Michael Y. Li
@michaelyli_
CS PhD @StanfordAILab @StanfordNLP advised by @noahdgoodman and Emily Fox.
Prev: undergrad @princeton
472 Following 1.3K Followers
Can an LM, starting from random init (!!), learn to generate all of its pretraining data?
Introducing Self-Play Pretraining with Zero Data. Two models start from random initialization: a generator proposes programs for a universal Turing machine and a learner trains on their outputs. We never train on any real data, but see predictable scaling on natural datasets: zero-shot val loss on images, text, audio, and melodies decreases predictably with self-play compute. And the learner develops in-context learning capabilities.
A fun proof-of-concept, co-led with @AdityaCowsik and @KfirDolev and co-authors @gbruno_dl, @ANourya @noahdgoodman, and @YoavLevine.
Show more
You're wasting FLOPs when scaling inference compute: by independently sampling parallel attempts, you burn compute rediscovering the same solutions.
Introducing QuasiMoTTo: we scale parallel sampling with correlated samples instead! These samples have higher coverage, are marginally exact draws from the LLM, and can be generated in parallel.
Result: same performance with 25-47% fewer samples in test-time scaling + 50% fewer training steps in RL!
In our new paper, we explore the design space of correlated samplers. Work with co-authors @probablynotaz9 (co-lead), @gandhikanishk, @noahdgoodman, and Emily Fox!
Show more
