Tao Lin

Even state-of-the-art dexterous manipulation policies fail to generalize, achieving only 38.5% success on novel tasks after training on related ones.

LLM agents can now learn complex multi-turn tasks with critic-free RL by dynamically shaping advantages based on an evolving "Agentic Reference Graph."

Current image generation models struggle with rendering dense, multi-region text, with fidelity dropping by over 14 points when clarity is prioritized.

Parallel LLM scientific agents collaborate far more effectively when sharing raw evidence rather than hypotheses, unlocking a 9.5-point discovery quality gain and nearly doubling exploration speed without succumbing to premature consensus.

Contrastive flow matching can upper-bound forward-KL divergence on fixed preference pairs, resolving the training instability of unbounded DPO surrogates without requiring online rollouts.

Frontier-scale agentic RL on 700B+ MoEs is no longer locked inside hyperscaler proprietary stacks, achieving stable 263-second step times across 64 GB300 GPUs via a verified, open-source training infrastructure.

High-fidelity image synthesis does not require paired text from day one: pre-training visual priors on uncaptioned images before multimodal alignment beats conventional joint training pipelines to establish a new open-source DiT benchmark.