🔥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

Deep Equilibrium Optical Flow Estimation Shaojie Bai, Zhengyang Geng, Yash Savani, Zico Kolter tl;dr: DEQ ("infinite depth aka single layer", arxiv.org/abs/1909.01377) look like natural fit for optical flow estimation. arxiv.org/abs/2204.08442 github.com/locuslab/deq-f…

Happy to share our latest DEQ work with Shaojie Bai, Yash Savani, and Zico Kolter! DEQ flow now sets SOTA zero-shot generalization performance on KITTI-15, with over 20% error reduction and super strong efficiency! Paper and Code available at paperswithcode.com/paper/deep-equ…. #CVPR2022

Excited about this work with Asher Trockman Yash Savani (and others) on antidistillation sampling. It uses a nifty trick to efficiently generate samples that makes student models _worse_ when you train on samples. I spoke about it at Simons this past week. Links below.

I’ll be at ICLR next week presenting this paper co-written with Alex Damian (Alex Damian). Would love to meet and chat about optimization in deep learning! My DMs are open - please reach out via DM or email. openreview.net/forum?id=sIE2r…

✨ Did you know that NOT using all generated rollouts in GRPO can boost your reasoning LLM? Meet PODS! We down-sample rollouts and train on just a fraction, delivering notable gains over vanilla GRPO. (1/7)

*Antidistillation Sampling* by Yash Savani Asher Trockman Zico Kolter et al. They modify the logits of a model with a penalty term that poisons potential distillation attempts (by estimating the downstream distillation loss). arxiv.org/abs/2504.13146