Kaggle

⭐️ KaggleX Project Spotlight: Cole Kainz Cole's project is built an Interactive Study Guide and Exam Generator utilizing ColPali & PaliGemma to create a personalized student learning tool for students. Check out Cole’s project presentation here: https://lnkd.in/eW5i3AHN

Introducing FACTS Grounding 🧠📐 A new benchmark we’re launching with Google DeepMind designed to evaluate the factual accuracy of large language models (LLMs) across over 1,700 tasks. Check it out! 👉 Leaderboard: https://lnkd.in/ewja5ntf  👉Technical Report: https://lnkd.in/eeZZx_5W 👉 Blog post: https://lnkd.in/eDkGZX28 FACTS Grounding is made up of 1719 factual grounding tasks. Each example includes: - a document - a system instruction requiring the LLM to exclusively reference the provided document - an accompanying user request LLM responses are evaluated by judge models to ensure they fulfill the user request and remain grounded in the provided context. To ensure accessibility, we’ve published 860 tasks as a public data set for anyone to use: https://lnkd.in/eg4McdYy We plan to expand and maintain FACTS over time with new models, tasks and evaluations. Our goal is to promote the development of AI systems that are both knowledgeable and trustworthy. We’ll also be working with model owners and the broader community to add more models to the leaderboard over time. Let us know what we should evaluate next! 👀 

Today at NeurIPS in Vancouver our Head of AI/ML Luis Oala together with industry colleagues are showcasing Croissant 🥐 - metadata format designed to make working with datasets in Machine Learning simpler, faster, and more responsible. By making datasets more discoverable, portable, and interoperable across tools and frameworks, Croissant eases the friction points in ML data management. Backed by popular dataset repositories and compatible with major ML frameworks, Croissant is ready to unlock the full potential of your data. Learn more: NeurIPS 🔗: https://lnkd.in/d6CdqFGA   Github 🔗: https://lnkd.in/dK4zDZx2 Poster Session: Wednesday, December 12, 11 am. – 2 pm. PST The MLCommons Croissant working group, the team that 'baked' this new standard, brings together AI/ML experts from the following organisations: 🥐  Bayer 🥐  cTuning foundation (founding member of MLCommons) 🥐  DANS-Knaw 🥐  Dotphoton  🥐  Google 🥐  Harvard University 🥐 Hugging Face 🥐 Kaggle 🥐 King's College London 🥐 Luxembourg Institute of Science and Technology (LIST) 🥐 Meta 🥐 NASA - National Aeronautics and Space Administration 🥐  NASA IMPACT - UAH 🥐 Open Data Institute 🥐 Universitat Oberta de Catalunya 🥐 Eindhoven University of Technology Carole-Jean Wu, Costanza Conforti (Stanze), D. Sculley, Dan Brickley, Drew Duglan, Ph.D., Elena Simperl, Goeff Thomas, Grigori Fursin, Iksha Gurung, Joan Giner Miguelez, Joaquin Vanschoren, Jos van der Velde, Luis Oala, Manil Maskey, Meg Risdal, Michael Kuchnik, Mubashara Akhtar, Natasha Noy, Nitisha Jain, Omar Benjelloun, Peter Mattson, Philip Durbin, Pierre Marcenac, Pierre Ruyssen, Pieter Gijsbers, Quentin Lhoest, Rajat Shinde, Slava Tykhonov, Stefano Maria Iacus, Steffen Vogler, Sylvain Lesage, and Yuhan (Douglas) Rao #neurIPS #neurIPS2024 #croissant #metadataformat #ML

🤖 Now on #KaggleModels! PaliGemma 2, the tunable vision-language model, brings the power of sight to Gemma 2. With scalable performance, detailed image captioning, and specialized task capabilities, building advanced, customizable AI that can "see" is now easier than ever. https://lnkd.in/ekzk9qHv 

Competition Reminder: Jane Street Real-Time Market Data Forecasting 🚨 Midway through the Jane Street competition—time to join 2,736 participants and refine the forecasts! 🎯 Develop an ML forecasting model using real-world data derived from production systems 💰 Prize Pool: $120,000 ⏰ Entry Deadline: 01/06/2025 This challenge highlights the difficulties in modeling financial markets, including fat-tailed distributions, non-stationary time series, and sudden shifts in market behavior. It's a chance to engage with problems similar to those faced by Jane Street's quantitative teams, who have spent decades innovating in this field. Learn More: https://lnkd.in/dg3MyruS

📣 Competition Launch Alert! Konwinski Prize hosted by Andy Konwinski 🎯Test AI’s capability to solve GitHub issues 💰 $1,225,000 Prize Pool ⏰ Entry Deadline: 3/5/25 Build AI models to help automate GitHub issue resolution so you can spend less time fixing bugs and more time building. Your work will help shape the future of open & effective software development. Learn more at https://lnkd.in/eEX6iEzF

⭐️ KaggleX Project Spotlight: Ayomide Zul Kazeem Ayomide's project explores how LLMs can optimize operations in the traditional water industry to promote sustainability & increase public awareness of conservation efforts. Check it out: https://lnkd.in/eZVSRaXJ

Build with Gemini 2.0 Flash Experimental. ⚡ 2.0 Flash is 2x as fast as 1.5 Pro while achieving stronger performance, and comes with the new Multimodal Live API and native tool use. 🧪 Now available to test through the Gemini API, Google AI Studio, and Vertex AI. → https://goo.gle/3ZMJoRz

Day one of #nuerips2024 is in the books, and we're all a little hoarse and footsore, but also just so happy to be around so many machine learning nerds. Here's some highlights from the Kaggle booth today. Come check us out if you're at the conference! Phil Culliton, Walter Reade, Brenda Flynn, D. Sculley, Paige Bailey, Abhishek Kaushik, Nate Keating

We’re excited to announce that Kaggle’s 5-Day Generative AI Intensive Course with Google is now available as a self-paced Learn Guide! This no-cost course was created by a team of Google ML researchers and engineers and designed to help you better understand some of the fundamental technologies and techniques behind Generative AI. Access whitepapers, podcasts, code labs, and recorded livestreams—all designed to help you build your Generative AI expertise. Additionally, there is a bonus assignment for you! Access Learn Guide Here: https://lnkd.in/enYP2Xxi Happy Learning!