirements engineering: The landscape and challenges,” IEEE software, vol. 34, no. 2, pp. 44–52, 2017. [23] W. Maalej and H. Nabil, “Bug report, feature request, or simply praise? on automatically classifying app reviews,” in 2015 IEEE 23rd international requirements engineering conference (RE). IEEE, 2015, pp. 116–125. [24] G. Williams and A. Mahmoud, “Mining twitter feeds for software user requirements,” in 2017 IEEE 25th International Requirements Engineering Conference (RE). IEEE, 2017, pp. 1–10. [25] T. Iqbal, M. Khan, K. Taveter, and N. Seyff, “Mining reddit as a new source for software requirements,” in 2021 IEEE 29th International Requirements Engineering Conference (RE). IEEE, 2021, pp. 128–138. [26] M. Fazzini, H. Khalajzadeh, O. Haggag, Z. Li, H. Obie, C. Arora, W. Hussain, and J. Grundy, “Characterizing human aspects in reviews of covid-19 apps,” in Proceedings of the 9th IEEE/ACM International Conference on Mobile Software Engineering and Systems, 2022, pp. 38–49. [27] L. Olson, E. Guzman, and F. Kunneman, “Along the margins: ´ Marginalized communities’ ethical concerns about social platforms,” arXiv preprint arXiv:2304.08882, 2023. [28] E. Guzman, M. Ibrahim, and M. Glinz, “A little bird told me: Mining tweets for requirements and software evolution,” in 2017 IEEE 25th International Requirements Engineering Conference (RE). IEEE, 2017, pp. 11–20. [29] Watchful1, “Subreddit comments/submissions 2005- 06 to 2022-12.” [Online]. Available: https: //www.reddit.com/r/pushshift/comments/11ef9if/separate dump files for the top 20k subreddits/ [30] N. N. Arony, Z. S. Li, B. Xu, and D. Damian, “Inclusiveness Matters: A Large-Scale Analysis of User Feedback,” Oct. 2023. [Online]. Available: https://doi.org/10.5281/zenodo.10050673 [31] A. Radford, J. W. Kim, T. Xu, G. Brockman, C. McLeavey, and I. Sutskever, “Robust speech recognition via large-scale weak supervision,” in International Conference on Machine Learning. PMLR, 2023, pp. 28 492–28 518. [32] J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pretraining of deep bidirectional transformers for language understanding,” arXiv preprint arXiv:1810.04805, 2018. [33] Y. Liu, M. Ott, N. Goyal, J. Du, M. Joshi, D. Chen, O. Levy, M. Lewis, L. Zettlemoyer, and V. Stoyanov, “Roberta: A robustly optimized bert pretraining approach,” arXiv preprint arXiv:1907.11692, 2019. [34] A. Conneau, K. Khandelwal, N. Goyal, V. Chaudhary, G. Wenzek, F. Guzman, E. Grave, M. Ott, L. Zettlemoyer, and V. Stoyanov, ´ “Unsupervised cross-lingual representation learning at scale,” arXiv preprint arXiv:1911.02116, 2019. [35] M. Lewis, Y. Liu, N. Goyal, M. Ghazvininejad, A. Mohamed, O. Levy, V. Stoyanov, and L. Zettlemoyer, “Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension,” arXiv preprint arXiv:1910.13461, 2019. [36] M. Sihag, Z. S. Li, A. Dash, N. N. Arony, K. Devathasan, N. Ernst, A. B. Albu, and D. Damian, “A data-driven approach for finding requirements relevant feedback from tiktok and youtube,” in 2023 IEEE 31st International Requirements Engineering Conference (RE). IEEE, 2023, pp. 111–122. [37] R. Akbani, S. Kwek, and N. Japkowicz, “Applying support vector machines to imbalanced datasets,” in Machine Learning: ECML 2004: 15th European Conference on Machine Learning, Pisa, Italy, September 20-24, 2004. Proceedings 15. Springer, 2004, pp. 39–50. [38] S. Parashar, S. Dhar, and U. Dhar, “Perception of values: a study of future professionals,” Journal of Human Values, vol. 10, no. 2, pp. 143–152, 2004. [39] S. H. Schwartz, “Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries,” in Advances in experimental social psychology. Elsevier, 1992, vol. 25, pp. 1–65. [40] Y. Brun and A. Meliou, “Software fairness,” in Proceedings of the 2018 26th ACM joint meeting on european software engineering conference and symposium on the foundations of software engineering, 2018, pp. 754–759. [41] S. Akter, G. McCarthy, S. Sajib, K. Michael, Y. K. Dwivedi, J. D’Ambra, and K. N. Shen, “Algorithmic bias in data-driven innovation in the age of ai,” p. 102387, 2021. [43] J. Larson, S. Mattu, L. Kirchner, and J. Angwin, “How We Analyzed the COMPAS Recidivism Algorithm.” [Online]. Available: https://www.propublica.org/article/ how-we-analyzed-the-compas-recidivism-algorithm [44] A. Tobin and A. Kofman, “Facebook Finally Agrees to Eliminate Tool That Enabled Discriminatory Advertising,” Jun. 2022. [Online]. Available: https://www.propublica.org/article/ facebook-doj-advertising-discrimination-settlement [45] J. Chakraborty, S. Majumder, and T. Menzies, “Bias in machine learning software: Why? how? what to do?” in Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, 2021, pp. 429–440. [46] S. Krishnan, J. Patel, M. J. Franklin, and K. Goldberg, “A methodology for learning, analyzing, and mitigating social influence bias in recommender systems,” in Proceedings of the 8th ACM Conference on Recommender systems, 2014, pp. 137–144. [47] V. Tsintzou, E. Pitoura, and P. Tsaparas, “Bias disparity in recommendation systems,” arXiv preprint arXiv:1811.01461, 2018. [48] G. Hofstede, “Cultures and organizations: Software of the mind (london and new york, mcgraw hill),” House, RJ, Hanges, PJ, Javidan, M., Dorfman, PW, & Gupta, V.(Eds. 2004), Airaksinen, pp. 1–25, 1991. [49] N. Al-Huwail, S. Al-Sharhan, and A. Al-Hunaiyyan, “Learning design for a successful blended e-learning environment: Cultural dimensions,” INFOCOMP Journal of Computer Science, vol. 6, no. 4, pp. 60–69, 2007. [50] R. A.-L. Fischer, R. Walczuch, and E. Guzman, “Does culture matter? impact of individualism and uncertainty avoidance on app reviews,” in 2021 IEEE/ACM 43rd International Conference on Software Engineering: Software Engineering in Society (ICSE-SEIS). IEEE, 2021, pp. 67–76. [51] D. Potnis, “Culture’s consequences: Economic barriers to owning mobile phones experienced by women in india,” Telematics and Informatics, vol. 33, no. 2, pp. 356–369, 2016. [52] A. De Angeli, U. Athavankar, A. Joshi, L. Coventry, and G. I. Johnson, “Introducing atms in india: a contextual inquiry,” Interacting with Computers, vol. 16, no. 1, pp. 29–44, 2004. [53] M. R. Roller and P. J. Lavrakas, Applied qualitative research design: A total quality framework approach, ser. Applied qualitative research design: A total quality framework approach. New York, NY, US: The Guilford Press, 2015, pages: xviii, 398. This paper is under CC 4.0 license. available on arxiv

A Large-Scale Analysis of Inclusiveness-Related User Feedback : Conclusion & References

Scientific American, Volume XXIV., No. 12, March 18, 1871 by Various, is part of the HackerNoon Books Series. You can jump to any chapter in this book . To Manufacturers of Knitted Goods. here To Manufacturers of Knitted Goods. is now in England, completing arrangements so as to be able to supply his American friends with his improved Knitting Machines with greater dispatch, and with all the latest improvements. He would beg to call especial attention to JOHN KENT The Improved Rib Top Frame, now so well known, and acknowledged to be the best rib top frame ever built, for speed and quality of goods produced. Price, delivered free in New York, $520, currency. The Improved Circular Web Frame, for drawers and shirts, built of any size and gage. Price for a 4-head set, 17 inch to 20 inch diameter, $810, currency, delivered free in New York. The Circular Stocking Frame, from 2 in. to 5½ in. diameter. These circular frames, with my last improvements, are as near perfection as possible. The Patent Full-fashioned Shirt, Drawers and Stocking Frames produce the most perfect goods ever made by steam-power machinery, and cost fifty per cent less to keep in repair than any other Knitting Machine. Built 10 to 24 gage, and from 30 to 140 inch wide, to order. The Improved Circular Looping Frame, for putting on shirt cuffs, drawers bands, clearing the top of circular shirts, &c., built to order, of any size, from 2 in. to 22 in. diameter, and of any gage. Steel Needles and Sinkers to pattern. Persons wishing to order while Mr. Kent is in England, will please address JOHN KENT, Nottingham, up to April 12th, or, if they prefer, may send through depot. Address JOHN KENT, 348 Pearl st., New York. PUMPS. —For Description, Price Lists etc., of the Best Centrifugal Pump ever invented, with Overwhelming Testimony in its favor, send for new illustrated pamphlet (40 pp.) to Messrs. HEALD, SISCO & CO., Baldwinsville, N. Y. BRICK PRESSES. **FOR RED AND FIRE BRICK.
\ Factory 309 S. Fifth street, Philadelphia, Pa. S. P. MILLER. ST. JOSEPH, Mo., Nov. 10, 1870. T. R. BAILEY & VAIL, Lockport, N. Y.: Gentlemen:—The Lathe you shipped me has arrived, and I have it in full operation. It works perfectly, and I think it the best lathe made in the world for Bedstead and Chair work. I would recommend it to any one desirous of obtaining such a lathe. Yours truly, H. R. BRISTOL. WOOLEN MILL FOR SALE. With House and 3 acres of land, Seymour, Ct., (Naugatuck Valley,) 2 miles from R. R. depot. Never-failing stream. 3 ft. fall, dam and wheel in good condition. Inquire of JAS. ORMSBEE, on the premises. Portable & Stationary Steam Engines AND HOISTING ENGINES. A good article at low prices. Every machine warranted. Send for descriptive Price List. H. B. BIGELOW & CO., New Haven, Conn. PATENT BANDSAW MACHINES Of the most approved kinds, of various sizes , to saw bevel as well as square, without inclining the table, by , 452 to 456 Tenth ave., New York. Price $250, $275, $350, and $400. At present (Oct. 16), there are in operation, in this city alone, 88 of our machines. Send for circular. Manufacture, also, an improved saw-filing apparatus; price, $30. Have also on hand a large stock of best French Bandsaw Blades. FIRST & PRYIBIL L. & J. W. FEUCHTWANGER, 55 Cedar st., New York, Chemists, Manufacturers, and Importers of Specialities, Silicates, Soda and Potash, Chloride of Calcium, Peroxide of Manganese, Hydrofluoric Acid, Metallic Oxides, Steel and Glass Makers' and Potters' Articles, Publishers of Treatises on "Soluble Glass," "Gems," and "Fermented Liquors." PIMLICO BRACES, SOMETHING NEW. THIS invention is based on a strictly scientific principle, and is a valuable improvement on old style suspenders. It is simple in construction, and combines the qualities of Brace and Suspender. They are unequaled for elegance, durability and comfort. Manufactured at the Monumental Silk Works, Baltimore. JOHN M. DAVIES & CO., Sole Agents, 384 & 386 B'd'y, N. Y. Dr. J. ARMSTRONG'S (Patent) Improved Heater, Filter, Lime Extractor, and Condenser Combined, For Steam Boilers. MANUFACTURED BY Toledo, Ohio. Armstrong & Starr, Send for Circulars. Formerly Armstrong & Welsh. —An Engine, 12x36 in. cylinder, and two Boilers, 4x15 feet, in good order, will be sold cheap. FOR SALE. J. J. TAYLOR & CO., 68 Courtlandt st., New York. Trapping and Fishing. All about it. Sent free. Address "Hunter," Hinsdale, N. H. HUNTING, First Premium awarded by Am. Inst., 1870 T. H. McALLISTER, Optician, 49 Nassau st., N. Y. Rare and Beautiful Flowers —AND— CHOICE VEGETABLES Can always be obtained by Sowing The Seventeenth Annual Edition of their celebrated "SEED CATALOGUE AND GUIDE TO THE FLOWER AND KITCHEN GARDEN," is now ready for distribution. It contains Four Beautifully Colored Lithographs, and about 300 choice Engravings of favorite Flowers and Vegetables, 136 pages of closely-printed matter, and a list of Twenty-five Hundred species and varieties of Flower and Vegetable Seeds, with explicit directions for their culture, and much other useful information upon the subject of Gardening. A copy will be mailed to all applicants inclosing 25 cts. Regular customers supplied gratis. Address B. K. BLISS & SONS, Nos. 23 Park Place, and 20 Murray st., P. O. Box No. 5712. New York. Science for the Million. THE Boston Journal of Chemistry, Devoted to the Science of Home Life, the Arts, Agriculture, and Medicine. JAMES R. NICHOLS, M.D., WILLIAM A. ROLFE, A.M., } Editors. ONE DOLLAR PER YEAR. A paper which commends itself at once to Physicians, Druggists, Chemists, Teachers, Farmers, Mechanics—in short, to Professional and Practical Men of every class. The Domestic Recipes and Formulæ for Art Processes are of themselves worth many times the cost of subscription. Specimen Copies Sent Free. Address **BOSTON JOURNAL OF CHEMISTRY,
\ 150 Congress st., Boston. FOOT LATHES, And all kinds of small tools. Illustrated catalogue free. GOODNOW & WIGHTMAN, 23 Cornhill, Boston, Mass. AGENTS WANTED.—To sell Stephens' which embraces a Rule, Level, Square, Plumb, Bevel, Slope Level, T Square, etc., in one compact tool. These instruments retail at $3.50 each, and energetic salesmen can make money by selling them among mechanics. We warrant them in every particular, as the construction and graduation is faultless. Send for descriptive circular, cuts, and terms. Patent Combination Rule, STEPHENS & CO., Riverton, Conn. FELT. THE BEST, CHEAPEST and MOST DURABLE non-conductor known, for sale by the Original Manufacturer, at the BOILER FELTING WORKS, 46 Courtland st., New York. Universal Wood Worker. For Agricultural, Railroad, Car, Carriage, and Wagon Works, Planing Mill, Sash, Door and Blind, Bedstead, Cabinet and Furniture Factories. McBETH, BENTEL & MARGEDANT, Hamilton, O. , Inventor's Exchange 245 Broadway, N. Y., "Am. Agriculturist" Building. TANGIBLE INVENTIONS NEGOTIATED. No goods received unless ordered. B. F. KEMP, Proprietor. MACHINISTS' TOOLS, at greatly reduced prices. Also, some Woodworth Planers and Second-hand Tools. 97 to 113 R. R. ave., Newark, N. J. E. & R. J. GOULD, successors to Gould Machine Co. N. B. PATENTED Articles introduced. Also, State and County Rights sold for Inventors. STONE, PUGH & CO., 55 N. 6th st., Philadelphia. G OLDEN HILL Seminary for young ladies, Bridgeport, Conn. Miss Emily Nelson, Principal. 1826 USE THE VEGETABLE PULMONARY BALSAM. 1870 The old standard remedy for Coughs, Colds, Consumption. "Nothing Better." Cutler Bros. & Co., Boston. T HE CALVERT IRON ROLLING MILLS are offered at private sale. These mills are situated in the city of Baltimore, and cover 1½ acres of ground. The Machinery is of the most approved description, for making all sizes of round and square bar iron, from ¼ in. to 3 in. diameter, and flat bars of all widths, up to 7 inches. The buildings are ample and commodious. In addition to the Rolling Mills are two brick buildings (50x125 feet and 40x90 feet), now containing an 80 H.P. Engine, and Spike Machinery, but which could be used for the manufacture of Nails, Horseshoes, or any other branch of heavy hardware. This property offers an unusual opportunity to capitalists, and will be sold at a reasonable price. For further description address MARSHALL  P. SMITH, P. O. Box 1158, Baltimore, Md. BURDON IRON WORKS.—Manufacturers of Pumping Engines for Water Works, High & Low Pressure Engines, Portable Engines and Boilers, of all kinds, Sugar Mills, Screw, Lever, Drop, & Hydraulic Presses, Machinery in general. HUBBARD & WHITTAKER, 102 Front st., Brooklyn. ENGINES AND MACHINERY FOR SALE, at a great sacrifice. Two new Steam Engines, 12 and 20 horse power; 1 Faribain's Riveting Machine; 1 large Power Shears; 1 ditto Table Punch; 2 ditto Flange Punches; 1 set Power Bending Rolls; together with a large lot of Turning Lathes, Drilling Machines, Machinists' and Smiths' Hand Tools, Pulleys, Hangers, and 6 Fairbanks' Platform Scales. Send for catalogue, or apply at the South Brooklyn Steam Engine Works, cor. Imlay and Summit sts., Brooklyn. SHINGLE AND HEADING MACHINE— Law's Patent with Trevor & Co.'s Improvements. The Simplest and Best in use. Also, Shingle, Heading and Stave Jointers, Equalizers, Heading Turners, Planers etc. AddressTREVOR & CO., Lockport, N. Y. AGENTS WANTED—($225 A MONTH) by the AMERICAN KNITTING MACHINE CO. Boston, Mass., or St. Louis, Mo. THE UNITED STATES Brick Machine IS THE BEST IN THE WORLD, BECAUSE IT MAKES THE , , and GREATEST NUMBER THE BEST THE CHEAPEST BRICKS. It is the Perfection of Simplicity. It is Durable, and not likely to get out of Repair. See Scientific American, Sept. 17, 1870. For Descriptive Circular apply to F. C. WELLS, President, Room 13, 98 Madison st.,\CHICAGO. MACHINES can be seen in operation at the Company's Works, Chicago; at rear 59 Ann st., New York city; and at Novelty Iron Works, corner of Delord and Peter sts., New Orleans. PREPARED Asphalte Roofing Felt. This new prepared production is ready coated, and can be applied on the roof without further trouble. It is easy of application, and does not require any repairs for a long time. It is more durable than some slates, and has been found a suitable substitute for iron or tin roofs. It has a sanded or stony surface, which renders it Exposed to the most intense fire, and sparks falling upon it, it will not propagate the fire. Under the influence of the sun it will not run, which makes it specially adapted to hot climates. Its easy application and pleasing appearance have made it a favorite roofing material throughout all the Indies and other colonies. Being not cumbrous for transport, it is of invaluable service to settlers and farmers in far remote districts. When used for temporary purposes it may be taken off and applied again to another construction. It replaces common Asphalting on Terraces, Lobbies, Counting-houses, Office Floors, etc.; is a great preservative against dampness and vermin, and equalizes the temperature. It is 32 inches wide, and made in rolls of 25 yards each. Send for circular to UNINFLAMMABLE and FIRE-PROOF. E. H. MARTIN, 70 Maiden Lane and 9 Liberty St., N. Y. J. J. H. GREGORY'S Seed Catalogue. My Annual Illustrated Catalogue, containing a list of many new and rare Vegetables, some of which are not found in any other catalogue, and all the standard vegetables of the farm and garden (over one hundred of which I grow on my three seed farms), with a carefully selected list of flower seed, will be sent free to all. All my seed is sold under three warrants: 1st. That all money sent shall reach me. 2d. That all seed ordered shall reach the purchaser. 3d. That my seeds shall be fresh and true to name. JAMES J. H. GREGORY, Marblehead, Mass. THE NEW WILSON Under-Feed Shuttle\SEWING MACHINES! $25 cheaper than any other! For Simplicity, Durability and Beauty they stand For Stitching, Hemming, Tucking, Felling, Quilting, Cording, Binding, Braiding, Gathering, Gathering & sewing on gathers, For particulars address unrivalled! they are unexcelled! Wilson Sewing Machine Co., Cleveland, O., or St. Louis, Mo. , **NEW and 2d-HAND.—
\ Send for Circular. Chas. PLACE & CO., 60 Vesey st., New York. MACHINERY MACHINISTS. Illustrated Catalogue and Price List of all kinds of small Tools and Materials sent free to any address. GOODNOW & WIGHTMAN, 23 Cornhill, Boston, Mass. P. BLAISDELL & CO. MANUFACTURERS of the "BLAISDELL" PATENT DRILL PRESSES, with quick return motion, Agricultural Drills, Improved Engine Lathes, from 12 in. to 28 in. swing, Planers, Gear Cutters, Boring Mills, Hand Lathes, and other first-class Machinists' Tools. Jackson st., Worcester, Mass. Patent BEDSTEAD FASTENING. The and ever invented. Best, Cheapest, Strongest FASTENING Rights for States and Territories for sale. Address JOHN DOMINGOS and                     BENJAMIN ESSIG, Sacramento, Cal. BENT, GOODNOW & CO., Boston, Mass., Publishers of , sell Patent Rights and goods of all kinds. Orders solicited. "PATENT STAR" AGENTS WANTED. Send stamp for copy. TO THE WORKING CLASS.—We are now prepared to furnish all classes with constant employment at home, the whole of the time or for the spare moments. Business new, light and profitable. Persons of either sex easily earn from 50c. to $5 per evening, and a proportional sum by devoting their whole time to the business. Boys and girls earn nearly as much as men. That all who see this notice may send their address, and test the business, we make this unparalleled offer: To such as are not well satisfied, we will send $1 to pay for the trouble of writing. Full particulars, a valuable sample which will do to commence work on, and a copy of —one of the largest and best family newspapers published—all sent free by mail. Reader, if you want permanent, profitable work, address The People's Literary Companion D. C. ALLEN & CO., Augusta, Maine. IMPORTANT TO  MACHINISTS.—The Best Metal for all Machine Uses is the MARTIN STEEL, made by The New Jersey Steel and Iron Co., Trenton, N. J. This steel is made by an entirely different process from any other and is tougher than wrought iron. It can be turned without annealing, being entirely free from hard spots. Every one who uses it pronounces it just what they have long wanted, for a multitude of uses, such as Crank Pins, Lathe Spindles and Screws, Cotton Machinery Rollers, Saw and Fan Spindles, etc., etc. Also, particularly adapted for Firebox Plates. Prices low. Send for further information, or a sample, stating use to which it is to be applied. OTIS' SAFETY HOISTING\Machinery. **OTIS, BROS. & CO.
\ No. 309 BROADWAY, NEW YORK. Union Emery Wheels. Solid and with Stone Center. UNION STONE CO., Boston, Mass. Branch Office, 93 Liberty st., N. Y. General Agents for the Am. Twist Drill Co.'s Superior Grinder and other Emery Wheel Machinery and Tools. Send for Circular. \ WOODBURY'S PATENT** Planing and Matching and Molding Machines, Gray & Wood's Planers, Self-oiling Saw Arbors, and other wood working machinery. S. A. WOODS, Send for Circulars. } 91 Liberty street, N. Y.; 67 Sudbury street, Boston. RI**CHARDSON, MERIAM & CO.,
\ Manufacturers of the latest improved Patent Daniels' and Woodworth Planing Machines, Matching, Sash, and molding, Tenoning, Mortising, Boring, Shaping, Vertical, and Circular Re-sawing Machines, Saw Mills Saw Arbors, Scroll Saws, Railway, Cut-off, and Rip-saw Machines, Spoke and Wood Turning Lathes, and various other kinds of Wood-working Machinery. Catalogues and price lists sent on application. Manufactory, Worcester, Mass. Warehouse, 107 Liberty st., New York. 17 1 Reynolds' TURBINE WATER WHEELS. The Oldest and Newest. All others only imitations of each other in their strife after complications to confuse the public. We do not boast but quietly excel them all in staunch reliable, economical power. Beautiful pamphlet free. Geo. TALLCOT, 96 Liberty st., New York. Gearing, Shafting.

To Manufacturers of Knitted Goods

152 Stories To Learn About Ethereum Blockchain

Steam, Its Generation and Use by Babcock & Wilcox Company, is part of the HackerNoon Books Series. You can jump to any chapter in this book . BOILER ROOM PIPING here BOILER ROOM PIPING In the design of a steam plant, the piping system should receive the most careful consideration. Aside from the constructive details, good practice in which is fairly well established, the important factors are the size of the piping to be employed and the methods utilized in avoiding difficulties from the presence in the system of water of condensation and the means employed toward reducing radiation losses. Engineering opinion varies considerably on the question of material of pipes and fittings for different classes of work, and the following is offered simply as a suggestion of what constitutes good representative practice. All pipe should be of wrought iron or soft steel. Pipe at present is made in “standard”, “extra strong” and “double extra strong” weights. Until recently, a fourth weight approximately 10 per cent lighter than standard and known as “Merchants” was built but the use of this pipe has largely gone out of practice. Pipe sizes, unless otherwise stated, are given in terms of nominal internal diameter. gives the dimensions and some general data on standard and extra strong wrought-iron pipe. [76] Table 62 Dimensions are nominal and except where noted are in inches. In connection with pipe sizes, , giving certain tube data may be found to be of service. Table 63 Dimensions are nominal and except where noted are in inches. Pipe Material and Thickness—For saturated steam pressures not exceeding 160 pounds, all pipe over 14 inches should be 3⁄8 inch thick O. D. pipe. All other pipe should be standard full weight, except high pressure feed and blow-off lines, which should be extra strong. [77] For pressures above 150 pounds up to 200 pounds with superheated steam, all high pressure feed and blow-off lines, high pressure steam lines having threaded flanges, and straight runs and bends of high pressure steam lines 6 inches and under having Van Stone joints should be extra strong. All piping 7 inches and over having Van Stone joints should be full weight soft flanging pipe of special quality. Pipe 14 inches and over should be 3⁄8 inch thick O. D. pipe. All pipes for these pressures not specified above should be full weight pipe. Flanges—For saturated steam, 160 pounds working pressure, all flanges for wrought-iron pipe should be cast-iron threaded. All high pressure threaded flanges should have the diameter thickness and drilling in accordance with the “manufacturer’s standard” for “extra heavy” flanges. All low pressure flanges should have diameter, thickness and drilling in accordance with “manufacturer’s standard” for “standard flanges.” The flanges on high pressure lines should be counterbored to receive pipe and prevent the threads from shouldering. The pipe should be screwed through the flange at least 1⁄16 inch, placed in machine and after facing off the end one smooth cut should be taken over the face of the flange to make it square with the axis of the pipe. For pressures above 160 pounds, where superheated steam is used, all high pressure steam lines 4 inches and over should have solid rolled steel flanges and special upset lapped joints. In the manufacture of such joints, the ends of the pipe are heated and upset against the face of a holding mandrel conforming to the shape of the flange, the lapped portion of the pipe being flattened out against the face of the mandrel, the upsetting action maintaining the desired thickness of the lap. When cool, both sides of the lap are faced to form a uniform thickness and an even bearing against flange and gasket. The joint, therefore, is a strictly metal to metal joint, the flanges merely holding the lapped ends of the pipe against the gasket. A special grade of soft flanging pipe is selected to prevent breaking. The bending action is a severe test of the pipe and if it withstands the bending process and the pressure tests, the reliability of the joint is assured. Such a joint is called a Van Stone joint, though many modifications and improvements have been made since the joint was originally introduced. The diameter and thickness of such flanges should be special extra heavy. Such flanges should be turned to diameter, their fronts faced and the backs machined in lieu of spot facing. In lines other than given for pressures over 150 pounds, all flanges for wrought-iron pipe should be threaded. All threaded flanges for high pressure superheated lines 3½ inches and under should be “semi-steel” extra heavy. Flanges for other than steam lines should be manufacturer’s standard extra heavy. Welded flanges are frequently used in place of those described with satisfactory results. Fittings—For saturated steam under pressures up to 160 pounds, all fittings 3½ inches and under should be screwed. Fittings 4 inches and over should have flanged ends. Fittings for this pressure should be of cast iron and should have heavy leads and full taper threads. Flanged fittings in high pressure lines should be extra heavy, and in low pressure lines standard weight. Where possible in high pressure flanges and fittings, bolt surfaces should be spot faced to provide suitable bearing for bolt heads and nuts. Fittings for superheated steam up to 70 degrees at pressures above 160 pounds are sometimes of cast iron. For superheat above 70 degrees such fittings should be “steel castings” and in general these fittings are recommended for any degree of superheat. Fittings for other than high pressure work may be of cast iron, except where superheated steam is carried, where they should be of “wrought steel” or “hard metal”. Fittings 3½ inches and under should be screwed, 4 inches and over flanged. [78] Flanges for pressures up to 160 pounds in pipes and fittings for low pressure lines, and any fittings for high pressure lines should have plain faces, smooth tool finish, scored with V-shaped grooves for rubber gaskets. High pressure line flanges should have raised faces, projecting the full available diameter inside the bolt holes. These faces should be similarly scored. All pipe ½ inch and under should have ground joint unions suitable for the pressure required. Pipe ¾ inch and over should have cast-iron flanged unions. Unions are to be preferred to wrought-iron couplings wherever possible to facilitate dismantling. Valves—For 150 pounds working pressure, saturated steam, all valves 2 inches and under may have screwed ends; 2½ inches and over should be flanged. All high pressure steam valves 6 inches and over should have suitable by-passes. All valves for use with superheated steam should be of special construction. For pressures above 160 pounds, where the superheat does not exceed 70 degrees, valve bodies, caps and yokes are sometimes made of cast iron, though ordinarily semi-steel will give better satisfaction. The spindles of such valves should be of bronze and there should be special necks with condensing chambers to prevent the superheated steam from blowing through the packing. For pressures over 160 pounds and degrees of superheat above 70, all valves 3 inches and over should have valve bodies, caps and yokes of steel castings. Spindles should be of some non-corrosive metal, such as “monel metal”. Seat rings should be removable of the same non-corrosive metal as should the spindle seats and plug faces. All salt water valves should have bronze spindles, sleeves and packing seats. The suggestions as to flanges for different classes of service made on page hold as well for valve flanges, except that such flanges are not scored. 311 Automatic stop and check valves are coming into general use with boilers and such use is compulsory under the boiler regulations of certain communities. Where used, they should be preferably placed directly on the boiler nozzle. Where two or more boilers are on one line, in addition to the valve at the boiler, whether this be an automatic valve or a gate valve, there should be an additional gate valve on each boiler branch at the main steam header. Relief valves should be furnished at the discharge side of each feed pump and on the discharge side of each feed heater of the closed type. Feed Lines—Feed lines should in all instances be made of extra strong pipe due to the corrosive action of hot feed water. While it has been suggested above that cast-iron threaded flanges should be used in such lines, due to the sudden expansion of such pipe in certain instances cast-iron threaded flanges crack before they become thoroughly heated and expand, and for this reason cast-steel threaded flanges will give more satisfactory results. In some instances, wrought-steel and Van Stone joints have been used in feed lines and this undoubtedly is better practice than the use of cast-steel threaded work, though the additional cost is not warranted in all stations. Feed valves should always be of the globe pattern. A gate valve cannot be closely regulated and often clatters owing to the pulsations of the feed pump. Gaskets—For steam and water lines where the pressure does not exceed 160 pounds, wire insertion rubber gaskets 1⁄16 inch thick will be found to give good service. For low pressure lines, canvas insertion black rubber gaskets are ordinarily used. For oil lines special gaskets are necessary. For pressure above 160 pounds carrying superheated steam, corrugated steel gaskets extending the full available diameter inside of the bolt holes give good satisfaction. For high pressure water lines wire inserted rubber gaskets are used, and for low pressure flanged joints canvas inserted rubber gaskets. Size of Steam Lines—The factors affecting the proper size of steam lines are the radiation from such lines and the velocity of steam within them. As the size of the steam line increases, there will be an increase in the radiation. As the size decreases, the steam velocity and the pressure drop for a given quantity of steam naturally increases. [79] There is a marked tendency in modern practice toward higher steam velocities, particularly in the case of superheated steam. It was formerly considered good practice to limit this velocity to 6000 feet per minute but this figure is to-day considered low. In practice the limiting factor in the velocity advisable is the allowable pressure drop. In the description of the action of the throttling calorimeter, it has been demonstrated that there is no loss accompanying a drop in pressure, the difference in energy between the higher and lower pressures appearing as heat, which, in the case of steam flowing through a pipe, may evaporate any condensation present or may be radiated from the pipe. A decrease in pipe area decreases the radiating surface of the pipe and thus the possible condensation. As the heat liberated by the pressure drop is utilized in overcoming or diminishing the tendency toward condensation and the heat loss through radiation, the steam as it enters the prime mover will be drier or more highly superheated where high steam velocities are used than where they are lower, and if enough excess pressure is carried at the boilers to maintain the desired pressure at the prime mover, the pressure drop results in an actual saving rather than a loss. The whole is analogous to standard practice in electrical distributing systems where generator voltage is adjusted to suit the loss in the feeder lines. In modern practice, with superheated steam, velocities of 15,000 feet per minute are not unusual and this figure is very frequently exceeded. Piping System Design—With the proper size of pipe to be used determined, the most important factor is the provision for the removal of water of condensation that will occur in any system. Such condensation cannot be wholly overcome and if the water of condensation is carried to the prime mover, difficulties will invariably result. Water is practically incompressible and its effect when traveling at high velocities differs little from that of a solid body of equal weight, hence impact against elbows, valves or other obstructions, is the equivalent of a heavy hammer blow that may result in the fracture of the pipe. If there is not sufficient water in the system to produce this result, it will certainly cause knocking and vibration in the pipe, resulting eventually in leaky joints. Where the water reaches the prime mover, its effect will vary from disagreeable knocking to disruption. Too frequently when there are disastrous results from such a cause the boilers are blamed for delivering wet steam when, as a matter of fact, the evil is purely a result of poor piping design, the most common cause of such an action being the pocketing of the water in certain parts of the piping from whence it is carried along in slugs by the steam. The action is particularly severe if steam is admitted to a cold pipe containing water, as the water may then form a partial vacuum by condensing the steam and be projected at a very high velocity through the pipes producing a characteristic sharp metallic knock which often causes bursting of the pipe or fittings. The amount of water present through condensation may be appreciated when it is considered that uncovered 6-inch pipe 150 feet long carrying 3600 pounds of high pressure steam per hour will condense approximately 6 per cent of the total steam carried through radiation. It follows that efficient means of removing condensation water are absolutely imperative and the following suggestions as to such means may be of service: The pitch of all pipe should be in the direction of the flow of steam. Wherever a rise is necessary, a drain should be installed. All main headers and important branches should end in a drop leg and each such drop leg and any low points in the system should be connected to the drainage pump. A similar connection should be made to every fitting where there is danger of a water pocket. Branch lines should never be taken from the bottom of a main header but where possible should be taken from the top. Each engine supply pipe should have its own separator placed as near the throttle as possible. Such separators should be drained to the drainage system. Check valves are frequently placed in drain pipes to prevent steam from entering any portion of the system that may be shut off. Valves should be so located that they cannot form water pockets when either open or closed. Globe valves will form a water pocket in the piping to which they are connected unless set with the stem horizontal, while gate valves may be set with the spindle vertical or at an angle. Where valves are placed directly on the boiler nozzle, a drain should be provided above them. High pressure drains should be trapped to both feed heaters and waste headers. Traps and meters should be provided with by-passes. Cylinder drains, heater blow-offs and drains, boiler blow-offs and similar lines should be led to waste. The ends of cylinder drains should not extend below the surface of water, for on starting up or on closing the throttle valve with the drains open, water may be drawn back into the cylinders. Covering—Magnesia, canvas covered. For calculating radiation for pressure and temperature other than 160 pounds, and 60 degrees, use B. t. u. figures for one degree difference. Radiation from Pipes—The evils of the presence of condensed steam in piping systems have been thoroughly discussed above and in some of the previous articles. Condensation resulting from radiation, while it cannot be wholly obviated, can, by proper installation, be greatly reduced. Bare pipe will radiate approximately 3 B. t. u. per hour per square foot of exposed surface per one degree of difference in temperature between the steam contained and the external air. This figure may be reduced to from 0.3 to 0.4 B. t. u. for the same conditions by a 1½ inch insulating covering. gives the radiation losses for bare and covered pipes with different thicknesses of magnesia covering. Table 64 Many experiments have been made as to the relative efficiencies of different kinds of covering. gives some approximately relative figures based on one inch covering from experiments by Paulding, Jacobus, Brill and others. Table 65 Based on one-inch covering. The following suggestions may be of service: Exposed radiating surfaces of all pipes, all high pressure steam flanges, valve bodies and fittings, heaters and separators, should be covered with non-conducting material wherever such covering will improve plant economy. All main steam lines, engine and boiler branches, should be covered with 2 inches of 85 per cent carbonate of magnesia or the equivalent. Other lines may be covered with one inch of the same material. All covering should be sectional in form and large surfaces should be covered with blocks, except where such material would be difficult to install, in which case plastic material should be used. In the case of flanges the covering should be tapered back from the flange in order that the bolts may be removed. All surfaces should be painted before the covering is applied. Canvas is ordinarily placed over the covering, held in place by wrought-iron or brass bands. Expansion and Support of Pipe—It is highly important that the piping be so run that there will be no undue strains through the action of expansion. Certain points are usually securely anchored and the expansion of the piping at other points taken care of by providing supports along which the piping will slide or by means of flexible hangers. Where pipe is supported or anchored, it should be from the building structure and not from boilers or prime movers. Where supports are furnished, they should in general be of any of the numerous sliding supports that are available. Expansion is taken care of by such a method of support and by the providing of large radius bends where necessary. It was formerly believed that piping would actually expand under steam temperatures about one-half the theoretical amount due to the fact that the exterior of the pipe would not reach the full temperature of the steam contained. It would appear, however from recent experiments that such actual expansion will in the case of well-covered pipe be very nearly the theoretical amount. In one case noted, a steam header 293 feet long when heated under a working pressure of 190 pounds, the steam superheated approximately 125 degrees, expanded 8¾ inches; the theoretical amount of expansion under the conditions would be approximately 935⁄64 inches. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Babcock & Wilcox Company (2007). Steam, Its Generation and Use. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/22657/pg22657-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

BOILER ROOM PIPING

How to Write Clearly: Rules and Exercises on English Composition by Edwin Abbott Abbott, is part of the HackerNoon Books Series. You can jump to any chapter in this book . BREVITY. here BREVITY. *46. Metaphor is briefer than literal statement.* See (13). "The cares and responsibilities of a sovereign often disturb his sleep," is not so brief as "Uneasy lies the head that wears a crown," where the effect of care on the mind is assimilated to the effect of a heavy crown pressing on the head. *47. General terms are briefer, though less forcible, than particular terms.* Thus: "He devours , no matter of what kind," is shorter than, "Novels or sermons, poems or histories, no matter what, he devours them all." literature *47 a. A phrase may be expressed by a word.* "These impressions , i.e. are ." can never be forgotten indelible "The style of this book is , i.e. ." of such a nature that it cannot be understood unintelligible The words "of such a nature that" are often unnecessarily inserted. See the extract from Sir Archibald Alison. *48. Participles can often be used as brief (though sometimes ambiguous) equivalents of phrases containing Conjunctions and Verbs.* "Hearing (when he heard) this, he advanced." See (7) for more instances. So "phrases conjunctions" means "phrases conjunctions." " , (for, ) he retired." containing that contain This done when this was done Sometimes the participle "being" is omitted. "France at our doors, he sees no danger nigh," for "France being" or "though France is." *49. Participles and participial adjectives may be used like Adjectives, as equivalents for phrases containing the Relative.* "The never- wind," "the ocean," "the rain," are instances. The licence of inventing participial adjectives by adding to a noun, is almost restricted to poetry. You could not write "the wind" in prose. ceasing clamouring drenching -ing crannying *50. A statement may sometimes be briefly implied instead of being expressed at length.* Thus, instead of "The spirit of Christianity was humanizing, and therefore &c.," or "Christianity, since it was (or being) of a humanizing spirit, discouraged &c.," we can write more briefly and effectively, "Gladiatorial shows were first discouraged, and finally put down, by the ." So instead of "The nature of youth is thoughtless and sanguine, and therefore &c.," we can write, "The danger of the voyage was depreciated and the beauty of the island exaggerated by ." humanizing spirit of Christianity the thoughtless nature of youth Sometimes a mere name or epithet implies a statement. "It was in vain that he offered the Swiss terms: war was deliberately preferred by the ," "by the Swiss, ." "The deed was applauded by all honest men, but the Government affected to treat it as murder, and set a price upon the head of (him whom they called) the might be expected to hold different language from ," "Napoleon when elated by the victory of Austerlitz," and "Napoleon when depressed by his imprisonment at St. Helena." hardy mountaineers i.e. because they were mountaineers and hardy assassin." "The conqueror of Austerlitz the prisoner of St. Helena i.e. CAUTION.—Different names must not be used for the same person unless each of them derives an appropriateness from its context. Thus, if we are writing about Charles II., it would be in very bad taste to avoid repeating "he" by using such periphrases as the following: "The third of the Stewarts hated business," "the Merry Monarch died in the fifty-fourth year of his age," &c. *51. Conjunctions may be omitted.* The omission gives a certain forcible abruptness, "You say this: I (on the other hand) deny it." e.g. When sentences are short, as in Macaulay's writings, conjunctions may be advantageously omitted. Where a contrast is intended, the conjunction usually prepares the way for the second of the two contrasted terms: "He is good dull." Where is used instead of , the incongruity savours of epigram: "He always talks truthfully prosily." "He is always amusing false." but but and but and and *51 a. The Imperative Mood may be used for "if."* " (for, ) Virtue of the awful authority she derives from the general reverence of mankind, and you rob her of half her majesty." Strip if you strip *52. Apposition may be used so as to convert two sentences into one.* "We called at the house of a person to whom we had letters of introduction, , and, what is more, a to all young students of music." This is as clear as, and briefer than, "He was a musician, &c." a musician good friend *53. Condensation may be effected by not repeating (1) the common subject of several verbs, (2) the common object of several verbs or prepositions.* (1) "He resided here for many years, and, after he had won the esteem of all the citizens, ( ) died," &c. So, (2) "He came to, and was induced to reside in, this city," is shorter than "He came to this city, and was induced to reside in it." he Such condensation often causes obscurity, and, even where there is no obscurity, there is a certain harshness in pausing on light, unemphatic words, such as , , &c., as in the first example. to in *54. Tautology.*—The fault of repeating the same word several times unnecessarily is called , e.g.: tautology "This is a painful ; it is a that I much , and he also will much the ." But the fault is not to be avoided by using different words to mean the same thing, as, "This is a painful ; it is a that I , and he also will the ." The true remedy is to arrange the words in such a manner that there may be no unnecessary repetition, thus: "This is a painful circumstance, a circumstance that causes me, and will cause him, deep regret." circumstance circumstance regret regret circumstance event circumstance much regret greatly lament occurrence The repetition of the same meaning in slightly different words is a worse fault than the repetition of the same word. See, for examples, the extract from Sir Archibald Alison, at the end of the book. Thus " for conquests is a characteristic of this nation. It is an that &c." Other instances are—"The opinion of men;" "His judgment is so that it is ," &c. A burning thirst ardent passion universal all infallible never deceived *55. Parenthesis may be used with advantage to brevity.* "We are all (and who would not be?) offended at the treatment we have received," is shorter and more forcible than the sentence would have been if the parenthesis had been appended in a separate sentence: "Who, indeed, would not be offended?" Extreme care must, however, be taken that a parenthesis may not obscure the meaning of a long sentence. *56. Caution: let clearness be the first consideration.* It is best, at all events for beginners, not to aim so much at being brief, or forcible, as at being perfectly clear. Horace says, "While I take pains to be brief, I fall into obscurity," and it may easily be seen that several of the rules for brevity interfere with the rules for clearness. Forcible style springs from (1) vividness and (2) exactness of thought, and from a corresponding (1) vividness and (2) exactness in the use of words. (1) When you are describing anything, endeavour to it and describe it as you see it. If you are writing about a man who was killed, the man before you, and ask, was he , , , , , or ? If you are writing about the capture of a city, was the city , , , , or ? Was an army , , , , or ? see see executed cut down run through the body butchered shot hanged stormed surprised surrendered starved out demolished before surrender repelled defeated routed crushed annihilated (2) Exactness in the use of words requires an exact knowledge of their meanings and differences. This is a study by itself, and cannot be discussed here.[17] FOOTNOTES: [17] See , pp. 1-53. English Lessons for English People About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Edwin Abbott Abbott (2007). How to Write Clearly: Rules and Exercises on English Composition. Urbana, Illinois: Project Gutenberg. Retrieved https://www.gutenberg.org/cache/epub/22600/pg22600.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at , located at . www.gutenberg.org https://www.gutenberg.org/policy/license.html

Stories, News & Mentions about HANGEH

Hangeh’s Stories on HackerNoon