🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

We trained a Vision Transformer to solve ONE single task from François Chollet and Mike Knoop’s ARC Prize. Unexpectedly, it failed to produce the test output, even when using 1 MILLION examples! Why is this the case? 🤔

just achieved a score of 53% on the ARC Prize - what a feeling! Jack Cole Mohamed Osman lets gooo!

yes

New SoTA on ARC-AGI. Nothing like the synergy of an awesome team (Michael Hodel Mohamed Osman). From 53 to 54.5 today. Onward and upward! 🚀 ARC Prize Mike Knoop François Chollet Bryan Landers Greg Kamradt Machine Learning Street Talk First. #kaggle - kaggle.com/competitions/a…

New ARC-AGI paper ARC Prize w/ fantastic collaborators Wen-Ding Li @ ICLR'25 Keya Hu Zenna Tavares evanthebouncy Basis For few-shot learning: better to construct a symbolic hypothesis/program, or have a neural net do it all, ala in-context learning? cs.cornell.edu/~ellisk/docume…

We got upto 55.5% on the ARC Prize leaderboard today! Progress towards the 60.2 % milestone of median human performance reported by arxiv.org/pdf/2409.01374 is not slowing down. Jack Cole Michael Hodel

I finally got to meet François Chollet in person recently to interview him about ARC Prize, intelligence vs memorization, human cognitive development, learning abstractions, limits of pattern recognition and consciousness development. These are the best bits. Full show released tomorrow

Have been working on my 2nd synthetic ARC riddle generator (agent: ideation -> prog generation). Got >1k diverse generator+solver pairs as PoC so far. Some nice examples:

very excited to win guys, it's been such a blast! let's goo

ARC prize 2024 🥈place paper by the ARChitects who scored 53.5 (56.5): github.com/da-fr/arc-priz… - Transformers/LLMs are for ARC what ConvNets were for Imagenet - strong base model, TTT, specialized datasets (e.g. Michael Hodel’s re-arc) + novel: DFS sampling with LLM critique

talk is still cheap

it's not solved

Consulting my heart... Ok, looks like you haven't. But whenever you have a SotA (or close) solution built on top of the OpenAI API we're more than happy to verify it and add it to the public ARC Prize leaderboard. Anything using less than $10k worth of API calls is eligible.

Great presentation on some unique TTT ideas and experiments by Jonas Hübotter Tufalabs. youtu.be/vei7uf9wOxI?si…

Today, MindsAI (Jack Cole Mohamed Osman Michael Hodel) becomes part of Tufalabs First assignment: complete the ARC Prize challenge

Happy to announce we outperformed OpenAI o1 with a 7B model :) We released two self-improvement methods for verifiable domains in our preliminary paper -->

🚨 NEW PAPER DROP! Wouldn't it be nice if LLMs could spot and correct their own mistakes? And what if we could do so directly from pre-training, without any SFT or RL? We present a new class of discrete diffusion models, called GIDD, that are able to do just that: 🧵1/12

Excited to advance our lead and SoTA score on ARC-AGI-2 (ARC Prize) by 3 points to 15.28. Dries Smit Mohamed Osman Michael Hodel Greg Kamradt Tufalabs kaggle.com/competitions/a…

📝 New research: AlphaWrite applies evolutionary algorithms to creative writing. Inspired by AlphaEvolve, we use iterative generation + Elo ranking to systematically improve story quality through inference-time compute scaling. Results: 72% preference over baseline generation

DM if you're interested in meeting!