A Survey on Deep Learning for Theorem Proving

Stream: Machine Learning for Theorem Proving

[Kaiyu Yang]

Hi all,

We released a survey on deep learning for theorem proving and autoformalization. Hopefully it can provide enough pointers for people who want to dig deeper into these topics.

We aimed to be as comprehensive as possible, but it's possible that we still missed some important papers. You're welcome to bring them up by opening issues in our GitHub repo, and we will consider incorporating them into the next version. Thanks!

arXiv: https://arxiv.org/abs/2404.09939
GitHub: https://github.com/zhaoyu-li/DL4TP

[Kaiyu Yang]

BTW, if you think this survey might be helpful for you, please consider re-tweeting us on X: https://twitter.com/KaiyuYang4/status/1780268062127669552. Thanks!

[Lasse Blaauwbroek]

Nice work @Kaiyu Yang and others! I guess in the dataset section, our recent Coq dataset meant for training graph models and language models is missing. For reference: website, preprint, zenodo and visualization

[Kaiyu Yang]

@Lasse Blaauwbroek Thanks! We'll add it to the next version of the paper. Meanwhile, could you submit a PR to https://github.com/zhaoyu-li/DL4TP and @yangky11 in the PR?