This is a periodically updated paper list of AI4Math.
Since papers can have different usages and significance for different people, I try to avoid only providing the ones that I personally find interesting. Instead, here is the long list that can be skimmed and filtered (more details below).
It is a by-product when I read papers, and is open sourced because I hope it may be a little bit helpful to others.
Here are some sub-lists that may be potentially interesting:
- ICLR 2025 Submissions
- NeurIPS 2024
- All top conferences
- Papers using proof assistants (Lean / Isabelle / ...)
- Highly cited papers
- Papers highly cited by others in this field
- Recent ArXiv
The full list of papers without any filtering is here (Markdown) / here (Dataset).
This list is also provided as a dataset, with each row being a paper. Example usages may possibly be:
- Filter the list: This can be done via keywords, labels, or even LLMs. For example, "citation > 50 OR top conference", or "filter papers if Llama3 thinks this contains things of my personal interest".
- Supported labels:
citation_count
,content_markdown
,publish_info
,publish_is_top
,tldr_text
,citation_count_filtered_math_and_top_conf
,theorem_provers
, as well as normaltitle
,abstract
,authors
,date
,url
, etc
- Supported labels:
- Feed it into RAG systems: Then, this dataset (or a chosen subset) can be a source of truth for LLMs when talking with it.
- Skim the missing pieces: For example, by excluding the papers already in your Zotero, the remaining may contain missing papers and may worth a skim.
- In the future, AI may be powerful enough to read these papers and output summarizations / trends / insights / ideas / ...
- ...
The number of AI4Math papers collected in this repository.
Here are some other great resources, and many thanks to them!