Center for AI and Digital Policy’s Post

Governments are increasingly turning to automated solutions. This calls for application of algorithmic transparency principles, standards, and rules by public bodies in their use of algorithms. This year, led by Juan David Gutiérrez Rodríguez and Alison Gillwald, the GPAI Data Governance and Responsible AI Working Groups have joined forces to guide public bodies in their compliance to these principles. The first step of their work on this project is summarized in this state of the art research that studies the algorithmic transparency concept and components, in addition to mapping and assessing existing transparency instruments. 📃Read the full report here: https://bit.ly/ATPS2024 📢Don’t forget to register for the webinar on this very topic, on July 23rd: https://lnkd.in/ecXj6u69 Bertrand Monthubert Shameek Kundu Avik Sarkar, Przemyslaw Biecek Seydina M. Ndiaye Juliana Sakai Thierry Warin Rosanna Fanni Irakli Khodeli #ArtificialIntelligence #ResponsibleAI #AlgorithmicTransparency

Unlocking LLM Potential with GraphRAG Microsoft Research introduces GraphRAG, a breakthrough technique combining LLMs and knowledge graphs to enhance data discovery on private datasets. By leveraging graph machine learning, GraphRAG improves upon traditional RAG methods, offering superior results in document analysis and answering complex queries. This innovation addresses challenges in connecting disparate information and understanding large data collections, making it a powerful tool for enterprises dealing with proprietary data. Key topics include: · Integration of LLMs and knowledge graphs · Enhanced data discovery techniques · Advanced graph machine learning applications · Explore the full potential of GraphRAG on Microsoft's blog. https://lnkd.in/dUEhkGsv #AI #MachineLearning #DataScience #MicrosoftResearch #GraphRAG

🌐 Revolutionizing Data Analysis with GraphRAG: Microsoft's New Frontier 🌐 In a groundbreaking development, Microsoft Research has introduced GraphRAG, a powerful new tool designed to transform how we interact with and analyze complex information sets. GraphRAG leverages knowledge graphs generated by large language models (LLMs) to enhance the retrieval-augmented generation process, marking a significant leap in handling narrative private data. 🔍 Why GraphRAG Matters GraphRAG represents an innovative approach to querying and analyzing complex datasets, particularly those that are unstructured and vast—common in fields like market research, security, and academic studies. Traditional methods often struggle to connect disparate pieces of information to deliver cohesive insights. GraphRAG addresses this by creating a dynamic knowledge graph that organizes data into interconnected entities and relationships, allowing for more sophisticated query responses and insights generation. 📊Transformative Capabilities: 🔹 Enhanced Question-Answering Performance: GraphRAG dramatically improves the ability of systems to handle complex questions that traditional models cannot, by synthesizing insights from broad data sets. 🔹 Provenance and Trust: Each response generated by GraphRAG is traceable back to its data sources, ensuring transparency and reliability in the insights provided. 🔹 Broad Industry Applications: From financial services deciphering market trends to academic researchers analyzing large volumes of text data, GraphRAG's implications are vast and versatile. 🔹 Looking Forward: The advent of GraphRAG not only enhances Microsoft's suite of AI tools but also sets a new standard for data analysis technologies. As companies increasingly rely on vast amounts of data to make decisions, tools like GraphRAG that can efficiently navigate and extract meaning from these resources will become invaluable. 👉 https://lnkd.in/dqzVD9Hr 👥 Let’s discuss: 🔹 How do you see GraphRAG transforming your industry? 🔹 What challenges and opportunities do you anticipate with the adoption of advanced AI tools like GraphRAG in data analysis? #LLM #RAG #DataScience #AI #MicrosoftResearch #Innovation #TechnologyTrends #GraphRAG

On April 17, 2024, the U.S. Department of Commerce announced the launch of a Request for Information RFI regarding AI-Ready Open Government Data Assets, initiated by the #AI and Open Government Data Assets Working Group. In particular, the Department noted that the working group, established by its #DataGovernance Board, aims to develop guidelines and best practices for achieving AI-ready open government #data. https://lnkd.in/gZDVthjU #worlddatasummit #dataprivacy #datascience #dataliteracy #opensource #dataanalytics #dataarchitecture #bigdata

I am excited to share that the Department of Commerce has released a Request for Information on AI and Open Government Data Assets. Commerce is interested in how generative AI systems can easily and responsibly utilize the department’s public data assets, with the goal of democratizing access to its public data. This RFI seeks valuable insights from industry experts, researchers, civil society organizations, and other members of the public on the development of AI-ready open data assets and modernized data dissemination standards. AI and data experts, please submit comments before July 16th! Learn more about the RFI here. https://lnkd.in/ewMp5MSc

Calling all data enthusiasts! 📢The Department of Commerce is on a mission to make government data more interpretable for AI applications. They've just released a Request for Information (RFI) seeking your input on data dissemination methods and data structures. This is a golden opportunity to help shape the future of AI-powered government data!This RFI is part of a larger initiative by the AI and Open Government Data Working Group to democratize access to Commerce data through Generative AI. Generative AI has the potential to revolutionize the way we interact with data, and your insights can help make it a reality.Head over to Request for Information (RFI) on AI-Ready Open Government Data Assets to learn more and share your thoughts. Let's work together to unlock the full potential of government data!#AI #OpenData #GovernmentData #GenerativeA #RFI #DataScience #MachineLearning #DemocratizeDatahttps://lnkd.in/efy_PhFa

Interested in the outcomes of the RFI. I brought this up as a discussion point this afternoon at a Canada School of Public Service | École de la fonction publique du Canada event on Using Generative AI in the Government of Canada [https://lnkd.in/gb7J66Z9]. I was remembering back to Paul Kishchuk's talk at the Canadian Open Data Society / Communauté canadienne des données ouvertes summit [https://lnkd.in/gzXgvFV8], and thought there has to be more granular best practices out there for making #opendata as consumable for LLMs. Or how can Open Data Portals better enable use of the data for RAG? #gcdigital #gcdata

I've been reading a lot lately about RAG vs long context windows on AI applications, and I couldn't help but dive into Microsoft Research's latest innovation—GraphRAG! Did you know GraphRAG is designed to supercharge the capabilities of large language models (LLMs) when it comes to handling private datasets securely? So far, it seems to me that, this tool could be a game-changer in the way we approach data synthesis from diverse sources while maintaining tight context, especially as the knowledge-based use for RAG applications grows. GraphRAG stands out by using knowledge graphs combined with cutting-edge graph machine learning, which significantly improves the Q&A performance and document inspections. This method isn't just about retrieving data; it’s about delivering precise, context-rich, and reliable answers, complete with verifiable sources. Intrigued by how knowledge graphs are transforming data retrieval? LangChain's blog provides a fantastic overview of integrating these graphs with RAG applications, revolutionizing both structured and unstructured data handling. These types of developments promise extensive enhancements in various sectors, promising a profound impact on how we manage complex datasets.  Do you have thoughts on how this might change your work or curiosity about data science innovations? Drop a comment below—I'd love to hear your views! Link to Microsoft research here: https://lnkd.in/gJ4Dk2er Link to langchain blog post here: https://lnkd.in/g2SS7ZhW #DataScience #Innovation #GraphRAG #FutureOfWork  #TechTrends #AI #MachineLearning

Interesting read.

Perhaps the greatest challenge – and opportunity – of LLMs is extending their powerful capabilities to solve problems beyond the data on which they have been trained, and to achieve comparable results with data the LLM has never seen. This opens new possibilities in data investigation, such as identifying themes and semantic concepts with context and grounding on datasets. In this post, we introduce GraphRAG, created by Microsoft Research, as a significant advance in enhancing the capability of LLMs. #AI #Microsoft #RAG