self play
3 stories and discussions about self play, aggregated from every source we track.
1.
Advances in language modeling have been driven by scaling pretraining on ever more data. Yet, the training data is still largely curated on the model's behalf. A more general approach to pretraining would let the model…
2.
Advances in language modeling have been driven by scaling pretraining on ever more data. Yet, the training data is still largely curated on the model's behalf. A more general approach to pretraining would let the model…
3.
A breakthrough post-training result for physical AI via self-play: a strong base model like S1 can learn to complete extremely dexterous and dynamic tasks, like soccer, by competing against itself in a simulation.