Abstract is an all-in-one workspace for government affairs in California

Abstract

Introduction Last chapter of this series. If you have followed up along these blog posts you are now fully capable of creating projects using the Object Oriented paradigm. I am glad you have followed all the sections and in this part, you are going to end up what you have started. In fact, now you will learn how to use abstract classes and several other less common feature of PHP such as reflection API and the singleton design pattern. Keep reading and you will see that as per a lot of other web developers, myself included, these resources will let you do that breakthrough you are looking for. Follow the series … This blog post is the sixth part of “ The complete guide to Object-Oriented Programming: Go from procedural programming to OOP expert in ” PHP If you did not read the other parts yet You can check the other blog posts following the links below , , , , and Introduction to Object-Oriented Programming Inheritance and Interfaces in PHP More Interfaces and Polymorphism Visibility and Static keyword Constructor and Magic methods Abstract Classes and Extra bits Table of Content Abstract classes Method chaining Autoloading Anonymous classes Reflection API Singleton Abstract Classes Abstract classes are very similar to interfaces but they have different rules that a web developer need to be aware of. The definition is quite straightforward, basically, an abstract class is a class that contains abstract methods, Abstract methods are methods that have been declared but not implemented in the code. This kind of classes cannot be constructed as an object, in fact, they need to be extended. A ‘normal’ class that extends an abstract one must implement all methods that are inside the parent. Something to notice is that a child class must define the abstract methods with the same or less restrictive visibility, This means that if in the abstract class there is a protected method it can be implemented only as a public or protected. Another rule is that the number and type or required arguments need to be the same as the parameters. abstract

    abstract

    public

    {

        echo

    }

}

class

    public

    {

        return

    }

}

$myOffice

// This command will return “Building cannot be constructed”

$myOffice->getWindowsCount();

// This command will return 12 I have defined the Office class that extends the abstract class . Building I have implemented and changed its visibility from protected to public. getWindowsCount() There are no abstract methods within the class thus, I can construct an object from it. Office I have added a constructor inside the abstract class so that you will be aware of the fact the parent’s constructor is called when the object is instantiated. What is the difference between an interface and an abstract class? To sum up, Consider an interface as an empty shell. It has only signatures of the methods, which means the methods do not have a body. An interface does not do anything is just a pattern, a set of methods that you need to replicate. On the other hands, abstract classes are proper classes. Even though they look like interfaces the have some specific characteristics. You can specify their use by what they do by a given method name. Here is an example that highlights the differences between the two: interface

{

     public

     public

}

class

{

    public

    public

    {

        echo

    }

    public

    {

        return

    }

} In here we are saying all the Building should look like this than is up to the class to change the functionality. Office abstract class Building{public $windows;public function getWindowsCount(){return $this->windows;}abstract public function build();} class Office extends Building{public function build(){echo “Office built”;}} In an abstract class instead, we are saying that all the Building must have . getWindowsCount() Then we say that the method can be different, so we obligate the class to provide the implementation according to its needs. build() Office Extra / Bonus We have almost finished here with this introduction to Object Oriented in PHP, programming Lot’s of information has been shared and I hope you understood all of it (if not just write your question below). However, there are a few things that have been just quickly mentioned and I would love to go deeper and give you a more wide comprehension on this topic. Method chaining Method chaining is an OOP technique that permits the programmer to perform several different but connected actions in a sequential way. Let’s break down this sentence,“Perform different actions“ means calling different methods or functions; “ ” can be translated into actions related to the same object or instance of the class; Perform connected actions “ ” simply suggest that the methods are called one after another. In a sequential way Here is an example that will clarify the concept: class

{

    public

    function

    {

        $this->message = "Foundation poured - ";

    }

    function

    {

        $this->message .= "framing, plumbing, electrical and HVAC completed -";

        return

    }

    function

    {

        $this->message .= "interior and exterior completed";

        return

    }

    function

    {

        return

    }

}

$myHome

echo Here you go, All the method of the class have one trait in common all of them return the object itself via the use of the keyword , Building $this If you need to brush up your memory I have explained what the keyword does in the . $this introduction to Object-Oriented Programming The last command is chaining down all the methods of the class necessary in order to build up our sentence. The outcome of the last command will look like this: “Foundation poured — framing, plumbing, electrical and HVAC completed — interior and exterior completed”. This is so easy that you can implement this to your code straight away without any problem, just be careful to return the object. Autoloading Any class must be defined before you need to use it, forget to do so will result in an error thrown. Autoload alongside PSR-4 standards allows you to define them when a class is required. I am sure that I do not need to say that this is an incredibly important feature. In PHP autoloading is done by the that accept a function as a parameter, spl_autoload_register() This is quite out-topic and just to explain how it works in detail I will need to write another article, so I will just give you an understandable example: Given that the class you want to autoload is within the “classes/” path, here is what you would do. This is more or less what happens behind the scene of most of the PHP frameworks out there. function

    include

}

spl_autoload_register(myAutoloader); spl_autoload_register has two more parameters, the second is a boolean and defines if the need to throw exceptions when the autoload_function cannot be registered, spl_autoload_register the third parameter is boolean as well and if true it will prepend the autoloader rather then attaching it. If an error occurred and the function was not able to find the specified class, it will throw a fatal error. Anonymous classes Wouldn’t be great if PHP allowed the web developer to create and instantiate objects on the fly? The release of PHP 7 included this present to the list of its features. $myAnonymousBuilding

{

    public

    {

        echo

    }

} Cool, right? Notice how easy is to implement this class inline and that it can even accept parameters like a normal class. The code in the example will output the string “a new class has been constructed in Main Street”. The main use of an anonymous class is the case you need to extend a named class. If, for instance, you will need to override properties or methods, instead of declaring it you can just create an in-line implementation in the same workflow you are currently having. Reflection API Reflection is a very useful feature that has been introduced with the release of PHP 5.0 and became a standard later on 5.3. This API permits you to examine PHP elements and fetch information about them at runtime. For me, one of the most useful places to use reflection API is during automated testing, in specific during PHP Unit. For example, when testing the values of private property in classes you can entree the properties via the API and then make assertion according to the test. There are different reflection classes, each of them allows you to inspect different types of variables. Here are some classes: ReflectionClass ReflectionObject ReflectionMethod ReflectionFunction ReflectionProperty Below a quick example of how the ReflectionClass is used: $ReflectionObject

print_r($ReflectionObject->getMethods()); This example will output an array of all the methods available within the Building class. The parameter passed to the ReflectionClass can be of two types, either a string with the name of the class or a true instance of the class itself. Each of the reflection classes has few methods that allow you to retrieve different information about the element you are inspecting. Singleton Any self-respecting web developer will encounter design patterns in its way to master the programming art sooner or later. Those patterns are nothing more than established methods that solve problems in a predefined way. 99% of PHP framework use them in order to increase the performances and the quality of the code. If you have never heard the words design pattern here is your shot. I want to introduce you the easiest among them all. We are going to use again our imaginary city, we have had several etc, Buildings, Offices, Schools What about the city hall? the mayor needs to keep on eye on all these classes and instantiation and method and properties. In a city, there is only one mayor and he works in only one building the city hall. There cannot be more that one city hall in a city. Well… How do we add this rule in programming? Singleton class cityHall{private static $instance = null;private function __construct(){...}public static function getInstance(){if (self::$instance == null){self::$instance = new Singleton();}return self::$instance;}} Here is it, this is an implementation of the Singleton design pattern. How does it actually work? As you can see the visibility of the constructor is set to private it blocks the creation of the object. In order to create an instance of the class, you need to invoke the static method . getInstance() It checks if the instance has been already created if not, it creates a new instance with the command self::$instance = new Singleton(); Otherwise, it just returns the object already created. The result is that in this case, you are always going to have only one single cityHall in all your program. Where to use the Singleton design pattern in a real case? Singleton is very easy but it can also become problematic to deal with, think at this design pattern whenever you must have a single instance of a class. and it has to be one in a really long term. An easy implementation could be when creating a database class to connect a DB to the code, in the majority of the cases you are going to have only one database per project. This is the perfect use for it. Conclusion The full version of “The complete guide to Object-Oriented Programming” in an all-in-one edition is available in a Kindle format on Amazon. If you have followed all the part of this article this has been quite a journey. You have realized the Object Oriented Programming or OOP is less difficult than it may seem from an outsider. The reality is that as web developers we love to use jargon. Properties, paramethers, interfaces, abstract classes, dependency injection. Those words can provide a headache very quickly. However, just keeping on going taking a little bit at a time you will find that apply these new knowledge is not only possible but easier than you thought. What’s the next step? If you understood the basic concept of this article you are ready to learn about PHP frameworks, There are plenty of them out there, and choose one of them can be overwhelming. Below there is a really deep comparison that I wrote that compare 24 of them, you will not find anything similar elsewhere. Guide to PHP frameworks. If you like this content and you are hungry for some more join the Facebook’s community in which we share info and news just like this one!

Abstract Classes and OOP extras

This paper is available on arxiv under CC0 1.0 DEED license. Authors: (1) Salim Chouaki, LIX, CNRS, Inria, Ecole Polytechnique, Institut Polytechnique de Paris; (2) Oana Goga, LIX, CNRS, Inria, Ecole Polytechnique, Institut Polytechnique de Paris; (3) Hamed Haddadi, Imperial College London, Brave Software; (4) Peter Snyder, Brave Software. Table of Links Abstract and Introduction Background Measurement Methodology Results Limitations Related Work Conclusion, Acknowledgments, and References Appendix ABSTRACT We present the first extensive measurement of the privacy properties of the advertising systems used by privacy-focused search engines. We propose an automated methodology to study the impact of clicking on search ads on three popular private search engines which have advertising based business models: StartPage, Qwant, and DuckDuckGo, and we compare them to two dominant data-harvesting ones: Google and Bing. We investigate the possibility of third parties tracking users when clicking on ads by analyzing firstparty storage, redirection domain paths, and requests sent before, when, and after the clicks. Our results show that privacy-focused search engines fail to protect users’ privacy when clicking ads. Users’ requests are sent through redirectors on 4% of ad clicks on Bing, 86% of ad clicks on Qwant, and 100% of ad clicks on Google, DuckDuckGo, and StartPage. Even worse, advertising systems collude with advertisers across all search engines by passing unique IDs to advertisers in most ad clicks. These IDs allow redirectors to aggregate users’ activity on ads’ destination websites in addition to the activity they record when users are redirected through them. Overall, we observe that both privacy-focused and traditional search engines engage in privacy-harming behaviors allowing cross-site tracking, even in privacy-enhanced browsers. CCS CONCEPTS • Security and privacy → Privacy protections; Privacy protections; • Networks → Network measurement; Network measurement. KEYWORDS Search engines, advertising systems, cross-site tracking, privacy, measurement. ACM Reference Format: Salim Chouaki, Oana Goga, Hamed Haddadi, and Peter Snyder. 2023. Understanding the Privacy Risks of Popular Search Engine Advertising Systems. In Proceedings of the 2023 ACM Internet Measurement Conference (IMC ’23), October 24–26, 2023, Montreal, QC, Canada. ACM, New York, NY, USA, 15 pages. https://doi.org/10. 1145/3618257.3624823 1 INTRODUCTION Privacy-focused search engines such as DuckDuckGo, StartPage, and Qwant [3, 9, 10] promote a strategy of respecting users’ privacy and promise not to track users’ search and browsing behavior, all while delivering relevant search results. However, private search engines rely on advertising for revenue, and use traditional advertising platforms to deliver ads: DuckDuckGo and Qwant use Microsoft’s advertising system, while StartPage uses Google’s advertising system. These search engines are often ambiguous on the privacy properties of the ads that appear on their search page, and their consequent privacy properties remain unexplored to the best of our knowledge. In this work, we aim to fill this gap by conducting the first study of the privacy properties of the advertising systems of three major privacy-focused search engines: DuckDuckGo, StartPage, and Qwant, and how they compare to more popular search engines: Bing and Google. We investigate the privacy properties of these search engines when they: (i) present search ads to users, (ii) when a user clicks on an ad, and (iii) when the user lands on the advertiser’s page. We implement an automated measurement methodology to measure if and how users can be re-identified (hence, their privacy is compromised) when clicking on search ads on each search engine (see Section 3). We build an opensource implementation of this methodology in the form of a Puppeteer-based pipeline that simulates search queries and ad clicks. We apply this crawling methodology to the five search engines, providing a full dataset with visited websites, cookies created, locally stored values, and web requests to search engines’ servers and/or other third parties when clicking ads. We use filter rules from several major open-source lists to detect web requests to online trackers, and we propose a methodology to differentiate user identifiers from non-tracking values in query parameters and cookies values. We then present in Section 4 a systematic analysis of our dataset to investigate privacy harms before clicking an ad, during clicking an ad, and after clicking an ad and reaching the advertiser’s website. We find that users’ privacy is not harmed until users click on an ad. Privacy-focused search engines do not appear to attempt to re-identify users across visits or queries and do not include resources from, or make network requests to known trackers. However, we find that users’ privacy is compromised by studied search engines in various ways once users click on an ad. all Disappointingly, we find that all search engines record additional information about the user and/or the users’ clicks after the user has clicked on an ad. Private search engines capture data related to the clicked ad, including the ad provider, destination URL, and the ad’s position within the search results page, along with the user’s browsing data, such as the search query, device type, and browser language. Private search engines do not store user identifiers upon ad clicks, in contrast to traditional search engines that record user identifying values. Furthermore, we find that all search engines engage in navigation-based tracking. Navigation based tracking refers to tracking techniques that are redirecting users through one or more redirectors when navigating from one website to another in order to share user information across sites [33]. Navigation-based tracking does not require third-party cookies and can be used to circumvent browsers’ privacy protections from cross-site tracking using partitioned cookies storage. Alarmingly, we observe that privacy-focused search engines engage in more navigationbased tracking than non-privacy-focused ones: We observe navigational tracking on 4% ad clicks on Bing, on 100% ad clicks on Google, on 100% ad clicks on DuckDuckGo, on 86% ad clicks on Qwant, and on 100% ad clicks on StartPage. On the destination page, we check whether the search engine requires advertisers to abide by privacy-respecting practices by measuring whether advertisers include trackers or other known privacy-harming resources. We found that 93% of ads destination pages (across all five search engines) included tracker and privacy-harming resources. Finally, we check whether search engines or redirectors aid advertisers in profiling visitors by measuring the data they receive in the form of user-describing query params. We find that advertisers receive user identifiers in 68%, 92%, and 53% of cases for DuckDuckGo, StartPage, and Qwant, respectively. This practice, known as UID smuggling, enables redirectors to aggregate more user behavior data if they have scripts on the ads’ destination websites and they store the user-identifying parameters they receive. Notably, in the case of private search engines, the user-identifying parameters are not set by the search engine but by the redirectors encountered between the search engine’s and the advertiser’s sites. Our results indicate that private search engines’ privacy protections do not sufficiently cover their advertising systems. Although these search engines refrain from identifying and tracking users and their ad clicks, the presence of ads from Google or Microsoft subjects users to the privacyinvasive practices performed by these two advertising platforms. When users click on ads on private search engines, they are often identified and tracked either by Google, Microsoft, or other third parties, through bounce tracking and UID smuggling techniques. Particularly, advertisers receive unique user identifiers through query parameters in most ad clicks, which can enable cross-site tracking even in privacyenhanced browsers that block third-party cookie tracking.

Understanding the Privacy Risks of Popular Search Engine Advertising Systems: Abstract and Intro

This paper is available on arxiv under CC 4.0 license. Authors: (1) Pietro Ghezzi, Brighton & Sussex Medical School, Falmer, Brighton, UK; (2) Peter G Bannister, Brighton & Sussex Medical School, Falmer, Brighton, UK; (3) Gonzalo Casino, Communication Department, Pompeu Fabra University, Barcelona, Spain and Iberoamerican Cochrane Center, Barcelona, Spain; (4) Alessia Catalani, Department of Biomolecular Sciences, University of Urbino Carlo Bo, Urbino, PU, Italy; (5) Michel Goldman, Institute for Interdisciplinary Innovation in healthcare (I3h), Université libre de Bruxelles; (6) Jessica Morley, Oxford Internet Institute, University of Oxford, Oxford, UK; (7) Marie Neunez, Institute for Interdisciplinary Innovation in healthcare (I3h), Université libre de Bruxelles; (8) Andreu Prados, Communication Department, Pompeu Fabra University, Barcelona, Spain, Iberoamerican Cochrane Center, Barcelona, Spain, Department of Biomolecular Sciences, University of Urbino Carlo Bo, Urbino, PU, Italy, Institute for Interdisciplinary Innovation in healthcare (I3h), Université libre de Bruxelles, Oxford Internet Institute, University of Oxford, Oxford, UK, and Blanquerna School of Health Sciences, Universitat Ramon Llull, Barcelona, Spain; (9) Mariarosaria Taddeo, Oxford Internet Institute, University of Oxford, Oxford, UK, Blanquerna School of Health Sciences, Universitat Ramon Llull, Barcelona, Spain, and The Alan Turing Institute, London, UK; (10) Tania Vanzolini, Department of Biomolecular Sciences, University of Urbino Carlo Bo, Urbino, PU, Italy; (11) Luciano Floridi, Oxford Internet Institute, University of Oxford, Oxford, UK, Blanquerna School of Health Sciences, Universitat Ramon Llull, Barcelona, Spain, and The Alan Turing Institute, London, UK. Table of Links Abstract and Introduction Methods Web search and content analysis Results Discussion Conclusion and References Table 1. Example of classification of webpages Table 2. Percentage of vaccine-negative webpages in different SERPs Corresponding author: Professor Pietro Ghezzi, Brighton and Sussex Medical School, Trafford Centre, University of Sussex, Falmer, Brighton, BN19RY, UK. Email: p.ghezzi@bsms.ac.uk; phone: +44-(0)1273-873112 Abstract The fact that Internet companies may record our personal data and track our online behaviour for commercial or political purpose has emphasized aspects related to online privacy. This has also led to the development of search engines that promise no tracking and privacy. Search engines also have a major role in spreading low-quality health information such as that of anti-vaccine websites. This study investigates the relationship between search engines’ approach to privacy and the scientific quality of the information they return. We analyzed the first 30 webpages returned searching “vaccines autism” in English, Spanish, Italian and French. The results show that “alternative” search engines (Duckduckgo, Ecosia, Qwant, Swisscows and Mojeek) may return more anti-vaccine pages (10-53%) than Google.com (0%). Some localized versions of Google, however, returned more anti-vaccine webpages (up to 10%) than Google.com. Our study suggests that designing a search engine that is privacy savvy and avoids issues with filter bubbles that can result from user-tracking is necessary but insufficient; instead mechanisms should be developed to test search engines from the perspective of information quality (particularly for health-related webpages), before they can be deemed trustworthy providers of public health information. Introduction The World Health Organization lists vaccine hesitancy as one of the top ten threats to global health in 2019 1, requiring ongoing global monitoring. Despite the fact that the 1998 study which incorrectly suggested that the MMR vaccine could cause autism in children and prompted antivaccine beliefs 2 has now been discredited, misinformation, and indeed, disinformation, about vaccines continues to be published on the Internet perpetuating such beliefs. It has been suggested that this misinformation plays a role in the current low uptake of vaccines in developed countries 3. Understanding whether this is actually the case, and how to address this issue, is crucial as web-based sources of health information may be instrumental to solve the sustainability challenge currently facing health systems across the globe. The accuracy of information provided by a website is a key indicator of its overall information quality (IQ). The broader aspects of IQ have been the subject of many studies 5, but health IQ and trustworthiness of the sources have only partially been characterized 6. Studies looking at the influence of variations in eHealth literacy levels 7 and trust in different sources of online health information2 indicate that the relationship is not linear in all cases, i.e. higher health IQ does not result automatically in higher perceived levels of trustworthiness. For example, in a study by Chen et al. 8, 618 people were recruited to complete a survey which tested their eHealth literacy level, asked them to identify which of 25 sources of health information they used, and how much they trusted each source. The study showed that people with lower eHealth literacy were less likely to trust medical websites (typically higher IQ) and more likely to trust social media, blogs, and celebrity webpages (typically lower IQ). This might seem a spurious result, were it not for the fact that research has shown that those with high eHealth literacy assess more accurately the credibility and relevance of online health information, whereas those with low eHealth literacy often struggle to locate and understand eHealth information 9. This difficulty lowers their self-efficacy 10, distorts their perception of source credibility, and impacts negatively perceived trustworthiness 9, ultimately creating a need for individuals with low eHealth literacy to find an alternative means of determining trustworthiness in online sources of information. One such alternative is social endorsement. Visible social endorsement, e.g. ‘likes,’ 11 enables those with low eHealth literacy to determine trust based on the bandwagon heuristic and assume that, if the source has already been deemed valid by others, then it is safe for them to trust it too 10,12. Traditional medical websites afford those with low eHealth literacy no such alternative means of determining credibility and trust. This suggests that those who are more vulnerable to the real-world effects of both disinformation and misinformation (e.g. declining to vaccinate their children) are more likely to rely on online sources with lower health IQ, which are more prone to spread such inflammatory and inaccurate information. Personalisation of online search results may favour this phenomenon and lead to a vicious cycle where the more one searches and reads dis- and mis-information about vaccines the less one finds and reads scientific information on the same topic. At the same time, there are increasing concerns about the privacy risks associated with Internet search engines storing potentially sensitive and private health information contained within users search histories, combining it with additional information collected for tracking purposes, and using these data for commercial or other purposes 13,14. This creates a public push back against the idea that search engines or public health providers should interfere in the results people see when searching for health information online. This makes it hard to address concerns about health disinformation/misinformation in a way that is at least socially acceptable if not ideally preferable 15. The UK’s National Health Service (NHS) discovered this when its announcement of Amazon Alexa responding with guaranteed high-IQ content from NHS.UK to user voice queries such as “Alexa, how do I treat a migraine,” resulted in a public outcry over privacy infringement 16. This raises the question whether, in the context of online vaccination information, it is possible at all to balance concerns about user privacy and IQ. The following pages seek to lay the foundations for answering this question by reporting on the methodology for, and results of, a study concerning the following research question: “What is the current relationship between search engines’ approach to privacy and the scientific quality of the information they return?”.

Online Information of Vaccines: Abstract and Introduction

This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Anh Nguyen-Duc, University of South Eastern Norway, BøI Telemark, Norway3800 and Norwegian University of Science and Technology, Trondheim, Norway7012;
(2) Beatriz Cabrero-Daniel, University of Gothenburg, Gothenburg, Sweden;
(3) Adam Przybylek, Gdansk University of Technology, Gdansk, Poland;
(4) Chetan Arora, Monash University, Melbourne, Australia;
(5) Dron Khanna, Free University of Bozen-Bolzano, Bolzano, Italy;
(6) Tomas Herda, Austrian Post - Konzern IT, Vienna, Austria;
(7) Usman Rafiq, Free University of Bozen-Bolzano, Bolzano, Italy;
(8) Jorge Melegati, Free University of Bozen-Bolzano, Bolzano, Italy;
(9) Eduardo Guerra, Free University of Bozen-Bolzano, Bolzano, Italy;
(10) Kai-Kristian Kemell, University of Helsinki, Helsinki, Finland;
(11) Mika Saari, Tampere University, Tampere, Finland;
(12)  Zheying Zhang, Tampere University, Tampere, Finland;
(13) Huy Le, Vietnam National University Ho Chi Minh City, Hochiminh City, Vietnam and Ho Chi Minh City University of Technology, Hochiminh City, Vietnam;
(14) Tho Quan, Vietnam National University Ho Chi Minh City, Hochiminh City, Vietnam and Ho Chi Minh City University of Technology, Hochiminh City, Vietnam;
(15) Pekka Abrahamsson, Tampere University, Tampere, Finland. Table of Links Abstract and Introduction
Background
Research Approach
Research Agenda
Outlook and Conclusions
References Abstract Context Generative Artificial Intelligence (GenAI) tools have become increasingly prevalent in software development, offering assistance to various managerial and technical project activities. Notable examples of these tools include OpenAI’s ChatGPT, GitHub Copilot, and Amazon CodeWhisperer. Objective. Although many recent publications have explored and evaluated the application of GenAI, a comprehensive understanding of the current development, applications, limitations, and open challenges remains unclear to many. Particularly, we do not have an overall picture of the current state of GenAI technology in practical software engineering usage scenarios. Method. We conducted a literature review and focus groups for a duration of five months to develop a research agenda on GenAI for Software Engineering. Results. We identified 78 open Research Questions (RQs) in 11 areas of Software Engineering. Our results show that it is possible to explore the adoption of GenAI in partial automation and support decision-making in all software development activities. While the current literature is skewed toward software implementation, quality assurance and software maintenance, other areas, such as requirements engineering, software design, and software engineering education, would need further research attention. Common considerations when implementing GenAI include industry level assessment, dependability and accuracy, data accessibility, transparency, and sustainability aspects associated with the technology. Conclusions. GenAI is bringing significant changes to the field of software engineering. Nevertheless, the state of research on the topic still remains immature. We believe that this research agenda holds significance and practical value for informing both researchers and practitioners about current applications and guiding future research. Keywords: Generative Artificial Intelligence, GenAI, ChatGPT, CoPilot, Software Engineering, Software Project, Software Development, Focus Group, Literature review, Structured review, Literature survey, Research Agenda, Research Roadmap 1. Introduction In recent years, Generative Artificial Intelligence (GenAI) has attracted significant attention and interest from Software Engineering (SE) research and practice. GenAI tools such as GitHub Copilot1 and ChatGPT2 have been rapidly explored in various professional contexts given their remarkable performance in producing human-like content. The emerging adoption of  these tools reintroduces a host of classic concerns about productivity and quality when adopting new technologies. Clearly, code generation and test case optimization are SE tasks that directly benefit from recent large language models (LLMs) [1, 2, 3, 4]. Beyond traditional research on applied. Machine Learning (ML), GenAI tools introduced usability and accessibility, leveraging AI-generated content to a broader range of professionals, requiring less technical competence to work and integrate them into existing work environment. GenAI tools are also showing their potential in non-coding tasks, such as assisting requirements engineering, software processes and project management. At the moment, GenAI is an active research areas with several challenges and open questions. To do well on specific tasks, LLMs require fine-tuning or training. Research has much focused so far on achieving reliable and scalable GenAI output for different SE tasks. GenAI models are inherently nondeterministic: the same prompt produces different answers on different inference executions [5]. Moreover, GenAI’s output can be very sensitive to the input parameters or settings [6, 7]. Hallucination is another common concern with AI-generated content, as they can be imaginary or fictional [8]. The promise of AI automation might be closer than ever in SE [9], when these open challenges are sufficiently addressed. Research agenda is a popular type of work for guiding and organizing research efforts in certain research areas [10, 11, 12, 13, 14, 15]. It often includes a review of existing work, and a presentation of directions, visions and priorities, enabling researchers to make meaningful contributions to their respective fields and to address pressing challenges. This research agenda is driven by past and current work on GenAI for SE. Based on focus groups and a literature review, we portray the current research on GenAI in SE and present the research agenda with open challenges for future research. While we do not emphasize the comprehensiveness of the literature review due to the fast-moving nature of the topic, the focus groups present practical expectations of the future roles of GenAI in software development. The research agenda is organized into 11 areas of concern, which are: (1) Requirements Engineering, (2) Software Design,(3) Sofware Implementation, (4) Quality Assurance, (5) Software Maintenance and Evolution, (6) Software Processes and Tools, (7) Engineering Management, (8) Professional Competencies, (9) Software Engineering Education, (10) Macro Aspects and (11) Fundamental concerns of GenAI. Two international events have provided a ground for this work. As a part of the first international workshop on AI-assisted Agile Software Development3 , we ran our first focus group to explore the benefits and challenges of AI-assisted Agile software development. At the Requirements Engineering (RE) conference, we organized the second international workshop on Requirements Engineering for Software Startups and Emerging Technologies (RESET) 4 with a special theme on GenAI and Requirements Engineering. Consequently, the identified open challenges reflect not just the existing gaps in literature but also draw from the practical, value-driven insights of the co-authors. The paper is structured as follows: we clarify the meaning of GenAI and what we know about GenAI in SE in Section 2, and we will present our research approach in Section 3. Section 4 presents our Research Agenda. Finally, Section 5 is Outlook and Conclusions. This paper is available on arxiv under CC 4.0 license. Authors: (1) Anh Nguyen-Duc, University of South Eastern Norway, BøI Telemark, Norway3800 and Norwegian University of Science and Technology, Trondheim, Norway7012; (2) Beatriz Cabrero-Daniel, University of Gothenburg, Gothenburg, Sweden; (3) Adam Przybylek, Gdansk University of Technology, Gdansk, Poland; (4) Chetan Arora, Monash University, Melbourne, Australia; (5) Dron Khanna, Free University of Bozen-Bolzano, Bolzano, Italy; (6) Tomas Herda, Austrian Post - Konzern IT, Vienna, Austria; (7) Usman Rafiq, Free University of Bozen-Bolzano, Bolzano, Italy; (8) Jorge Melegati, Free University of Bozen-Bolzano, Bolzano, Italy; (9) Eduardo Guerra, Free University of Bozen-Bolzano, Bolzano, Italy; (10) Kai-Kristian Kemell, University of Helsinki, Helsinki, Finland; (11) Mika Saari, Tampere University, Tampere, Finland; (12)  Zheying Zhang, Tampere University, Tampere, Finland; (13) Huy Le, Vietnam National University Ho Chi Minh City, Hochiminh City, Vietnam and Ho Chi Minh City University of Technology, Hochiminh City, Vietnam; (14) Tho Quan, Vietnam National University Ho Chi Minh City, Hochiminh City, Vietnam and Ho Chi Minh City University of Technology, Hochiminh City, Vietnam; (15) Pekka Abrahamsson, Tampere University, Tampere, Finland. This paper is available on arxiv under CC 4.0 license. Authors: Authors: (1) Anh Nguyen-Duc, University of South Eastern Norway, BøI Telemark, Norway3800 and Norwegian University of Science and Technology, Trondheim, Norway7012; (2) Beatriz Cabrero-Daniel, University of Gothenburg, Gothenburg, Sweden; (3) Adam Przybylek, Gdansk University of Technology, Gdansk, Poland; (4) Chetan Arora, Monash University, Melbourne, Australia; (5) Dron Khanna, Free University of Bozen-Bolzano, Bolzano, Italy; (6) Tomas Herda, Austrian Post - Konzern IT, Vienna, Austria; (7) Usman Rafiq, Free University of Bozen-Bolzano, Bolzano, Italy; (8) Jorge Melegati, Free University of Bozen-Bolzano, Bolzano, Italy; (9) Eduardo Guerra, Free University of Bozen-Bolzano, Bolzano, Italy; (10) Kai-Kristian Kemell, University of Helsinki, Helsinki, Finland; (11) Mika Saari, Tampere University, Tampere, Finland; (12)  Zheying Zhang, Tampere University, Tampere, Finland; (13) Huy Le, Vietnam National University Ho Chi Minh City, Hochiminh City, Vietnam and Ho Chi Minh City University of Technology, Hochiminh City, Vietnam; (14) Tho Quan, Vietnam National University Ho Chi Minh City, Hochiminh City, Vietnam and Ho Chi Minh City University of Technology, Hochiminh City, Vietnam; (15) Pekka Abrahamsson, Tampere University, Tampere, Finland. Table of Links Abstract and Introduction Background Research Approach Research Agenda Outlook and Conclusions References Abstract and Introduction Abstract and Introduction Background Background Research Approach Research Approach Research Agenda Research Agenda Outlook and Conclusions Outlook and Conclusions References References Abstract Context Generative Artificial Intelligence (GenAI) tools have become increasingly prevalent in software development, offering assistance to various managerial and technical project activities. Notable examples of these tools include OpenAI’s ChatGPT, GitHub Copilot, and Amazon CodeWhisperer. Context Objective. Although many recent publications have explored and evaluated the application of GenAI, a comprehensive understanding of the current development, applications, limitations, and open challenges remains unclear to many. Particularly, we do not have an overall picture of the current state of GenAI technology in practical software engineering usage scenarios. Objective. Method . We conducted a literature review and focus groups for a duration of five months to develop a research agenda on GenAI for Software Engineering. Method Results . We identified 78 open Research Questions (RQs) in 11 areas of Software Engineering. Our results show that it is possible to explore the adoption of GenAI in partial automation and support decision-making in all software development activities. While the current literature is skewed toward software implementation, quality assurance and software maintenance, other areas, such as requirements engineering, software design, and software engineering education, would need further research attention. Common considerations when implementing GenAI include industry level assessment, dependability and accuracy, data accessibility, transparency, and sustainability aspects associated with the technology. Results Conclusions. GenAI is bringing significant changes to the field of software engineering. Nevertheless, the state of research on the topic still remains immature. We believe that this research agenda holds significance and practical value for informing both researchers and practitioners about current applications and guiding future research. Conclusions. Keywords: Generative Artificial Intelligence, GenAI, ChatGPT, CoPilot, Software Engineering, Software Project, Software Development, Focus Group, Literature review, Structured review, Literature survey, Research Agenda, Research Roadmap Keywords: 1. Introduction In recent years, Generative Artificial Intelligence (GenAI) has attracted significant attention and interest from Software Engineering (SE) research and practice. GenAI tools such as GitHub Copilot1 and ChatGPT2 have been rapidly explored in various professional contexts given their remarkable performance in producing human-like content. The emerging adoption of  these tools reintroduces a host of classic concerns about productivity and quality when adopting new technologies. Clearly, code generation and test case optimization are SE tasks that directly benefit from recent large language models (LLMs) [1, 2, 3, 4]. Beyond traditional research on applied. Machine Learning (ML), GenAI tools introduced usability and accessibility, leveraging AI-generated content to a broader range of professionals, requiring less technical competence to work and integrate them into existing work environment. GenAI tools are also showing their potential in non-coding tasks, such as assisting requirements engineering, software processes and project management. At the moment, GenAI is an active research areas with several challenges and open questions. To do well on specific tasks, LLMs require fine-tuning or training. Research has much focused so far on achieving reliable and scalable GenAI output for different SE tasks. GenAI models are inherently nondeterministic: the same prompt produces different answers on different inference executions [5]. Moreover, GenAI’s output can be very sensitive to the input parameters or settings [6, 7]. Hallucination is another common concern with AI-generated content, as they can be imaginary or fictional [8]. The promise of AI automation might be closer than ever in SE [9], when these open challenges are sufficiently addressed. Research agenda is a popular type of work for guiding and organizing research efforts in certain research areas [10, 11, 12, 13, 14, 15]. It often includes a review of existing work, and a presentation of directions, visions and priorities, enabling researchers to make meaningful contributions to their respective fields and to address pressing challenges. This research agenda is driven by past and current work on GenAI for SE. Based on focus groups and a literature review, we portray the current research on GenAI in SE and present the research agenda with open challenges for future research. While we do not emphasize the comprehensiveness of the literature review due to the fast-moving nature of the topic, the focus groups present practical expectations of the future roles of GenAI in software development. The research agenda is organized into 11 areas of concern, which are: (1) Requirements Engineering, (2) Software Design,(3) Sofware Implementation, (4) Quality Assurance, (5) Software Maintenance and Evolution, (6) Software Processes and Tools, (7) Engineering Management, (8) Professional Competencies, (9) Software Engineering Education, (10) Macro Aspects and (11) Fundamental concerns of GenAI. Two international events have provided a ground for this work. As a part of the first international workshop on AI-assisted Agile Software Development3 , we ran our first focus group to explore the benefits and challenges of AI-assisted Agile software development. At the Requirements Engineering (RE) conference, we organized the second international workshop on Requirements Engineering for Software Startups and Emerging Technologies (RESET) 4 with a special theme on GenAI and Requirements Engineering. Consequently, the identified open challenges reflect not just the existing gaps in literature but also draw from the practical, value-driven insights of the co-authors. The paper is structured as follows: we clarify the meaning of GenAI and what we know about GenAI in SE in Section 2, and we will present our research approach in Section 3. Section 4 presents our Research Agenda. Finally, Section 5 is Outlook and Conclusions.

Generative Artificial Intelligence for Software Engineering: Abstract and Introduction

This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Ghazaleh H. Torbati, Max Planck Institute for Informatics Saarbrucken, Germany & ghazaleh@mpi-inf.mpg.de;
(2) Andrew Yates, University of Amsterdam Amsterdam, Netherlands & a.c.yates@uva.nl;
(3) Anna Tigunova, Max Planck Institute for Informatics Saarbrucken, Germany & tigunova@mpi-inf.mpg.de;
(4) Gerhard Weikum, Max Planck Institute for Informatics Saarbrucken, Germany & weikum@mpi-inf.mpg.de. Table of Links Abstract and Introduction
Related Work
Methodology
Experimental Design
Experimental Results
Conclusion
Ethics Statement and References Abstract Recommender systems are most successful for popular items and users with ample interactions (likes, ratings etc.). This work addresses the difficult and underexplored case of supporting users who have very sparse interactions but post informative review texts. Our experimental studies address two book communities with these characteristics. We design a framework with Transformer-based representation learning, covering user-item interactions, item content, and user-provided reviews. To overcome interaction sparseness, we devise techniques for selecting the most informative cues to construct concise user profiles. Comprehensive experiments, with datasets from Amazon and Goodreads, show that judicious selection of text snippets achieves the best performance, even in comparison to LLMgenerated rankings and to using LLMs to generate user profiles. Index Terms—recommender, sparse data, user text, language model, transformer I. INTRODUCTION A. Motivation Recommender systems have become a mature technology and are ubiquitous in applications [1]. Methodologies fall into two major families or hybrid combinations: i) interactionbased recommenders that leverage binary signals (e.g., membership in personal playlists or libraries) or numeric ratings for user-item pairs, and ii) content-based recommenders that exploit item features and user-provided content, ranging from meta-data attributes (e.g., item categories) all the way to textually rich reviews. In many settings, where user-item signals are easily collectable and abundant, interaction-based recommenders achieve excellent performance and are hard to beat. Some applications combine interaction data with informative item features, again relying on massive data (e.g., [2]). In contrast, the data-poor regime of long-tail users and long-tail items has received less attention. Some works advocate countering data sparseness by clustering long-tail items (e.g., [3]). However, this is not sufficient to gear up for users with few interactions but diverse tastes. Approaches for cold-start and cross-domain recommendations and zero-shot learning (e.g., [4]–[7]), are often focused on the item side only: handling new items by transferring (latent) knowledge from existing ones with dense data. Note that sparseness is not just a lack of user-item interactions, but also refers to the users and items themselves: users with merely a handful of items do not exhibit a clear picture of their tastes, and items that are completely unseen at training time are a challenging case. Moreover, even users with a good number of items can be a major problem when the preferred items are highly diverse, such as less known books from all kinds of genres. To address these difficulties, the most promising approach is to leverage review texts by users. In settings where users spend substantial time per item, such as books or travel destinations (as opposed to items with short attention spans, like music or restaurants), even data-poor users with few interactions may leave rich texts that express their interests and tastes. The approach has been studied before (e.g., [8]–[11]), but the performance is much lower than for settings with ample interaction data. This paper aims to advance the state of the art on this theme, by tackling both issues of data sparseness and diversity of tastes via textbased concise user profiles. B. Research Questions Our approach constructs and encodes concise user profiles from reviews of data-poor text-rich users, focusing on the book domain for concreteness (using data from Goodreads and Amazon). Unlike movies or restaurants, books have a much longer tail of popularity and exhibit huge diversity of user tastes, making books a particularly challenging domain for data-poor recommendations. Also, negative reviews are very rare (most have ratings like 4 or 5 stars), so that we deal with binary data and have no negative samples. This approach raises several research questions (RQs): RQ1: Learning with Language Models. How can we best incorporate language models, like BERT, T5 or ChatGPT, for item ranking or for encoding the gist of a user’s reviews, enhanced with the LM’s world knowledge? RQ2: Long Text. Since neural architectures can consume only limited input size (e.g., 512 tokens), we have to select a subset of snippets. Even with a large capacity, the computational and energy cost is typically quadratic in the number of input tokens, so that being selective has benefits. What are the most informative cues that should be fed into the model? RQ3: Review Aspects. User reviews express a mix of aspects, including personal background (e.g., “I’m a retired teacher”), sentiment expressions (e.g., “what a great story”), general emotions (e.g., “brings back lovely memories”), and comments about the book contents. As sentiments do not add value over the already known fact that the book was liked, only the content comments yield cues towards new recommendations. How should this issue be considered in constructing a concise user profile with informative cues? C. Approach We devise a general framework, called CUP, for constructing Concise User Profiles, and encoding them in a recommender system, which allows us to configure a variety of methods. We propose a Tranformer-based architecture that supports end-to-end learning of item and user encodings, making use of a language model. Our choice for the language model is BERT; alternatives such as T5 can be easily plugged in but incur higher computational cost. The end-to-end learning operates on sufficiently short, judiciously constructed profiles, and leverages the available user item interactions as well. On top of the Tranformer, we place feed-forward layers for a supervised classifier. The classifier and the representation learning are jointly trained end-to-end. The prediction scores for user-item pairs yield the per-user ranking of items. This architecture can take as input a spectrum of user and item information including user reviews and item descriptions and metadata (e.g., categories). The rationale for this design is twofold: i) ensuring that inference-time computations are efficient, leveraging trainingtime-computed vectors for users and items and solely computing a scalar product at run-time, and ii) being able to express a wide spectrum of user profiling techniques within the same architecture. Alternative architectures, such as CNNbased, or LLM-based with smart prompts, are baselines in our experiments. We specifically devise new techniques for coping with the case where users have few reviews with long and noisy text (see RQ2 and RQ3), such that their total length exceeds the number of tokens that the encoders can consume. Our techniques are based on information-retrieval-style statistics (like idf weights) for n-grams or sentences, or on neural similarity measures (using SentenceBERT), or on running review texts through a generative language model, like T5 or ChatGPT, for generating user profiles. Table VII illustrates the role of concise user profiles, by giving examples for two users of Amazon books and the Goodreads community, with profiles computed by different methods. In experiments, we limit the number of tokens per user profile, as a stress-test. The main argument for this choice is to keep the system as light-weight as possible. The computational cost of the Transformer, and thus also the environmental footprint, are quadratic in the number of input tokens. Section V-D experimentally shows that a small token budget is competitive to larger ones, while having advantages in efficiency. D. Contributions The salient contributions of this work are the following: • a new framework, called CUP, for Transformer-based recommenders that leverage language models and concise user profiles from reviews; • judicious techniques for selecting and encoding informative cues from the long and noisy text of user reviews, outperforming prior baselines as well as methods based on large language models (LLMs); • comprehensive experiments with focus on data-poor but text-rich users with highly diverse preferences, and difficult predictions (e.g., recommending new authors, not just new books by previously seen authors). We make code, data and experimental details available on our project web page[1]. [1] https://personalization.mpi-inf.mpg.de/CUP This paper is available on arxiv under CC 4.0 license. Authors: (1) Ghazaleh H. Torbati, Max Planck Institute for Informatics Saarbrucken, Germany & ghazaleh@mpi-inf.mpg.de; (2) Andrew Yates, University of Amsterdam Amsterdam, Netherlands & a.c.yates@uva.nl; (3) Anna Tigunova, Max Planck Institute for Informatics Saarbrucken, Germany & tigunova@mpi-inf.mpg.de; (4) Gerhard Weikum, Max Planck Institute for Informatics Saarbrucken, Germany & weikum@mpi-inf.mpg.de. This paper is available on arxiv under CC 4.0 license. Authors: (1) Ghazaleh H. Torbati, Max Planck Institute for Informatics Saarbrucken, Germany & ghazaleh@mpi-inf.mpg.de; (2) Andrew Yates, University of Amsterdam Amsterdam, Netherlands & a.c.yates@uva.nl; (3) Anna Tigunova, Max Planck Institute for Informatics Saarbrucken, Germany & tigunova@mpi-inf.mpg.de; (4) Gerhard Weikum, Max Planck Institute for Informatics Saarbrucken, Germany & weikum@mpi-inf.mpg.de. Table of Links Abstract and Introduction Related Work Methodology Experimental Design Experimental Results Conclusion Ethics Statement and References Abstract and Introduction Abstract and Introduction Related Work Related Work Methodology Methodology Experimental Design Experimental Design Experimental Results Experimental Results Conclusion Conclusion Ethics Statement and References Ethics Statement and References Abstract Recommender systems are most successful for popular items and users with ample interactions (likes, ratings etc.). This work addresses the difficult and underexplored case of supporting users who have very sparse interactions but post informative review texts. Our experimental studies address two book communities with these characteristics. We design a framework with Transformer-based representation learning, covering user-item interactions, item content, and user-provided reviews. To overcome interaction sparseness, we devise techniques for selecting the most informative cues to construct concise user profiles. Comprehensive experiments, with datasets from Amazon and Goodreads, show that judicious selection of text snippets achieves the best performance, even in comparison to LLMgenerated rankings and to using LLMs to generate user profiles. Index Terms —recommender, sparse data, user text, language model, transformer Index Terms I. INTRODUCTION A. Motivation A. Motivation Recommender systems have become a mature technology and are ubiquitous in applications [1]. Methodologies fall into two major families or hybrid combinations: i) interactionbased recommenders that leverage binary signals (e.g., membership in personal playlists or libraries) or numeric ratings for user-item pairs, and ii) content-based recommenders that exploit item features and user-provided content, ranging from meta-data attributes (e.g., item categories) all the way to textually rich reviews. In many settings, where user-item signals are easily collectable and abundant, interaction-based recommenders achieve excellent performance and are hard to beat. Some applications combine interaction data with informative item features, again relying on massive data (e.g., [2]). In contrast, the data-poor regime of long-tail users and long-tail items has received less attention. Some works advocate countering data sparseness by clustering long-tail items (e.g., [3]). However, this is not sufficient to gear up for users with few interactions but diverse tastes. Approaches for cold-start and cross-domain recommendations and zero-shot learning (e.g., [4]–[7]), are often focused on the item side only: handling new items by transferring (latent) knowledge from existing ones with dense data. Note that sparseness is not just a lack of user-item interactions, but also refers to the users and items themselves: users with merely a handful of items do not exhibit a clear picture of their tastes, and items that are completely unseen at training time are a challenging case. Moreover, even users with a good number of items can be a major problem when the preferred items are highly diverse, such as less known books from all kinds of genres. To address these difficulties, the most promising approach is to leverage review texts by users. In settings where users spend substantial time per item, such as books or travel destinations (as opposed to items with short attention spans, like music or restaurants), even data-poor users with few interactions may leave rich texts that express their interests and tastes. The approach has been studied before (e.g., [8]–[11]), but the performance is much lower than for settings with ample interaction data. This paper aims to advance the state of the art on this theme, by tackling both issues of data sparseness and diversity of tastes via textbased concise user profiles. B. Research Questions B. Research Questions Our approach constructs and encodes concise user profiles from reviews of data-poor text-rich users, focusing on the book domain for concreteness (using data from Goodreads and Amazon). Unlike movies or restaurants, books have a much longer tail of popularity and exhibit huge diversity of user tastes, making books a particularly challenging domain for data-poor recommendations. Also, negative reviews are very rare (most have ratings like 4 or 5 stars), so that we deal with binary data and have no negative samples. This approach raises several research questions (RQs): RQ1: Learning with Language Models. How can we best incorporate language models, like BERT, T5 or ChatGPT, for item ranking or for encoding the gist of a user’s reviews, enhanced with the LM’s world knowledge? Learning with Language Models. RQ2: Long Text. Since neural architectures can consume only limited input size (e.g., 512 tokens), we have to select a subset of snippets. Even with a large capacity, the computational and energy cost is typically quadratic in the number of input tokens, so that being selective has benefits. What are the most informative cues that should be fed into the model? Long Text. RQ3: Review Aspects. User reviews express a mix of aspects, including personal background (e.g., “I’m a retired teacher”), sentiment expressions (e.g., “what a great story”), general emotions (e.g., “brings back lovely memories”), and comments about the book contents. As sentiments do not add value over the already known fact that the book was liked, only the content comments yield cues towards new recommendations. How should this issue be considered in constructing a concise user profile with informative cues? Review Aspects. C. Approach C. Approach We devise a general framework, called CUP, for constructing Concise User Profiles, and encoding them in a recommender system, which allows us to configure a variety of methods. We propose a Tranformer-based architecture that supports end-to-end learning of item and user encodings, making use of a language model. Our choice for the language model is BERT; alternatives such as T5 can be easily plugged in but incur higher computational cost. The end-to-end learning operates on sufficiently short, judiciously constructed profiles, and leverages the available user item interactions as well. On top of the Tranformer, we place feed-forward layers for a supervised classifier. The classifier and the representation learning are jointly trained end-to-end. The prediction scores for user-item pairs yield the per-user ranking of items. This architecture can take as input a spectrum of user and item information including user reviews and item descriptions and metadata (e.g., categories). The rationale for this design is twofold: i) ensuring that inference-time computations are efficient, leveraging trainingtime-computed vectors for users and items and solely computing a scalar product at run-time, and ii) being able to express a wide spectrum of user profiling techniques within the same architecture. Alternative architectures, such as CNNbased, or LLM-based with smart prompts, are baselines in our experiments. We specifically devise new techniques for coping with the case where users have few reviews with long and noisy text (see RQ2 and RQ3), such that their total length exceeds the number of tokens that the encoders can consume. Our techniques are based on information-retrieval-style statistics (like idf weights) for n-grams or sentences, or on neural similarity measures (using SentenceBERT), or on running review texts through a generative language model, like T5 or ChatGPT, for generating user profiles. Table VII illustrates the role of concise user profiles, by giving examples for two users of Amazon books and the Goodreads community, with profiles computed by different methods. In experiments, we limit the number of tokens per user profile, as a stress-test. The main argument for this choice is to keep the system as light-weight as possible. The computational cost of the Transformer, and thus also the environmental footprint, are quadratic in the number of input tokens. Section V-D experimentally shows that a small token budget is competitive to larger ones, while having advantages in efficiency. D. Contributions D. Contributions The salient contributions of this work are the following: • a new framework, called CUP, for Transformer-based recommenders that leverage language models and concise user profiles from reviews; • judicious techniques for selecting and encoding informative cues from the long and noisy text of user reviews, outperforming prior baselines as well as methods based on large language models (LLMs); • comprehensive experiments with focus on data-poor but text-rich users with highly diverse preferences, and difficult predictions (e.g., recommending new authors, not just new books by previously seen authors). We make code, data and experimental details available on our project web page[1]. [1] https://personalization.mpi-inf.mpg.de/CUP

Recommendations by Concise User Profiles from Review Text: Abstract and Introduction

This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Xiaofei Sun, Zhejiang University & xiaofei_sun@zju.edu.cn;
(2) Xiaoya Li, Shannon.AI, Bytedance & xiaoya_li@shannonai.com;
(3) Shengyu Zhang, Zhejiang University & sy_zhang@zju.edu.cn;
(4) Shuhe Wang, Peking University & wangshuhe@stu.pku.edu.cn;
(5) Fei Wu, Zhejiang University & wufei@zju.edu.cn;
(6) Jiwei Li, Zhejiang University &  jiwei_li@zju.edu.cn;
(7) Tianwei Zhang, Nanyang Technological University & tianwei.zhang@ntu.edu.sg;
(8) Guoyin Wang, Bytedance & guoyin.wang@bytedance.com. Table of Links Abstract and Introduction
Related Work
LLM Negotiation for Sentiment Analysis
Experiments
Ablation Studies
Conclusion and References Abstract A standard paradigm for sentiment analysis is to rely on a singular LLM that makes the decision in a single round under the framework of in-context learning. This framework suffers the key disadvantage that the single-turn output generated by a single LLM might not deliver the perfect decision, just as humans sometimes need multiple attempts to get things right. This is especially true for the task of sentiment analysis where deep reasoning is required to address the complex linguistic phenomenon (e.g., clause composition, irony, etc) in the input. To address this issue, this paper introduces a multi-LLM negotiation framework for sentiment analysis. The framework consists of a reasoning-infused generator to provide decision along with rationale, a explanationderiving discriminator to evaluate the credibility of the generator. The generator and the discriminator iterate until a consensus is reached. The proposed framework naturally addressed the aforementioned challenge, as we are able to take the complementary abilities of two LLMs, have them use rationale to persuade each other for correction. Experiments on a wide range of sentiment analysis benchmarks (SST-2, Movie Review, Twitter, yelp, amazon, IMDB) demonstrate the effectiveness of proposed approach: it consistently yields better performances than the ICL baseline across all benchmarks, and even superior performances to supervised baselines on the Twitter and movie review datasets. 1 Introduction Sentiment analysis (Pang and Lee, 2008; Go et al., 2009; Maas et al., 2011a; Zhang and Liu, 2012; Baccianella et al., 2010; Medhat et al., 2014; Bakshi et al., 2016; Zhang et al., 2018) aims to extract opinion polarity expressed by a chunk of text. Recent advances in large language models (LLMs) (Brown et al., 2020; Ouyang et al., 2022; Touvron et al., 2023a,b; Anil et al., 2023; Zeng et al., 2022b; OpenAI, 2023; Bai et al., 2023) open a new door for the resolving the task (Lu et al., 2021; Kojima et al., 2022; Wang et al., 2022b; Wei et al., 2022b; Wan et al., 2023; Wang et al., 2023; Sun et al., 2023b,a; Lightman et al., 2023; Li et al., 2023; Schick et al., 2023): under the paradigm of in-context learning (ICL), LLMs are able to achieve performances comparable to supervised learning strategies (Lin et al., 2021; Sun et al., 2021; Phan and Ogunbona, 2020; Dai et al., 2021) with only a small number of training examples. Existing approaches that harness LLMs for sentiment analysis usually rely on a singular LLM, and make a decision in a single round under ICL. This strategy suffers from the following disadvantage: the single-turn output generated by a single LLM might not deliver the perfect response: Just as humans sometimes need multiple attempts to get things right, it might take multiple rounds before an LLM makes the right decision. This is especially true for the task of sentiment analysis, where LLMs usually need to articulate the reasoning process to address the complex linguistic phenomenon (e.g., clause composition, irony, etc) in the input sentence. To address the this issue, in this paper, we propose a multi-LLM negotiation strategy for sentiment analysis. The core of the proposed strategy is a generator-discriminator framework, where one LLM acts as the generator (G) to produce sentiment decisions, while the other acts as a discriminator (D), tasked with evaluating the credibility of the generated output from the first LLM. The proposed method innovates on three aspects: (1) Reasoning-infused generator (G): an LLM that adheres to a structured reasoning chain, enhancing the ICL of the generator while offering the discriminator the evidence and insights to evaluate its validity; (2) Explanation deriving discriminator (D); other LLM designed to offer post-evaluation rationales for its judgments; (3) Negotiation: two LLMs act as the roles of the generator and the discriminator, and perform the negotiation until a consensus is reached. This strategy harnesses the collective abilities of the two LLMs and provide with the channel for the model to correct imperfect responses, and thus naturally resolves the issue that a single LLM cannot yield the correct decision on its first try. The contributions of this work can be summarized as follows: 1) we provide a novel perspective on how sentiment analysis can benefit from multi-LLM negotiation. 2) we introduce a Generator Discriminator Role-switching DecisionMaking framework that enables multi-LLM collaboration through iteratively generating and validating sentiment categorizations. 3) our empirical findings offer evidence for the efficacy of the proposed approach: experiments on a wide range of sentiment analysis benchmarks (SST-2, Movie Review, Twitter, yelp, amazon, IMDB) demonstrate that the proposed method consistently yields better performances than the ICL baseline across all benchmarks, and even superior performances to supervised baselines on the Twitter and movie review datasets. This paper is available on arxiv under CC 4.0 license. Authors: (1) Xiaofei Sun, Zhejiang University & xiaofei_sun@zju.edu.cn; (2) Xiaoya Li, Shannon.AI, Bytedance & xiaoya_li@shannonai.com; (3) Shengyu Zhang, Zhejiang University & sy_zhang@zju.edu.cn; (4) Shuhe Wang, Peking University & wangshuhe@stu.pku.edu.cn; (5) Fei Wu, Zhejiang University & wufei@zju.edu.cn; (6) Jiwei Li, Zhejiang University &  jiwei_li@zju.edu.cn; (7) Tianwei Zhang, Nanyang Technological University & tianwei.zhang@ntu.edu.sg; (8) Guoyin Wang, Bytedance & guoyin.wang@bytedance.com. This paper is available on arxiv under CC 4.0 license. Authors: (1) Xiaofei Sun, Zhejiang University & xiaofei_sun@zju.edu.cn; (2) Xiaoya Li, Shannon.AI, Bytedance & xiaoya_li@shannonai.com; (3) Shengyu Zhang, Zhejiang University & sy_zhang@zju.edu.cn; (4) Shuhe Wang, Peking University & wangshuhe@stu.pku.edu.cn; (5) Fei Wu, Zhejiang University & wufei@zju.edu.cn; (6) Jiwei Li, Zhejiang University &  jiwei_li@zju.edu.cn; (7) Tianwei Zhang, Nanyang Technological University & tianwei.zhang@ntu.edu.sg; (8) Guoyin Wang, Bytedance & guoyin.wang@bytedance.com. Table of Links Abstract and Introduction Related Work LLM Negotiation for Sentiment Analysis Experiments Ablation Studies Conclusion and References Abstract and Introduction Abstract and Introduction Related Work Related Work LLM Negotiation for Sentiment Analysis LLM Negotiation for Sentiment Analysis Experiments Experiments Ablation Studies Ablation Studies Conclusion and References Conclusion and References Abstract A standard paradigm for sentiment analysis is to rely on a singular LLM that makes the decision in a single round under the framework of in-context learning. This framework suffers the key disadvantage that the single-turn output generated by a single LLM might not deliver the perfect decision, just as humans sometimes need multiple attempts to get things right. This is especially true for the task of sentiment analysis where deep reasoning is required to address the complex linguistic phenomenon (e.g., clause composition, irony, etc) in the input. To address this issue, this paper introduces a multi-LLM negotiation framework for sentiment analysis. The framework consists of a reasoning-infused generator to provide decision along with rationale, a explanationderiving discriminator to evaluate the credibility of the generator. The generator and the discriminator iterate until a consensus is reached. The proposed framework naturally addressed the aforementioned challenge, as we are able to take the complementary abilities of two LLMs, have them use rationale to persuade each other for correction. Experiments on a wide range of sentiment analysis benchmarks (SST-2, Movie Review, Twitter, yelp, amazon, IMDB) demonstrate the effectiveness of proposed approach: it consistently yields better performances than the ICL baseline across all benchmarks, and even superior performances to supervised baselines on the Twitter and movie review datasets. 1 Introduction Sentiment analysis (Pang and Lee, 2008; Go et al., 2009; Maas et al., 2011a; Zhang and Liu, 2012; Baccianella et al., 2010; Medhat et al., 2014; Bakshi et al., 2016; Zhang et al., 2018) aims to extract opinion polarity expressed by a chunk of text. Recent advances in large language models (LLMs) (Brown et al., 2020; Ouyang et al., 2022; Touvron et al., 2023a,b; Anil et al., 2023; Zeng et al., 2022b; OpenAI, 2023; Bai et al., 2023) open a new door for the resolving the task (Lu et al., 2021; Kojima et al., 2022; Wang et al., 2022b; Wei et al., 2022b; Wan et al., 2023; Wang et al., 2023; Sun et al., 2023b,a; Lightman et al., 2023; Li et al., 2023; Schick et al., 2023): under the paradigm of in-context learning (ICL), LLMs are able to achieve performances comparable to supervised learning strategies (Lin et al., 2021; Sun et al., 2021; Phan and Ogunbona, 2020; Dai et al., 2021) with only a small number of training examples. Existing approaches that harness LLMs for sentiment analysis usually rely on a singular LLM, and make a decision in a single round under ICL. This strategy suffers from the following disadvantage: the single-turn output generated by a single LLM might not deliver the perfect response: Just as humans sometimes need multiple attempts to get things right, it might take multiple rounds before an LLM makes the right decision. This is especially true for the task of sentiment analysis, where LLMs usually need to articulate the reasoning process to address the complex linguistic phenomenon (e.g., clause composition, irony, etc) in the input sentence. To address the this issue, in this paper, we propose a multi-LLM negotiation strategy for sentiment analysis. The core of the proposed strategy is a generator-discriminator framework, where one LLM acts as the generator (G) to produce sentiment decisions, while the other acts as a discriminator (D), tasked with evaluating the credibility of the generated output from the first LLM. The proposed method innovates on three aspects: (1) Reasoning-infused generator (G): an LLM that adheres to a structured reasoning chain, enhancing the ICL of the generator while offering the discriminator the evidence and insights to evaluate its validity; (2) Explanation deriving discriminator (D); other LLM designed to offer post-evaluation rationales for its judgments; (3) Negotiation: two LLMs act as the roles of the generator and the discriminator, and perform the negotiation until a consensus is reached. This strategy harnesses the collective abilities of the two LLMs and provide with the channel for the model to correct imperfect responses, and thus naturally resolves the issue that a single LLM cannot yield the correct decision on its first try. The contributions of this work can be summarized as follows: 1) we provide a novel perspective on how sentiment analysis can benefit from multi-LLM negotiation. 2) we introduce a Generator Discriminator Role-switching DecisionMaking framework that enables multi-LLM collaboration through iteratively generating and validating sentiment categorizations. 3) our empirical findings offer evidence for the efficacy of the proposed approach: experiments on a wide range of sentiment analysis benchmarks (SST-2, Movie Review, Twitter, yelp, amazon, IMDB) demonstrate that the proposed method consistently yields better performances than the ICL baseline across all benchmarks, and even superior performances to supervised baselines on the Twitter and movie review datasets.

Sentiment Analysis through LLM Negotiations: Abstract and Introduction 

Loss of Amazon Rainforest Resilience: Abstract and Introduction

This paper is under CC 4.0 license. available on arxiv Authors: (1) Jinrui Yang, School of Computing & Information Systems, The University of Melbourne (Email: jinruiy@student.unimelb.edu.au); (2) Timothy Baldwin, School of Computing & Information Systems, The University of Melbourne and Mohamed bin Zayed University of Artificial Intelligence, UAE (Email: (tbaldwin,trevor.cohn)@unimelb.edu.au); (3) Trevor Cohn, School of Computing & Information Systems, The University of Melbourne. Table of Links Abstract and Intro Background and Related Work Multi-EuP Experiments and Findings Language Bias Discussion Conclusion, Limitations, Ethics Statement, Acknowledgements, References, and Appendix Abstract We present Multi-EuP, a new multilingual benchmark dataset, comprising 22K multilingual documents collected from the European Parliament, spanning 24 languages. This dataset is designed to investigate fairness in a multilingual information retrieval (IR) context to analyze both language and demographic bias in a ranking context. It boasts an authentic multilingual corpus, featuring topics translated into all 24 languages, as well as cross-lingual relevance judgments. Furthermore, it offers rich demographic information associated with its documents, facilitating the study of demographic bias. We report the effectiveness of Multi-EuP for benchmarking both monolingual and multilingual IR. We also conduct a preliminary experiment on language bias caused by the choice of tokenization strategy. 1 Introduction Information retrieval (IR) classically uses a retrieval model to query a document collection and return a ranked list of documents which are predicted to be (decreasingly) relevant to the query. Retrieval models have increasingly been based on supervised learning, involving the annotation of documents with relevance scores relative to a given query, and the training of models to predict the relative association between a query and document (Karpukhin et al., 2020; Khattab and Zaharia, 2020). In parallel with these advances, the democratisation of the internet has led to a surge of individual contributors serving as information disseminators, hailing from various countries and regions, and posting in different languages. This has created possibilities for exploration of cross-lingual and multilingual text retrieval. Cross-lingual retrieval pertains to scenarios where queries are formulated in one language but documents are retrieved from another language. On the other hand, multilingual retrieval involves a query in one language but retrieval of documents across multiple languages simultaneously. An important consideration in any such work is both robustness and fairness across different combinations of languages – for instance, are results from one language consistently ranked higher than another for certain types of query. While progress towards multilingual retrieval through the release of datasets such as Mr. TYDI (Zhang et al., 2021) and mMARCO (Bonifacio et al., 2021), both are limited in that they evaluate monolingual retrieval for a range of languages, rather than true multilingual retrieval, using multiple languages simultaneously. Additionally, mMARCO was created by machine translation of MS MARCO (Nguyen et al., 2016), introducing a confounding factor of translation errors. We present a multilingual dataset based on the European Parliament debate archive with queries in 24 distinct languages, and relevance judgements also across all 24 languages. This ensures the “multilingual” nature of the dataset in terms of both query-to-document and document-to-query associations. We additionally augment each document with comprehensive metadata of the author, including gender, nationality, political affiliation, and age, for use in exploring fairness with respect to protected attributes. Our work contributes to the field in three main ways: (1) we construct and release the Multi-EuP dataset, a resource for multilingual retrieval over 24 languages, effectively capturing the multilingual nature of both queries and documents; (2) we explore language bias within the realm of multilingual retrieval, revealing that multilingual IR using BM25 indeed exhibits notable language bias; and (3) we supplement the dataset with rich author metadata to enable research on fairness and demographic bias in IR.[1] [1] The Multi-EuP dataset is available for download from https://github.com/jrnlp/Multi-EuP.

ABSTRACT

abstract.us

-- employees

Since n.d.

COMPANY RANKING

ABSTRACT

EVERGREEN INDEX #1544

Representing this company? Click here to claim and customize this page!

Abstract WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!

ABSTRACT

abstract.us

-- employees

Since n.d.

COMPANY RANKING

ABSTRACT

EVERGREEN INDEX #1544

RELATED COMPANIESRANK

STORIES ABOUT ABSTRACT

Representing this company? Click here to claim and customize this page!

Abstract WIKI

Apply to get your Tech Company News Page live on HackerNoon.com today!