Ahmed Maher Aly’s Post

This is a great application of the Gemma models! read the post and follow the colab to learn more how to ground your queries on real data with two methods: RAG (that you may already know) and RIG!!! #gemmaverse

✨ Introducing DataGemma, open models that enhance LLM factuality by grounding them with real-world data from Google’s Data Commons. → https://goo.gle/4dWFyKn Learn how DataGemma uses two distinct RIG and RAG approaches to: 📖 Harness the knowledge of Data Commons 🧠 Fact check numeric data against trusted sources 🛠️ Lead the way in developing responsible AI And more for the #Gemmaverse!

Sounds #incredible - I mean why have LLM #hallucination when you can just... RIG & RAG ;-) 1. RIG (Retrieval-Interleaved Generation) proactively querying trusted sources and fact-checking against information in #Google #DataCommons 2. RAG (Retrieval-Augmented Generation) incorporate relevant information beyond their training data, absorb more context, and enable more comprehensive and informative outputs. 

DataGemma is designed to help address the challenges of hallucination by grounding LLMs in the vast, real-world statistical data of Google's Data Commons. One of the major hurdles in generative AI today is how 𝐋𝐋𝐌𝐬 sometimes present incorrect information. Google has launched 𝐃𝐚𝐭𝐚𝐆𝐞𝐦𝐦𝐚, the first open models connected to the vast 𝐃𝐚𝐭𝐚 𝐂𝐨𝐦𝐦𝐨𝐧s repository, offering over 𝟐𝟒𝟎 𝐛𝐢𝐥𝐥𝐢𝐨𝐧 reliable data points from sources like the UN, WHO, and CDC. With 𝐃𝐚𝐭𝐚𝐆𝐞𝐦𝐦𝐚, AI models are now grounded in real-world data, leveraging two key approaches: 1. 𝐑𝐈𝐆 (𝐑𝐞𝐭𝐫𝐢𝐞𝐯𝐚𝐥-𝐈𝐧𝐭𝐞𝐫𝐥𝐞𝐚𝐯𝐞𝐝 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧): Queries trusted sources and retrieves data from Data Commons for more accurate outputs. 2. 𝐑𝐀𝐆 (𝐑𝐞𝐭𝐫𝐢𝐞𝐯𝐚𝐥-𝐀𝐮𝐠𝐦𝐞𝐧𝐭𝐞𝐝 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧): Extends context, allowing models to generate more informed, factual responses.

Interesting step forward in Explainable AI.

✨ GOOGLE is introducing DataGemma, open models that enhance LLM factuality by grounding them with real-world data from Google’s Data Commons. Learn how DataGemma uses two distinct RIG and RAG approaches to: 📖 Harness the knowledge of Data Commons 🧠 Fact check numeric data against trusted sources 🛠️ Lead the way in developing responsible AI And more for the hashtag #Gemmaverse!

Ready, set....AI! For data, the emergence of AI signals a paradigm shift. How do we harness and interpret our vast oceans of data now that AI has entered the chat? To answer: we created The Big Book of AI-Ready Data. Check out our high-level overview on AI-ready data. It's chock full of case studies, use cases, and a checklist for assessing your AI-ready data framework. Read on: https://lnkd.in/eEezEF9z

A common theme from our last Gen AI event was the topic of combining LLMs with structured/unstructured data sources, such as spreadsheets. I thought I'd reshare this tweet that Tim Elder, our CPO, shared internally from Ethan Mollick. 📊🧠 https://lnkd.in/eZEXi7WA #AIInnovation #FinancialServices #SpreadsheetLLM #MultiplyAI