Gu Dandan
Professor at the College of Artificial Intelligence, Jilin University. Her research covers agent systems and self-improvement technologies. She has contributed to survey research in the field of agent self-evolution.
Timeline (last 90 days)
The survey "Self-Improvement in Modern Agentic Systems" was jointly released by Professor Gu Dandan of the School of Artificial Intelligence at Jilin University and the AI team at King Abdullah University of Science and Technology.
Professor Gu's team has an ICLR paper building on the Gödel Machine, which treats the entire scaffolding of a Coding Agent as an evolvable program and designs an evolution strategy by evaluating the long-term potential of modified nodes.
Professor Gu pointed out that the biggest bottleneck in current self-evolution lies in verifying whether an improvement is effective, as it is difficult to determine in real time whether the current improvement is better than the previous one in open-ended tasks.
All Facts (3)
The survey "Self-Improvement in Modern Agentic Systems" was jointly released by Professor Gu Dandan of the School of Artificial Intelligence at Jilin University and the AI team at King Abdullah University of Science and Technology.
50%UnverifiedProfessor Gu's team has an ICLR paper building on the Gödel Machine, which treats the entire scaffolding of a Coding Agent as an evolvable program and designs an evolution strategy by evaluating the long-term potential of modified nodes.
50%UnverifiedProfessor Gu pointed out that the biggest bottleneck in current self-evolution lies in verifying whether an improvement is effective, as it is difficult to determine in real time whether the current improvement is better than the previous one in open-ended tasks.
50%