Paper Refinement Methods

[BradKML]

• If generated papers are rejected by the LLM reviewers, then the code should still be recyclable into a growing database of code
• LLM papers' code could be adjusted when the LLM reviewers critiques are minor and does not require scope expansions

[BradKML]

Found a repo that may be relevant to this, and yes AI can be more creative than a human https://github.com/NoviScl/AI-Researcher

[BradKML]

And another one, but this one is pretty fresh as well https://github.com/zjunlp/OmniThink