Who Verifies AI Is Who It Claims to Be? New Paper Defines the AI Identity Problem
A new arXiv paper proposes a conceptual framework for "AI identity," systematically exploring the challenges of identifi…
📂 Research
Latest articles in Research
AdaMamba: Adaptive Frequency-Gated Mamba Revolutionizes Long-Term Time Series Forecasting
A latest paper proposes the AdaMamba architecture, combining an adaptive frequency gating mechanism with the Mamba state…
CAP-CoT: Enhancing LLM Chain-of-Thought Reasoning Stability with Cycle Adversarial Prompting
A new study proposes CAP-CoT (Cycle Adversarial Prompt), a method that significantly improves the instability of chain-o…
Active Inference: A New 'Phenotyping' Framework for AI Agents
A latest arXiv paper proposes a phenotyping method for evaluating AI system agency based on active inference, using thre…
With Just a 1-Bit Danger Signal, LLM Agents Autonomously Learn Safety Norms
A latest arXiv paper proposes the EPO-Safe framework, enabling large language model agents to autonomously evolve natura…
NVIDIA BioNeMo Introduces Context Parallelism to Break Through Biomolecular Modeling Bottlenecks
NVIDIA has introduced Context Parallelism technology into its BioNeMo framework, breaking through single-GPU memory limi…
AI-Assisted Research Suggests Fish Oil May Harm Brain Health
A latest study leveraging AI-powered big data analysis suggests that long-term use of fish oil supplements may negativel…
StoryTR: Using Theory of Mind to Help AI Understand Video Narratives
A research team has proposed StoryTR, the first video moment retrieval benchmark for narrative content. By introducing T…
Who Judges the Judges? Bias Mitigation Strategies for LLM Evaluators Receive First Systematic Assessment
A large-scale empirical study systematically compared 9 debiasing strategies across 5 mainstream LLM judges, revealing t…
From Coarse to Fine: A New Breakthrough in Self-Adaptive Hierarchical Planning for LLM Agents
A latest arXiv paper proposes a self-adaptive hierarchical planning framework that addresses the core bottleneck of fixe…
PhySE Framework Reveals Social Engineering Attack Threats Combining AR Glasses with Large Language Models
A latest arXiv paper proposes the PhySE psychological framework, systematically analyzing the emerging threat of real-ti…
New Research: Don't Let LLMs Read Graphs — Let the Graphs Think for Themselves
A latest arXiv paper, through over 3,000 controlled experiments, finds that belief graphs as prompt context are nearly u…