🔥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

😍

Happening now at ICL workshop!!

They are more like "in-context inferencing"😉

Thanks for visiting our poster!!!

Can LLMs understand the semantic meaning of a graphics programs? Checkout our latest benchmark for the answer! **SGP-Bench** 🔗 sgp-bench.github.io

One of our unique insights is that, instruction-tuning your LLM with graphics program descriptions can effectively improve general reasoning ability! This is because the way we created the description of graphics programs distills both the language and vision knowledge of

Since some people are particularly interested in how/why symbolic graphcis programs can improve general reasoning, I want to elaborate more on this. Different from normal programs where the output is typically some numerical values, symbolic graphics programs output a 2D/3D

How could online learning apps adapt to learners and improve over time? Even if you're not a machine learning expert, Hanqi Zhou's blog post on our last ICLR paper explains new approaches in simple terms (joint work with 🐘 Álvaro Tejero-Cantero @[email protected] and Charley Wu (@thecharleywu.bsky.social), supported by Theresa Authaler).

📢Reminder to submit your cool results to our @NeurIPS2024 workshop on ML & Compression!! Deadline: Sept 30 (AoE). Submission link: openreview.net/group?id=NeurI… Fill out this form if you'd like to become a reviewer: forms.gle/UjHFrL4C3nc9ab…

Learning to play the piano with two robot hands is super challenging, even in simulation! It requires coping with bimanual coordination at high speed to achieve human-level dexterity. We introduce RP1M, a large-scale robot piano-playing motion dataset, featuring ~1M trajectories

🧐 Interesting finding: we test Reflection-70B on our SGP-Bench (sgp-bench.github.io, a benchmark evaluating symbolic program understanding). Despite Reflection-70B claims to outperform many LLMs, we find it actually still worse than Llama3.1-70B.

🚨New open-access paper alert! 🚨We investigated whether English listeners can use measure phrases like "a pile of" to revise their prior prediction on the fly. tandfonline.com/doi/full/10.10…

Biggest joy and honour leading this project at the intersection of visual neuroscience and ML to a successful finish! Paper: nature.com/articles/s4158…

Exciting News! Our cluster member Jakob Macke, PhD student Janne Lappalainen’s study with Srini Turaga et al. on the simulation of the fruit fly visual system with deep learning has been released in nature! Read more about it on our website: uni-tuebingen.de/en/research/co…

😍Lots of interesting and valuable feedback from the poster session of 2024 IMPRS-IS Boot camp for our work - Verbalized Machine Learning! Intelligent Systems Cyber Valley 🔗 arxiv.org/abs/2406.04344

New work on DrMMD gradient flow 📣 arxiv.org/pdf/2409.14980 1️⃣ tractable numerical implementation 2️⃣ interpolation between MMD flow and chi2 flow 3️⃣ Near-global convergence Joint work with @Aratrika Moustafi Pierre Glaser Anna Korba Arthur Gretton @Bharath Sriperumbudur

Congratulations Yao Feng for being named one of MIT's Rising Stars in EECS! Her work includes regressing 3D faces & bodies (DECA, PIXIE), avatars (TECA), mesh representations (Ghost-on-the-shell, MeshDiffusion), and human foundation models (ChatPose) risingstars-eecs.mit.edu/participants/y…