SALINA, Kan. - Producers have until Friday, Nov. 17, to submit applications for the Environmental Quality Incentives Program (EQIP) and the Conservation Stewardship Program (CSP). The announcement recently was made in a news release from Kris R. Ethridge ...

Deadline announced for EQIP and CSP applications

Event Name: Designing Chips with CHIPS: West Coast Pre-Silicon Summit Date: November 3, 2023 Venue: Irwin M Jacobs Qualcomm Hall, Qualcomm Headquarters – Building N 5775 Morehouse Drive, San Diego, CA 92121 Event Website: https://chipdesign.ieeeusa.org/

Designing Chips with CHIPS: West Coast Pre-Silicon Summit to Convene Industry Leaders in San Diego

Companies introduce Siemens Industrial Copilot, a generative AI-powered assistant, designed to enhance human-machine collaboration and boost productivity.Companies will work together to build additional copilots for manufacturing,

Siemens and Microsoft partner to drive cross-industry AI adoption

Advantage Solutions Inc. (NASDAQ:ADV), a leading provider of outsourced sales and marketing services to consumer goods manufacturers and retailers, has announced the

Crisp Acquires Atlas Technology Group from Advantage Solutions to Empower CPG Brands with Better Data

CCC® Amplify Helps Repairers Launch Customized Websites in Minutes and Strengthen their Online Presence CCC Intelligent Solutions Inc. (CCC), a leading cloud platform powering the P&C insurance and co

CCC Introduces Integrated Website Builder for Collision Repairers

Establishing the true cost to develop an advanced chip is complicated, but headline numbers appear to be significantly inflated.

What Will That Chip Cost?

In this week’s tech roundup, Nothing’s CEO, Carl Pei, revealed the Power (1) charger, a multi-device concept that never reached production, bearing design similarities to the Nothing Ear (2). In the meantime,

Gov. Cooper visits Schneider Electric hub in Morrisville

The woman who filed the lawsuit on Thursday claimed she ate chips, rice, guacamole, and drank an iced tea at Miguel's Cocina on Oct. 12. A few days later, she began experiencing symptoms.

Lawsuit filed against operator of Miguel's Cocina in 4S Ranch over E. coli cases

By sharing access to expensive infrastructure and equipment at ThermoGenesis' new coworking laboratory, startups benefit from a fully outfitted lab without having to pay to open or equip it.

ThermoGenesis opens shared lab space for immunotherapy medical startups

Caspio

caspio

As the request for digitalization continues to grow at an accelerated rate, more and more CIOs are putting low-code technology at the top of their priority list given the inherent strategic benefit it has in improving operational efficiency and accelerating digital transformation projects. All without requiring too much time or money invested in the development process. According to published in June 2021, low-code software development platforms are becoming increasingly popular as businesses strive to streamline workflows and digitalize. Low-code technology is already used in nearly half of the studied enterprises (47%) and one-fifth are planning to adopt a low-code platform within the next 12 months. The majority of respondents use low-code platforms to: TechRepublic research automate workflows (17%), create new apps (15%), shorten development time (15%), automate data collecting and reporting (15%), 0% of respondents mentioned reducing the burden on developers. also expect that by 2025, 70% of new enterprise apps would be built using minimal or no-code technologies, up from below 25% in 2020, as low-code has evolved to offer immediate and very high impact in enterprises. Gartner analysts This growing popularity of low-code development platforms is unsurprising, as they put process innovation tools in the hands of problem solvers, and both non-developers and developers are seeing an increase in the sophistication of the apps they can create using low-code approaches. And as more businesses discover what is possible with low-code development, the adoption of these tools will not only continue to grow but it will also become more structured. So, let’s look at 3 different low-code development approaches the LCP Market has contoured and the main things every CIO should know about them when it comes to adopting a low-code platform to drive their digital transformation strategy. IT centric low-code development platforms Designed for skilled developers to make their jobs easier and faster. They were a big hit once developers figured them out. They have built-in development, testing, production, licensing, and data governance. The tools are built on well-known development principles and can be used to construct various apps. But this is also their downfall, at least from a business point of view.  Because line-of-business stakeholders are unfamiliar with the tools, their involvement in the development process is generally limited to translating requirements. Resulting in a single-threaded delivery model where developers need to comprehend business requirements, generate a technical specification, and then go through blueprinting budgeting development user acceptance testing (UAT), go live, and then modify requests resuming the process. Businesses also struggled to grasp what was going on "behind the scenes," leaving them unable to assess how the application "behaved" in connection to their original business requirements. This is because they can't see what they're dealing with and in the end this slows down the development cycle. In this category I identified the following low-code platforms: Pega Agilepoint Unqork Caspio Aura Quantic Low-code development platforms for citizen developers Low-code platforms for citizen developers can be used to produce low to mid-complexity applications for 12-24 users inside an organization to achieve a certain goal, ideal for start-ups. Anyone who understands the business subject matter can build and change it fast. However, because these solutions are designed for small workgroups, they often lacked the scalability and interoperability required for big user communities, complicated technological environments, and enterprise deployments. They either hit severe limits or required extensive technical intervention from a range of technical disciplines to try and extend what wasn't inherent to the tool itself. Low-code platforms posed new issues for enterprises and IT professionals entrusted with assuring governance over application deployment and business utility. Many strengths we see in the IT centric low-code platforms, such as numerous staging environments, access rights, data for technical and business audits, important security features, license control, and on-premises deployment choices are frequently absent from their core set of functionalities. It’s no surprise that many IT professionals are still hesitant of implementing these toolsets. In this category I identified the following low-code platforms: PowerApps Betty Blocks Creatio Kissflow 3. Low-code development platforms for collaboration Low-code platforms developed for collaboration combine the best of both types of platforms listed above, allowing enterprises to build simple to complex apps that expand from dozens to thousands of users and transactions. They include all the controls and functionalities required for enterprise governance while also fostering a more collaborative environment in which subject matter experts can contribute to the development of a software application, understand what is being built as it is being built, and contribute in a way that is intuitive to their specific expertise. With low-code platforms built for collaboration, it is possible to have the business stakeholder work on aspects of the build that align with their expertise (automation flows, data models, UI design), while leaving the technical aspects (integrations, access control, staging environments, technical audits) to the IT subject matter experts. This eliminates business and IT silos and enable faster time to value for both parties. Enterprises can have technical and business teams working together on a same platform for a common goal. It's business-grade, has built-in governance, is intuitive for all users, and can scale quickly from departmental to corporate apps and it meets the needs of today's time-pressed, multi-faceted, ever-growing businesses. In this category I identified the following low-code platforms: Aurachain Mendix In 2022, digital transformation will reach new heights, but are we about to make the same software failure mistakes all over again? The past two years have been a wake-up call for firms without a clear digital transformation strategy.  The epidemic has emphasized the necessity for businesses to produce digital apps swiftly and easily adaptive to shifting needs. Businesses needed to change quickly and were looking for more cost-effective and agile solutions, but they must be careful when choosing their low-code approach. As Jonathan Wiener writes in TechRegister in his article ( SEE: Are we about to make the same software failure mistakes all over again?) “The risk ahead lies in the fact that many of the tools being used in low-code development environments when they divide rather than unite the various specialists in the process. They create ‘fiefdoms’ between tech and business rather than bringing the two sides together into one ‘development conversation.’ On the one hand we hear about developers wanting to retain control of their knowledge (and rightly so in many ways), while trends such as the citizen developer suggest it is OK for non-techies to build their own systems in isolation. It is OK at a basic workflow level, but not when you’re addressing mission-critical problems and enterprise application use. ” Collaboration should extend beyond teams working together to create a new digital app. It should exist at all stages of the development cycle, from process design to documentation, API design, publication, and app performance analysis. It is a collaborative effort between all essential parties, aided by technology. When work is 'democratized' among all relevant business and technical subject matter experts, a team can build together within the platform rather than working in isolation on its composite elements. To do so, a platform that is both user-friendly and enterprise-grade – with built-in governance, mission-critical change support, and cloud-native scalability – is required. The good news is that these types of platforms are now available on the LCP Market in the form of collaboration platforms - they are business-grade, have built-in governance, are intuitive for all users, and can scale quickly from departmental to corporate apps, and they meet the needs of today's time-pressed, multi-faceted, ever-growing businesses.

3 Types of Low-Code Development Platform That Every CIO Should Know About

Building a product is often the most challenging part of building a startup. That’s why lots of founders want to rid themselves of it by outsourcing this to other companies. And there’s nothing wrong with that if you’ve got the time and the money. Dev shops are great in terms of experience having built hundreds of similar or even identical projects from scratch, they know what you’re coming for. It also means that Is there this kind of technology? The short answer is no. The better answer is we’re building it right this second. Let’s first dive into the available options there are. This is for complicated projects when you’ve really got something robust going on and can’t find other sustainable solutions in the long run. But if your thing pretty much copies another idea (and there’s nothing wrong with that, the best ideas are stolen) then it’s not necessary at all. So, if you’re pitching a “next Uber/Instagram/Airbnb/…” or anything else you’re going to replicate to your investors, don’t ask for 250$k for development, Believe it or not, developers can write less code and get the same results. Why is your dev shop not interested in that? You guessed it, they can’t charge you as much. It’s as simple as that. If you’re a tech founder it’s the perfect one for you, though. Learn how a low-code platform works, or two. Low-code platforms: , , , No-code is the easiest and the fastest alternative to them all. Or, at least, that’s what the belief is. Sure, you can use drag-and-drop solutions for your early versions/prototypes/MVP but if you need something a little more sophisticated then the learning curve is fairly steep. Have you noticed that there are actually not so many Webflow developers? I mean, good Webflow developers. While no-code is sustainable for the idea validation stage and maybe a little bit further along, it won’t carry you through the rapid growth stage. And you’ll eventually have to scratch all your beautiful work and go to custom code. No-code platforms: , , , If you’re selling something online, there’s Shopify, Gumroad, WooCommerce, and even Wix. All you need is a few hours away from Netflix and you have an online business. And these are great at what they do. They are limited but you get enough features to get you going, you’re supported, safe, and pretty much self-explanatory. Platforms: , , , Or white label, if you will. Building a taxi app? There’s probably something on the market, buy it, stick your brand name on it and you’re good to go. Will it give you the freedom to grow and add innovative ideas? Probably not. Will it still be fine if your competitive advantage is exceptional service or a unique business model? Yes. with the right technology at hand, they’d be able to reuse parts or even projects almost to the “T” for other projects. 1. Custom code you can do it way cheaper. 2. Low-code It’ll make you way more efficient and give you all the control over your project. Caspio Retool Creatio OutSystems 3. No-code Because you’ve got to learn a lot. A LOT. And then once you’ve learned it all, you hit the ceiling. XPODA Bubble Lansa Airtable 4. Platforms You just need to play by their rules. Shopify Gumroad WooCommerce Wix 5. Out of the box Out of the box: , Twilio Cyclr 6. Mars Choosing among the above options might not be an easy task. Especially if you are a first-time startup maker. But fear not! There is a tool that combines the benefits of all the previous items – Mars. Mars is an open-source platform for making web and mobile apps. It gives you flexibility on demand. You get access to the custom code only when you want to implement a specific feature. As for the regular functionalities, such as login page, payments, cart, dropdown menu, footer, photo galleries, and profile page, you can plug them into your startup with little-to-zero coding knowledge. This makes Mars the perfect tool for making MVP: you are building fast most of the time, and you go to custom code only to make unique changes. You’re saving time but you’re not trading anything away for that. However, there is a disadvantage to using the platform. It is a new technology and it lacks good documentation. The learning curve may be a little steep, especially for a complete beginner. But once you get used to Mars, you will be able to build complex applications such as Airbnb, UBER, and Instagram amazingly easily. The tool: . Mars

Low-code vs No-code: Selecting the Best Stack For Your Startup

How I Hacked and Turned My CASIO F-91W Into a Contactless Payment Device

কিভাবে আমি আমার CASIO F-91W কে হ্যাক করে একটি কন্টাক্টলেস পেমেন্ট ডিভাইসে পরিণত করেছি

Wie ich meinen CASIO F-91W gehackt und in ein kontaktloses Zahlungsgerät verwandelt habe

Cómo hackeé y convertí mi CASIO F-91W en un dispositivo de pago sin contacto

Comment j'ai piraté et transformé mon CASIO F-91W en appareil de paiement sans contact

मैंने कैसे हैक किया और अपने CASIO F-91W को संपर्क रहित भुगतान डिवाइस में बदल दिया

CASIO F-91W をハッキングして非接触型決済デバイスに変えた方法

CASIO F-91W를 해킹하여 비접촉 결제 장치로 바꾸는 방법

Como hackeei e transformei meu CASIO F-91W em um dispositivo de pagamento sem contato

CASIO F-91W'mi Nasıl Hackledim ve Temassız Ödeme Cihazına Dönüştürdüm

Cách tôi hack và biến chiếc CASIO F-91W của mình thành một thiết bị thanh toán không cần tiếp xúc

我如何将 CASIO F-91W 改造为非接触式支付设备

If you’re a frequent user of ChatGPT, you know the tendency it has to wander off into what is known as hallucinations. A great collection of statistically correct words that have no basis in reality. A few months ago, a prompt about using Apache Cassandra for large language models (LLMs) and LangChain resulted in a curious response. ChatGPT reported that not only was Cassandra a good tool choice when creating LLMs, OpenAI used Cassandra with an MIT-licensed Python library they called CassIO. Into the rabbit hole we went, and through more prompting, ChatGPT described many details about how CassIO was used. It even included some sample code and a website. Subsequent research found no evidence of CassIO outside of ChatGPT responses, but the seed was sown. If this library didn’t exist, it needed to, and we started work on it shortly after. Best hallucination ever. Will the real CassIO please stand up? What was this great idea that ChatGPT (and, by association, OpenAI) inspired? A great Python library enables developers to do more with less. and combined forces in developing to make the integration of Cassandra with generative artificial intelligence and other machine learning workloads seamless. Its principal purpose is to abstract the process of accessing the Cassandra database, including its capabilities, offering a set of ready-to-use tools that minimize the need for additional code. As a result, developers can focus on designing and implementing their AI systems, knowing that CassIO has taken care of the underlying database complexities. The result is access to a proven database for affordable scale and low latency. The essence of CassIO is all about facilitating and simplifying the implementation process. DataStax Anant CassIO vector search CassIO's strength lies in its agnosticism toward specific AI frameworks. It doesn't concern itself with the specific implementation details of interfaces like , , , or various other generative AI toolkits. Instead, it provides a set of "thin adapters" that conform to the framework's interfaces while using the capabilities of CassIO. This enables CassIO to bridge the gap between your AI application and the database, thus enabling the application to leverage the power of Cassandra without getting entangled in its details. LangChain LlamaIndex Microsoft Semantic Kernel Integration with LangChain automates the majority of management tasks and interactions with LLMs. It provides support for memory, vector-based similarity search, advanced prompt templating abstraction, and a wealth of other features. CassIO integrates seamlessly with LangChain, extending Cassandra-specific tools to streamline tasks such as: LangChain A memory module for LLMs that uses Cassandra for storage, which can remember recent exchanges in a chat interaction, or even keep a summary of the entire past conversation. A feature to cache LLM responses on Cassandra, thereby saving on latency and tokens where possible.Automatic injection of data from Cassandra into a prompt or within a longer LLM conversation. Support for "partialing" of prompts, leaving some input unspecified for future supply. Automatic injection of data from a feature store (potentially ) into a prompt. Feast backed by Cassandra These components work together to streamline the process of incorporating data into prompts and ensure smooth interaction between the LLM and the database. Integration with vector search The inclusion of capabilities in Cassandra and DataStax Astra DB recently (read about the news ) has integrated a key feature into an already popular database for transactional data. Cassandra's reputation for high scale means that you have a single place to store and process data without moving data around in costly operations. The addition of vector search has opened doors to a suite of "semantically aware" tooling made available in CassIO, such as: vector search here A cache of LLM responses that are not dependent on the exact phrasing of a query. A "semantic index" that can store a knowledge base and retrieve relevant parts to construct the best answer to a given question. This tool can be adapted to suit many specific needs and can be configured to retrieve diverse information to maximize the actual information flowing into the answer. A "semantic memory" element for LLM chat interactions, which can retrieve relevant past exchanges even if they occurred in the distant past. The combination of CassIO and LangChain continues to expand and refine these capabilities over time to meet the ever-evolving needs of LLM management. The current state-of-the-art is in chaining prompts to get more accurate responses from LLMs. In a recent paper describing a technique called , the role of vector search plays a critical role in persistence from one prompt to the next. As these ideas move from academia to production, Cassandra will serve as an important part of the implementation. tree-of-thought Next prompt: What's ahead for CassIO As an evolving tool, CassIO is growing rapidly, with new developments and updates frequently added. At the time of writing, CassIO supports LangChain, with LlamaIndex coming soon. The long-term goal of this project is to support high-scale memory for autonomous AI agents such as the project. Agents with LLMs are an exciting development that will have an incredible impact on many industries with complex task handling. These agents will need to keep track of many aspects of data and interactions, and Cassandra is the right database for the job. Reliable and performant. JARVIS An upcoming boot camp, “ ,” will offer developers a chance to work hands-on with the library to build a chat bot. Look for more activities like this coming to a city near you! We encourage users exploring to , participate in the and help us improve this rapidly materializing hallucination. NoCode, Data & AI: LLM Bootcamp with Cassandra CassIO file issues forums Who knows how history will judge this moment? Was it a leak of internal information from OpenAI? Or, thinking a bit more darkly, is this the first step of AI to get humans to do its bidding?  Either way, developers now have a simple-to-use library to tap into the near-infinite scale of Cassandra when striking off into the world of generative AI. ChatGPT has given us a gift, so what are you going to build with this? I’m going to be diving into vector search in an upcoming webinar (register !), and if you just want to get in and start working today, has some great tutorials. here DataStax Astra By Patrick McFadin, DataStax Patrick McFadin is the co-author of the O’Reilly book 'Managing Cloud Native Data on Kubernetes.' He currently works at DataStax in developer relations and as a contributor to the Apache Cassandra project. Patrick has worked as chief evangelist for Apache Cassandra (he’s also a newly minted Cassandra committer!) and as a consultant for DataStax, where he had a great time building some of the largest deployments in production.

CassIO: The Best Library for Generative AI, Inspired by OpenAI

CassIO: জেনারেটিভ AI এর জন্য সেরা লাইব্রেরি, OpenAI দ্বারা অনুপ্রাণিত

CassIO: Die beste Bibliothek für generative KI, inspiriert von OpenAI

CassIO: la mejor biblioteca para IA generativa, inspirada en OpenAI

CassIO : la meilleure bibliothèque pour l'IA générative, inspirée par OpenAI

CassIO: जनरेटिव AI के लिए सर्वश्रेष्ठ लाइब्रेरी, OpenAI से प्रेरित

CassIO: OpenAI からインスピレーションを得た、生成 AI に最適なライブラリ

CassIO: OpenAI에서 영감을 받은 생성적 AI를 위한 최고의 라이브러리

CassIO: a melhor biblioteca para IA generativa, inspirada na OpenAI

CassIO: OpenAI'den Esinlenen, Üretken Yapay Zeka için En İyi Kütüphane

CassIO: Thư viện tốt nhất dành cho AI sáng tạo, lấy cảm hứng từ OpenAI

CassIO：受 OpenAI 启发的最佳生成 AI 库

The world of cryptocurrency should be doing everything it can to attract institutional investors. With $130 trillion of assets under management worldwide, institutional investors could have a huge positive impact if they moved even a tiny fraction of those funds into crypto, whose market cap remains under $300 billion. — Yet diving into the world of crypto can be incredibly daunting. On top of understanding the workings of the underlying technology, newcomers wishing to trade need to adapt to an incredibly volatile market and learn the ropes across a handful of exchanges. Many struggle to reconcile the fragmentation of these platforms there’s no one-stop shop, meaning that traders must interact with a plethora of exchanges to acquire the crypto assets they want. , it’s clear that investors . We’ve seen institutional players engaging — managing portfolios through Excel spreadsheets, writing Python scripts to execute trades, and using multiple web interfaces. There are, of course, a couple of exceptions to this rule, but for the vast majority, this rudimentary setup is fast becoming the norm. On the institutional front have been struggling in practices almost unheard of at that level in traditional markets through a comprehensive solution for asset management. Institutions are able to access a fully-fledged platform worthy of traditional markets. Complete with a suite of tools such as OEMS, PMS and RMS, as well as a number of sophisticated trading algorithms, the Caspian solution integrates with the backend of numerous exchanges so that users can access them from a single intuitive interface. Caspian aims to set the bar somewhat higher Tools like Caspian should be of great interest, not only to these entities, but to crypto evangelists and speculators at large — the long-term viability of the space is dependent on increasing investments, but many funds will not enter until more streamlined trading platforms can be offered. The introduction of products such as Caspian will therefore bridge an important gap between incumbent institutions, and the rapidly-evolving crypto space. Institutional players will be seminal in pushing the cryptocurrency market cap to the next level. Supporting as many exchanges as possible has been our main focus for a number of reasons — the first being arbitrage: there’s a lot of money to be made in leveraging prices discrepancies across exchanges. Another motivation for this is a lack of liquidity on a single exchange when executing high-volume orders. Tying the different platforms together in the same interface makes it much easier to fill orders, by drawing on not one, but several pools. This has the added benefit of spreading risk — many exchanges are unregulated, so Caspian mitigates counterparty risk. With decades of experience in banking, creating institutional-grade trading software and operating hedge funds, the Caspian team brings together a wealth of talent from around the world. It currently has over 40 engineers working on understanding and interacting with the various different APIs offered by exchanges. It’s an arduous task, due to the lack of widely-accepted standards in the industry, as well as all the different technology and language used for their API connections, but it’s also a necessary one to deliver the sophisticated platform that Caspian aims to roll out, and to bring new investors into crypto. About The Author is CTO & Co-founder of Caspian. He has served as Managing Director at Tora for the past 12 years. Prior to that Gerrit managed the development of high volume trading applications and analytical tools for various hedge funds and bulge bracket investment banks. Gerrit holds a B.S. in Computer Science from Stanford University Gerrit Van Wingerden About Caspian is a full-stack cryptoasset management platform tying together the biggest crypto exchanges in a single interface, so as to facilitate investments in crypto instruments for newcomers and veterans alike. The joint venture between heavyweights Tora and Kenetic brings to the table a wealth of experience in asset management, accumulated over decades of building and operating trading platforms and technologies. Caspian

How Caspian Is Interfacing With 40+ Crypto Exchanges to Facilitate Institutional Trading

Emotion As A Service: AI’s Next Market Is Your Heart

العاطفة كخدمة: السوق القادمة للذكاء الاصطناعي هي قلبك

Емоцията като услуга: Следващият пазар на AI е вашето сърце

পরিষেবা হিসাবে আবেগ: AI এর পরবর্তী বাজার আপনার হৃদয়

La emoción como servicio: el próximo mercado de la IA es tu corazón

Emosyon kòm yon sèvis: pwochen mache AI a se kè ou

感情をサービスとして: AIの次の市場はあなたの心

អារម្មណ៍ជាសេវាកម្ម៖ ទីផ្សារបន្ទាប់របស់ AI គឺជាបេះដូងរបស់អ្នក។

Кызмат катары эмоция: AIнын кийинки рыногу - бул сиздин жүрөгүңүз

Emocje jako usługa: następnym rynkiem dla sztucznej inteligencji jest Twoje serce

Emotion seSevhisi: AI Inotevera Musika Mwoyo Wako

ስምዒት ከም ኣገልግሎት፡ ቀጻሊ ዕዳጋ AI ልብኻ’ዩ።

Cảm xúc như một dịch vụ: Thị trường tiếp theo của AI là trái tim của bạn

Using knowledge graphs to improve the results of retrieval-augmented generation (RAG) applications has become a hot topic. Most examples demonstrate how to build a knowledge graph using a relatively small number of documents. This might be because the typical approach – extracting fine-grained, entity-centric information – just doesn’t scale. Running each document through a model to extract the entities (nodes) and relationships (edges) takes too long (and costs too much) to run on large datasets. We’ve argued that content-centric knowledge graphs – a vector-store allowing links between chunks – are an easier to use and more efficient approach. Here, we put that to the test. We load a subset of the Wikipedia articles from the 2wikimultihop dataset using both techniques and discuss what this means for loading the entire dataset. We demonstrate the results of some questions over the loaded data. We’ll also load the entire dataset – nearly 6 million documents – into a content-centric GraphVectorStore. Entity-centric: LLMGraphTransformer Loading documents into an entity-centric graph store like Neo4j was done using LangChain’s LLMGraphTransformer. The code is based on LangChain's "How to construct knowledge graphs." from langchain_core.documents import Document
from langchain_experimental.graph_transformers import LLMGraphTransformer
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(temperature=0, model_name="gpt-4-turbo")
llm_transformer = LLMGraphTransformer(llm=llm)

from time import perf_counter
start = perf_counter()

documents_to_load = [Document(page_content=line) for line in lines_to_load]
graph_documents = llm_transformer.convert_to_graph_documents(documents_to_load)

end = perf_counter()
print(f"Loaded (but NOT written) {NUM_LINES_TO_LOAD} in {end - start:0.2f}s") Content-centric: GraphVectorStore Loading the data into GraphVectorStore is roughly the same as loading it into a vector store. The only addition is that we compute metadata indicating how each page links to other pages. import json
from langchain_core.graph_vectorstores.links import METADATA_LINKS_KEY, Link

def parse_document(line: str) -> Document:
    para = json.loads(line)

    id = para["id"]
    links = {
        Link.outgoing(kind="href", tag=id)
        for m in para["mentions"]
        if m["ref_ids"] is not None
        for id in m["ref_ids"]
    }
    links.add(Link.incoming(kind="href", tag=id))
    return Document(
        id = id,
        page_content = " ".join(para["sentences"]),
        metadata = {
            "content_id": para["id"],
            METADATA_LINKS_KEY: list(links)
        },
    ) This is also a good example of how you can add your own links between nodes. from langchain_openai import OpenAIEmbeddings
from langchain_community.graph_vectorstores.cassandra import CassandraGraphVectorStore
import cassio

cassio.init(auto=True)
TABLE_NAME = "wiki_load"
store = CassandraGraphVectorStore(
embedding = OpenAIEmbeddings(),
node_table=TABLE_NAME,
insert_timeout = 1000.0,
)

from time import perf_counter
start = perf_counter()

from datasets.wikimultihop.load import parse_document
kg_documents = [parse_document(line) for line in lines_to_load]
store.add_documents(kg_documents)

end = perf_counter()
print(f"Loaded (and written) {NUM_LINES_TO_LOAD} in {end - start:0.2f}s") Loading benchmarks Running at 100 rows, the entity-centric approach using GPT-4o took 405.93s to extract the GraphDocuments and 10.99s to write them to Neo4j, while the content-centric approach took 1.43s. Extrapolating, it would take 41 weeks to load all 5,989,847 pages using the entity-centric approach and about 24 hours using the content-centric approach. But thanks to parallelism, the content-centric approach runs in only 2.5 hours! Assuming the same parallelism benefits, it would still take over four weeks to load everything using the entity-centric approach. I didn’t try it since the estimated cost would be $58,700 — assuming everything worked the first time! Bottom-line: the entity-centric approach of extracting knowledge graphs from content using an LLM was both time and cost prohibitive at scale. On the other hand, using GraphVectorStore was fast and inexpensive. Example answers In this section, a few questions, drawn from the subset of loaded documents, are asked to address the quality of answers. Entity-centric used 7324 prompt tokens and cost $0.03 to produce basically useless answers, while content-centric used 450 prompt tokens and cost $0.002 to produce concise answers directly answering the questions. It may be surprising that the fine-grained Neo4j graph returns useless answers. Looking at the logging from the chain, we see some of why this happens: > Entering new GraphCypherQAChain chain...

Generated Cypher:

cypher

MATCH (a:Album {id: 'The Circle'})-[:RELEASED_BY]->(r:Record_label)

RETURN a.id, r.id


Full Context:

[{'a.id': 'The Circle', 'r.id': 'Restless'}]


> Finished chain.

{'query': "When was 'The Circle' released?", 'result': "I don't know the answer."} So, the fine-grained schema only returned information about the record label. It makes sense that the LLM wasn’t able to answer the question based on the retrieved information. Conclusion Extracting fine-grained, entity-specific knowledge graphs is time- and cost-prohibitive at scale. When asked questions about the subset of data that was loaded, the additional granularity (and extra cost of loading the fine-grained graph) returned more tokens to include the prompt but generated useless answers! GraphVectorStore takes a coarse-grained, content-centric approach that makes it fast and easy to build a knowledge graph. You can start with your existing code for populating a VectorStore using LangChain and add links (edges) between chunks to improve the retrieval process. Graph RAG is a useful tool for enabling generative AI RAG applications to retrieve more deeply relevant contexts. But using a fine-grained, entity-centric approach does not scale to production needs. If you're looking to add knowledge graph capabilities to your RAG application, try GraphVectorStore. By Ben Chambers, DataStax Using knowledge graphs to improve the results of retrieval-augmented generation (RAG) applications has become a hot topic. Most examples demonstrate how to build a knowledge graph using a relatively small number of documents. This might be because the typical approach – extracting fine-grained, entity-centric information – just doesn’t scale. Running each document through a model to extract the entities (nodes) and relationships (edges) takes too long (and costs too much) to run on large datasets. We’ve argued that content-centric knowledge graphs – a vector-store allowing links between chunks – are an easier to use and more efficient approach. Here, we put that to the test. We load a subset of the Wikipedia articles from the 2wikimultihop dataset using both techniques and discuss what this means for loading the entire dataset. We demonstrate the results of some questions over the loaded data. We’ll also load the entire dataset – nearly 6 million documents – into a content-centric GraphVectorStore . content-centric knowledge graphs content-centric knowledge graphs 2wikimultihop 2wikimultihop GraphVectorStore GraphVectorStore Entity-centric: LLMGraphTransformer Loading documents into an entity-centric graph store like Neo4j was done using LangChain’s LLMGraphTransformer. The code is based on LangChain's "How to construct knowledge graphs." "How to construct knowledge graphs." "How to construct knowledge graphs." from langchain_core.documents import Document
from langchain_experimental.graph_transformers import LLMGraphTransformer
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(temperature=0, model_name="gpt-4-turbo")
llm_transformer = LLMGraphTransformer(llm=llm)

from time import perf_counter
start = perf_counter()

documents_to_load = [Document(page_content=line) for line in lines_to_load]
graph_documents = llm_transformer.convert_to_graph_documents(documents_to_load)

end = perf_counter()
print(f"Loaded (but NOT written) {NUM_LINES_TO_LOAD} in {end - start:0.2f}s") from langchain_core.documents import Document
from langchain_experimental.graph_transformers import LLMGraphTransformer
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(temperature=0, model_name="gpt-4-turbo")
llm_transformer = LLMGraphTransformer(llm=llm)

from time import perf_counter
start = perf_counter()

documents_to_load = [Document(page_content=line) for line in lines_to_load]
graph_documents = llm_transformer.convert_to_graph_documents(documents_to_load)

end = perf_counter()
print(f"Loaded (but NOT written) {NUM_LINES_TO_LOAD} in {end - start:0.2f}s") Content-centric: GraphVectorStore Loading the data into GraphVectorStore is roughly the same as loading it into a vector store. The only addition is that we compute metadata indicating how each page links to other pages. import json
from langchain_core.graph_vectorstores.links import METADATA_LINKS_KEY, Link

def parse_document(line: str) -> Document:
    para = json.loads(line)

    id = para["id"]
    links = {
        Link.outgoing(kind="href", tag=id)
        for m in para["mentions"]
        if m["ref_ids"] is not None
        for id in m["ref_ids"]
    }
    links.add(Link.incoming(kind="href", tag=id))
    return Document(
        id = id,
        page_content = " ".join(para["sentences"]),
        metadata = {
            "content_id": para["id"],
            METADATA_LINKS_KEY: list(links)
        },
    ) import json
from langchain_core.graph_vectorstores.links import METADATA_LINKS_KEY, Link

def parse_document(line: str) -> Document:
    para = json.loads(line)

    id = para["id"]
    links = {
        Link.outgoing(kind="href", tag=id)
        for m in para["mentions"]
        if m["ref_ids"] is not None
        for id in m["ref_ids"]
    }
    links.add(Link.incoming(kind="href", tag=id))
    return Document(
        id = id,
        page_content = " ".join(para["sentences"]),
        metadata = {
            "content_id": para["id"],
            METADATA_LINKS_KEY: list(links)
        },
    ) This is also a good example of how you can add your own links between nodes. from langchain_openai import OpenAIEmbeddings
from langchain_community.graph_vectorstores.cassandra import CassandraGraphVectorStore
import cassio

cassio.init(auto=True)
TABLE_NAME = "wiki_load"
store = CassandraGraphVectorStore(
embedding = OpenAIEmbeddings(),
node_table=TABLE_NAME,
insert_timeout = 1000.0,
)

from time import perf_counter
start = perf_counter()

from datasets.wikimultihop.load import parse_document
kg_documents = [parse_document(line) for line in lines_to_load]
store.add_documents(kg_documents)

end = perf_counter()
print(f"Loaded (and written) {NUM_LINES_TO_LOAD} in {end - start:0.2f}s") from langchain_openai import OpenAIEmbeddings
from langchain_community.graph_vectorstores.cassandra import CassandraGraphVectorStore
import cassio

cassio.init(auto=True)
TABLE_NAME = "wiki_load"
store = CassandraGraphVectorStore(
embedding = OpenAIEmbeddings(),
node_table=TABLE_NAME,
insert_timeout = 1000.0,
)

from time import perf_counter
start = perf_counter()

from datasets.wikimultihop.load import parse_document
kg_documents = [parse_document(line) for line in lines_to_load]
store.add_documents(kg_documents)

end = perf_counter()
print(f"Loaded (and written) {NUM_LINES_TO_LOAD} in {end - start:0.2f}s") Loading benchmarks Running at 100 rows, the entity-centric approach using GPT-4o took 405.93s to extract the GraphDocuments and 10.99s to write them to Neo4j, while the content-centric approach took 1.43s. Extrapolating, it would take 41 weeks to load all 5,989,847 pages using the entity-centric approach and about 24 hours using the content-centric approach. But thanks to parallelism, the content-centric approach runs in only 2.5 hours! Assuming the same parallelism benefits, it would still take over four weeks to load everything using the entity-centric approach. I didn’t try it since the estimated cost would be $58,700 — assuming everything worked the first time! Bottom-line: the entity-centric approach of extracting knowledge graphs from content using an LLM was both time and cost prohibitive at scale. On the other hand, using GraphVectorStore was fast and inexpensive. Example answers Example answers In this section, a few questions, drawn from the subset of loaded documents, are asked to address the quality of answers. Entity-centric used 7324 prompt tokens and cost $0.03 to produce basically useless answers, while content-centric used 450 prompt tokens and cost $0.002 to produce concise answers directly answering the questions. It may be surprising that the fine-grained Neo4j graph returns useless answers. Looking at the logging from the chain, we see some of why this happens: > Entering new GraphCypherQAChain chain...

Generated Cypher:

cypher

MATCH (a:Album {id: 'The Circle'})-[:RELEASED_BY]->(r:Record_label)

RETURN a.id, r.id


Full Context:

[{'a.id': 'The Circle', 'r.id': 'Restless'}]


> Finished chain.

{'query': "When was 'The Circle' released?", 'result': "I don't know the answer."} > Entering new GraphCypherQAChain chain...

Generated Cypher:

cypher

MATCH (a:Album {id: 'The Circle'})-[:RELEASED_BY]->(r:Record_label)

RETURN a.id, r.id


Full Context:

[{'a.id': 'The Circle', 'r.id': 'Restless'}]


> Finished chain.

{'query': "When was 'The Circle' released?", 'result': "I don't know the answer."} So, the fine-grained schema only returned information about the record label. It makes sense that the LLM wasn’t able to answer the question based on the retrieved information. Conclusion Extracting fine-grained, entity-specific knowledge graphs is time- and cost-prohibitive at scale. When asked questions about the subset of data that was loaded, the additional granularity (and extra cost of loading the fine-grained graph) returned more tokens to include the prompt but generated useless answers! GraphVectorStore takes a coarse-grained, content-centric approach that makes it fast and easy to build a knowledge graph. You can start with your existing code for populating a VectorStore using LangChain and add links (edges) between chunks to improve the retrieval process. Graph RAG is a useful tool for enabling generative AI RAG applications to retrieve more deeply relevant contexts. But using a fine-grained, entity-centric approach does not scale to production needs. If you're looking to add knowledge graph capabilities to your RAG application, try GraphVectorStore . GraphVectorStore GraphVectorStore By Ben Chambers , DataStax Ben Chambers

How to Save $70K Building a Knowledge Graph for RAG on 6M Wikipedia Pages

Hoe om $70 000 te bespaar Die bou van 'n kennisgrafiek vir JOOL op 6M Wikipedia-bladsye

كيفية توفير 70 ألف دولار من خلال إنشاء رسم بياني للمعرفة لـ RAG على 6 ملايين صفحة ويكيبيديا

Com estalviar 70.000 dòlars Construint un gràfic de coneixement per a RAG a 6 milions de pàgines de Viquipèdia

Cómo ahorrar $70 000 creando un gráfico de conocimiento para RAG en 6 millones de páginas de Wikipedia

600 万の Wikipedia ページで RAG のナレッジ グラフを構築して 7 万ドルを節約する方法

របៀបសន្សំ $70K បង្កើត Knowledge Graph សម្រាប់ RAG នៅលើទំព័រ Wikipedia 6M

6 сая Wikipedia хуудсан дээр RAG-д зориулсан мэдлэгийн график бүтээхэд 70 мянган долларыг хэрхэн хэмнэх вэ

6M विकिपिडिया पृष्ठहरूमा RAG को लागि ज्ञान ग्राफ निर्माण गरेर $ 70K कसरी बचत गर्ने

په 6M ويکيپېډيا پاڼو کې د RAG لپاره د پوهې ګراف جوړول $70K څنګه خوندي کول

Cum să economisiți 70.000 USD Construirea unui grafic de cunoștințe pentru RAG pe 6 milioane de pagini Wikipedia

ከመይ ጌርና $70K ንቑጥብ ንRAG ዝኸውን ናይ ፍልጠት ግራፍ ኣብ 6M ዊኪፒዲያ ገጻት ንምህናጽ

Indlela yokugcina i-$ 70K yokwakha iGrafu yoLwazi lwe-RAG kwi-6M yamaphepha e-Wikipedia

Supply Chain - The Achilles Heel of Cybersecurity

Recollections of My Life: Part 3

Physical Properties of the Eclipsing Binary KIC 9851944: Abstract and Intro

Stories, News & Mentions about CASPIO

Caspio’s Stories on HackerNoon

Caspio's News & Mentions