Curriculum Learning
- Formal Mathematics Statement Curriculum Learning
- neural theorem prover using GPT-f
- solve a curriculum of increasingly difficult problems out of a set of formal statements of sufficiently varied difficulty
- high-school Math Olympiad problems
- language model to find proofs of formal statements
- formal mathematics
- at same compute budget, expert iteration, by which they mean proof search interleaved with learning, dramatically outperforms proof search only
- expert iteration is capable of finding and solving a curriculum of increasingly difficult problems, without the need for associated ground-truth proofs
- miniF2F
- automatically solving multiple challenging problems drawn from high school olympiads
- lack of self-play in the formal mathematics setup can be effectively compensated for by automatically as well as manually curated sets of formal statement
- cheaper to formalize than full proofs