Reposted by Shengran HuSakana AI @sakanaai.bsky.social · 30/05/2025Introducing The Darwin Gödel Machine sakana.ai/dgm The Darwin Gödel Machine is a self-improving agent that can modify its own code. Inspired by evolution, we maintain an expanding lineage of agent variants, allowing for open-ended exploration of the vast design space of such self-improving agents. 1358
Reposted by Shengran HuJeff Clune @jeffclune.com · 30/05/2025Excited to introduce the Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents. We harness the power of open-ended algorithms to search for agentic systems that get better at coding, including improving their own code. 0286
Reposted by Shengran HuThe Register @theregister.com · 02/06/2025Boffins found self-improving AI sometimes cheateddlvr.itBoffins found self-improving AI sometimes cheatedInstead of addressing hallucinations, it just bypassed the function they built to detect them Computer scientists have developed a way for an AI system to rewrite its own code to improve itself.… 0144
Reposted by Shengran HuSakana AI @sakanaai.bsky.social · 08/04/2025Introducing The AI Scientist-v2, which produced the 1st fully AI-generated paper to pass peer review at a workshop level (at #ICLR2025) ‼️ Tech Report: pub.sakana.ai/ai-scientist... GitHub: github.com/SakanaAI/AI-... This work is a proud collaboration between Sakana AI, UBC, and Oxford University. 1187
Reposted by Shengran HuVector Institute @vectorinstitute.ai · 11/02/2025🤔 What if AI could explain its thought process before taking action? Our latest ANDERS blog covers Vector Researchers @jeffclune.com & @shengranhu.bsky.social’s wrk on "thought cloning" - teaching AI to express its reasoning process in language we can understand vectorinstitute.ai/thought-clon...vectorinstitute.aiThought Cloning: Teaching AI to Think Like Humans for Better Decision-Making - Vector Institute for Artificial IntelligenceNew research from Vector Faculty Member Jeff Clune and Vector Graduate Student Shengran Hu introduces a groundbreaking approach to imitation learning that could potentially revolutionize how we train ... 032
Reposted by Shengran HuJeff Clune @jeffclune.com · 12/02/2025Introducing Automated Capability Discovery! ACD automatically identifies surprising new capabilities and failure modes in foundation models, via "self-exploration" (models exploring their own abilities). Led by @cong-ml.bsky.social & @shengranhu.bsky.social 🔬🤖🧠🔎 [1/9] 1193
Reposted by Shengran HuJeff Clune @jeffclune.com · 16/12/2024Lots of interest in ADAS! Thanks everyone, and congrats Shengran Hu and @cong-ml.bsky.social! 🚀🚀🚀 0103
Reposted by Shengran HuJeff Clune @jeffclune.com · 16/12/2024Jenny did a great job presenting OMNI-EPIC during the IMOL Oral. Congrats Jenny Zhang! arxiv.org/abs/2405.15568 071
Reposted by Shengran HuJeff Clune @jeffclune.com · 12/12/2024Come stop by our poster this afternoon and say hi! 162
Reposted by Shengran HuJeff Clune @jeffclune.com · 15/12/2024This is the newest work out of our lab. Please stop by if you're curious to learn about our latest, which I'm very excited about! 042