[PDG 443] Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents

Link to article: https://arxiv.org/pdf/2505.22954
Title: Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents
Content: Current AI systems cannot continuously improve themselves due to fixed architectures. The researchers propose the Darwin Gödel Machine (DGM), which iteratively modifies its own code and validates changes using coding benchmarks. Using an evolution-inspired approach, it maintains an archive of coding agents and samples from it to create improved versions. The DGM achieved significant performance gains, improving from 20.0% to 50.0% on SWE-bench and 14.2% to 30.7% on Polyglot. This represents an important step toward truly self-improving AI systems with safety precautions in place.
Slack link: ml-ka.slack.com, channel: #pdg. Please join us -- if you cannot join, please message us here or to mlpaperdiscussiongroupka@gmail.com.

In the Paper Discussion Group (PDG) we discuss recent and fundamental papers in the area of machine learning on a weekly basis. If you are interested, please read the paper beforehand and join us for the discussion. If you have not fully understood the paper, you can still participate – everyone is welcome! You can join the discussion or simply listen in. The discussion is in German or English depending on the participants.