AI Achieves End-to-End Autonomous Scientific Discovery on a Real Optical Platform for the First Time
A latest arXiv paper demonstrates that an LLM-based AI agent has completed end-to-end autonomous scientific discovery on…
🏷️ STEM
155 articles about 'STEM'
Microsoft Research: What Happens When AI Agents Interact at Scale
Microsoft Research publishes new findings revealing that individually safe AI agents don't guarantee the overall safety …
Open Source Documentary Cult.Repo: Telling the Stories of the Internet's Unsung Heroes Through Film
The documentary project Cult.Repo shines a spotlight on open source software and the developer communities behind it, us…
Multi-Agent Collaboration: The Hardest Engineering Challenge in Scaling AI
Intuit's engineering team shares practical experience with multiple AI agents working collaboratively in complex systems…
Airbnb Launches AI Destination Recommendation Model, Revolutionizing Travel Planning
Airbnb's engineering team has unveiled its in-house destination recommendation model that leverages AI technology to ins…
New Method Enables Provably Safe Constraint Satisfaction for Black-Box Hybrid Dynamical Systems
A latest paper on arXiv proposes a control policy learning method for black-box hybrid dynamical systems that rigorously…
New Framework Significantly Boosts Energy Efficiency in Multi-Robot Coverage Path Planning
A latest arXiv paper proposes an energy-efficient multi-robot coverage path planning framework for non-convex regions, o…
"Wiggle Then Throw": A New Zero-Shot Approach to Dynamic Rope Manipulation
A research team has proposed a novel method for zero-shot dynamic rope manipulation using learned simulation priors. By …
L-System Genetic Encoding: A New Path for Neural Network Evolution
Researchers propose Lsys, an L-System-based genetic encoding method for neural networks. By comparing it with traditiona…
Dune-Themed Malicious Code Discovered in PyTorch Lightning Library
The AI training framework PyTorch Lightning was found to contain malicious code themed after the sandworm "Shai-Hulud" f…
New Method Enables Robots to Learn Complex Motions from Few Demonstrations
A latest arXiv paper proposes "Phase-varying Neural Potential Functions," addressing the limitations of dynamical system…
The Multi-Agent Resource Allocation Dilemma: Fewer and Stronger, or More and Simpler?
A new arXiv study introduces the 'Split over n' resource-sharing problem, exploring whether limited resources in multi-a…