arXiv cs.AI
GRAIL: AI translation for scientists application workflow on satellite data
The paper introduces GRAIL, an AI translation system designed to assist scientists in converting Python geospatial…
Hub / Ai Research
ai-research · WeSearch
Ai Research news.
Page 8 of Ai Research headlines on WeSearch — deduped and updated continuously from 10+ editorial sources.
arXiv cs.AI
PRIMA: Operational Patterns for Resilient Multi-Agent Research with Verifiable Identity and Convergent Feedback
The paper titled 'PRIMA: Operational Patterns for Resilient Multi-Agent Research with Verifiable Identity and…
arXiv cs.AI
Uncertainty Decomposition via Cyclical SG-MCMC and Soft-label Learning for Subjective NLP
The paper discusses a novel approach to uncertainty decomposition in subjective natural language processing (NLP). It…
arXiv cs.AI
Proper Scoring Rules for Agentic Uncertainty Quantification
The paper introduces the Trajectory Proper Score (TPS) for evaluating agentic uncertainty quantification in AI. It…
arXiv cs.AI
Automated Detection and Classification of Delusion-related Content in Naturalistic Audio Diaries Using Multi-Agent Language Models
A new study presents an automated pipeline using multi-agent language models to detect and classify delusion-related…
arXiv cs.AI
Hylos: Operability Contracts for Model-Native Spatial Intelligence
The paper titled 'Hylos: Operability Contracts for Model-Native Spatial Intelligence' introduces a new systems…
arXiv cs.AI
Fundamental Limitation in Explaining AI
A recent paper discusses the inherent limitations in explaining AI systems, particularly large-scale models. The…
arXiv cs.AI
MDIA: A Multi-Agent Diagnostic Intelligence Pipeline on HealthBench Professional
The article discusses the MDIA, a Multi-Agent Diagnostic Intelligence Pipeline designed for clinical reasoning. It…
arXiv cs.AI
Emotional intelligence in large language models is fragmented across perception, cognition, and interaction
A recent study highlights the fragmented nature of emotional intelligence in large language models (LLMs). The…
arXiv cs.AI
Exploration of Perceptual Speech Features for Clinical Decision-Support in Mental Health Care
A new study explores the use of perceptual speech features to support clinical decision-making in mental health care.…
arXiv cs.AI
When Mean CE Fails: Median CE Can Better Track Language Model Quality
The paper discusses the limitations of mean cross-entropy (CE) as a metric for evaluating language model quality. It…
arXiv cs.AI
Measuring Reasoning Quality in LLMs: A Multi-Dimensional Behavioral Framework
A new study proposes a multi-dimensional framework for evaluating reasoning quality in large language models (LLMs).…
arXiv cs.AI
Beyond Inference-Only Deployment: Comparing Weight-Based Consolidation Against Cascading Compaction
The article discusses a new approach to deploying large language models (LLMs) that goes beyond inference-only…
arXiv cs.AI
AVBench: Human-Aligned and Automated Evaluation Benchmark for Audio-Video Generative Models
AVBench is a newly introduced benchmark aimed at improving the evaluation of audio-video generative models,…
arXiv cs.AI
GlobalDentBench: A Multinational Benchmark for Evaluating LLM Clinical Reasoning in Dentistry with Expert Calibration
GlobalDentBench is introduced as the first multinational benchmark for evaluating large language models (LLMs) in…
arXiv cs.AI
Lattice theory and algebraic models for deep convolutional learning based on mathematical morphology
A new paper presents an algebraic framework for deep convolutional learning based on lattice theory and mathematical…
arXiv cs.AI
Agent-as-Peer-Debriefer: A Multi-Agent Framework with Perspective-Based Refinement for Qualitative Analysis
The article presents a new framework called Agent-as-Peer-Debriefer designed to enhance qualitative data analysis…
arXiv cs.AI
Hera: Learning Long-Horizon Coordination for Device-Cloud Collaborative LLM Agents
The article discusses a new approach called Hera for coordinating device-cloud collaborative large language model…
arXiv cs.AI
Learning to Reason Efficiently with A* Post-Training
A recent study explores the use of A* search algorithms to improve reasoning in large language models (LLMs). The…
arXiv cs.AI
HeartBeatAI: An Interpretable and Robust Deep Learning Framework for Multi-Label ECG Arrhythmia Detection
HeartBeatAI is a new deep learning framework designed for multi-label ECG arrhythmia detection. It addresses…
arXiv cs.AI
Associations between echocardiographic traits and AI-ECG predictions of heart failure
A recent study explored the relationship between echocardiographic traits and AI-ECG predictions of heart failure. The…
arXiv cs.AI
Summoning the Oracle to Slay It: Mitigating Look-Ahead Bias in Financial Backtesting with Large Language Models
A new paper proposes a method to mitigate look-ahead bias in financial backtesting using large language models. The…
arXiv cs.AI
Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models
The paper discusses a new framework for safe fine-tuning of large language models (LLMs) called Buffer-and-Reinforce.…
arXiv cs.AI
PALoRA: Projection-Adaptive LoRA for Preserving Reasoning in Large Language Models
The paper introduces PALoRA, a framework designed to enhance the integration of new knowledge into Large Language…
arXiv cs.AI
Beyond Control-Flow: Integrating the Resource Perspective into Multi-Collaborative Process Modeling from Text
The paper discusses a novel approach to process modeling in Business Process Management (BPM) that integrates resource…
arXiv cs.AI
Emission-Aware Reinforcement Learning for Sustainable Electric Vehicle Charging and Carbon Dioxide Reduction Under Varying Renewable Penetration
A new study proposes an emission-aware reinforcement learning strategy for electric vehicle charging. This approach…
arXiv cs.AI
DemoEvolve: Overcoming Sparse Feedback in Agentic Harness Evolution with Demonstrations
The paper introduces DemoEvolve, a method for enhancing agent harness evolution using demonstrations. This approach…
arXiv cs.AI
Hypothesis Generation and Inductive Inference in Children and Language Models
The study explores hypothesis generation and inductive inference in children and language models. It compares how both…
arXiv cs.AI
Reasoning as an Attack Surface: Adaptive Evolutionary CoT Jailbreaks for LLMs
The paper discusses the vulnerabilities of Large Reasoning Models (LRMs) to jailbreak attacks due to their…
arXiv cs.AI
Market Regime Council for Dynamic Credit Assignment in Multi-Agent LLM Decision Systems
The article presents a new cooperative multi-agent decision system called Market Regime Council (MRC) for portfolio…
arXiv cs.AI
TIGER: Text-Informed Generalized Enzyme-Reaction Retrieval
The article presents TIGER, a framework designed for enzyme-reaction retrieval in computational biology. It addresses…
arXiv cs.AI
AgentFugue: Agent Scaling for Long-Horizon Tasks through Collective Reasoning
The paper introduces AgentFugue, a framework designed for scaling agent capabilities in long-horizon tasks through…
arXiv cs.AI
SPACE: Unifying Symmetric and Asymmetric Routing Problems for Generalist Neural Solver
The paper titled 'SPACE: Unifying Symmetric and Asymmetric Routing Problems for Generalist Neural Solver' presents a…
arXiv cs.AI
SAM: State-Adaptive Memory for Long-Horizon Reasoning Agent
The article introduces State-Adaptive Memory (SAM), a framework designed for long-horizon reasoning in artificial…
arXiv cs.AI
Benchmarking the Limits of In-Context Reinforcement Learning for Ad-Hoc Teamwork
The paper explores the limitations of In-Context Reinforcement Learning (ICRL) in the context of Ad-Hoc Teamwork…
arXiv cs.AI
JT-SAFE-V2: Safety-by-Design Foundation Model with World-Context Data
The article introduces JT-Safe-V2, a large language model aimed at enhancing the safety and trustworthiness of…
arXiv cs.AI
The Model Is Not the Product: A Dual-Pillar Architecture for Local-First Psychological Coaching
The paper introduces Psych LM, an iOS application designed for psychological coaching using a local-first…
arXiv cs.AI
Advancing Graph Few-Shot Learning via In-Context Learning
The article discusses advancements in graph few-shot learning through a novel model called VISION. This model…
arXiv cs.AI
ConceptM$^3$oE: Concept-Guided Multimodal Mixture of Experts for Interpretable Computational Pathology
The article introduces ConceptM$^3$oE, a new framework for computational pathology that integrates multimodal…
arXiv cs.AI
Understanding and Mitigating Premature Confidence for Better LLM Reasoning
The paper discusses the issue of premature confidence in language models, which leads to flawed reasoning. It…
arXiv cs.AI
A governance horizon for ethical-use constraints in open-weight AI models
The paper discusses ethical-use constraints in open-weight AI models and their implications for governance policy. It…
arXiv cs.AI
Distilling Game Code World Model Generation into Lightweight Large Language Models
The paper discusses the generation of Game Code World Models (GameCWMs) using Large Language Models (LLMs). It…
arXiv cs.AI
Partner-Aware Hierarchical Skill Discovery for Robust Human-AI Collaboration
The article discusses a new framework called Partner-Aware Skill Discovery (PASD) designed to enhance human-AI…
arXiv cs.AI
Adaptive Human-AI Coordination via Hierarchical Action Disentanglement
The paper presents a new framework for adaptive human-AI coordination called Intrinsic Action Disentanglement (IAD).…
arXiv cs.AI
When Does Synthetic Patent Data Help? Volume-Fidelity Trade-offs in Low-Resource Multi-Label Classification
The study investigates the effectiveness of LLM-generated synthetic data in low-resource multi-label patent…
arXiv cs.AI
Safety-Oriented Routing Analysis of Mixtral MoE Under Benign and Harmful Prompts
The paper analyzes the routing behavior of the Mixtral 8x7B-Instruct model under different prompt conditions. It finds…
arXiv cs.AI
Toward Enactive Artificial Intelligence
The paper titled 'Toward Enactive Artificial Intelligence' advocates for integrating enactive approaches to perception…
arXiv cs.AI
How Well Do Models Follow Their Constitutions?
The paper examines how well AI models adhere to their specified behavioral guidelines. It introduces a multi-method…
arXiv cs.AI
Beyond Final Answers: Auditing Trajectory-Level Hallucinations in Multi-Agent Industrial Workflows
The paper discusses the limitations of current hallucination benchmarks for Large Language Models (LLMs) in…
arXiv cs.AI
Identifying and Mitigating Systemic Measurement Bias in Production LLM Inference Benchmarks
The paper discusses the challenges of measuring performance in Large Language Models (LLMs) as they move into…