About
Research & Development
Beyond Code: New Benchmark Tests AI’s Environment Setup Skills
TLDR: SetupBench is a new benchmark with 93 tasks designed to evaluate how well AI software engineering agents can set up development environments from...
Research & Development
AI Agents Uncover Human-Like Behaviors in Simulated Auctions
TLDR: A new research paper demonstrates that large language models (LLMs) can effectively simulate human behavior in various auction formats, replicating established economic findings...
Research & Development
New Gradient-Based Methods Bring Stability and Speed to Deep Reinforcement Learning
TLDR: This research introduces GPBE(λ), an extension of the Generalized Projected Bellman Error, to enable multistep credit assignment in deep reinforcement learning. It derives...
Research & Development
CompassJudger-2: Advancing LLM Evaluation with a New Generalist Judge Model
TLDR: CompassJudger-2 is a new generalist judge model designed to overcome the limitations of existing specialized and less robust LLM evaluators. It uses a...
Research & Development
Optimizing Graph Neural Networks with Adaptive Prompt Pruning
TLDR: A new framework, GPAWP, enhances Graph Neural Networks by combining graph prompt learning with adaptive weight pruning. It evaluates the importance of different...
Research & Development
POIFormer: Enhancing Location Attribution with Advanced AI for Real-World Mobility Data
TLDR: POIFormer is a new AI framework that uses a Transformer-based model to accurately link GPS data to specific Points of Interest (POIs). It...