AgentSearchBench: The First AI Agent Search Benchmark Arrives
A research team has released the AgentSearchBench benchmark, designed to address the challenge of finding the right AI a…
📂 Research
Latest articles in Research
'Background Temperature' Concept Reveals Hidden Randomness in Large Language Models
Researchers introduce the concept of 'background temperature,' formally quantifying the phenomenon where large language …
When Does LLM Self-Correction Actually Work? A New Cybernetics Framework Offers a Diagnostic Formula
Researchers model LLM self-correction as a cybernetic feedback loop, proposing a concise diagnostic criterion based on M…
Memanto: Reshaping AI Agent Long-Term Memory with Information-Theoretic Retrieval
A research team introduces the Memanto framework, which solves memory bottleneck problems for long-horizon AI agents in …
MolClaw: Autonomous AI Agent Revolutionizes Drug Molecule Screening and Optimization
A research team has released MolClaw, an autonomous intelligent agent that integrates over 30 specialized tools. Through…
Academic Publishing Faces Transformation: AI Research Demands a New Certification Framework
As AI-automated research pipelines generate a growing volume of academic output, researchers propose a dual-layer certif…
New Testing Framework 'Math Takes Two' Challenges the Authenticity of AI Mathematical Reasoning
A latest arXiv paper proposes the 'Math Takes Two' testing framework, which examines whether language models possess gen…
AI Agents Can Reproduce Social Science Research Results Using Only Paper Method Descriptions
A research team developed an agentic reproduction system that automatically extracts structured methods, writes code, an…
Artifact-Based Agent Framework Revolutionizes Medical Image Processing
Researchers have proposed an artifact-based agent framework designed to address the challenges of adaptability and repro…
Training an mRNA Language Model Across 25 Species for Just $165
Researchers successfully trained an mRNA language model covering 25 species at a computational cost of just $165, openin…
Deep Research Max Launch: A Qualitative Leap for Autonomous Research Agents
Google has launched Deep Research Max, a major upgrade built on the Gemini Deep Research Agent architecture, marking a s…
AI Tools Are Redefining the Role of Software Engineers
A survey of 158 professional software engineers by researcher Annie Vella reveals that AI tools are shifting engineers' …