What Matters for Dexterous RL?
Reading notes on dexterous RL: reset distributions, goal-conditioned pretraining, structured exploration, motion tracking, contact priors and simulation.
Paper-Reading
How to train your Robo-centric VLMs
Reading notes on Robo-centric VLMs: data construction, spatial grounding, multi-task training, and closed-loop robot execution.
Fantastic Mega-Scale RL Recipes and where to find them
Reading notes on Mega-Scale RL: batches, multiple policies, reset distributions and off-policy methods.
Loops Inside the Denoiser
Reading notes on ELT, Looped-DiT and LiFT, with comparisons to hidden states across steps and DEQ.
A Hitchhiker's Guide To Diffusion RL
My reading notes on diffusion / flow post-training, comparing the approaches used in image/video generation and robotics.
Titans: Learning to Memorize at Test Time
This article introduces Titans, a novel architecture that as a meta in-context learner, learns to memorize at test time. Through designing a long-term memory module, and proposing three variants of Titans (MAC, MAG, MAL), the model achieves superior performance compared to Transformers and other baselines, especially in long-context tasks.