arXiv cs.AI
From Prompt to Service: An SLM-Based Agent Orchestration Gateway for AI-Driven Virtual Worlds
The paper introduces an SLM-based Agent Orchestration Gateway designed for AI-driven virtual worlds. This gateway…
Hub / Ai Research
ai-research · WeSearch
Ai Research news.
Page 5 of Ai Research headlines on WeSearch — deduped and updated continuously from 10+ editorial sources.
arXiv cs.AI
Cross-Lingual Token Arbitrage: Optimizing Code Agent Context Windows via Local LLM Preprocessing
The paper discusses a new approach to optimize coding agents by reducing input-token costs. It introduces a middleware…
arXiv cs.AI
Bridging Auxiliary Constraints to Resolve Instruction Following in Large Reasoning Models
The paper introduces a framework to improve instruction following in Large Reasoning Models (LRMs) by addressing the…
arXiv cs.AI
TSQAgent: Rating Time Series Data Quality via Dedicated Agentic Reasoning
The paper introduces TSQAgent, a framework designed to improve the assessment of time series data quality using large…
arXiv cs.AI
Gender-Dependent Diagnostic Substitution in LLM Medical Triage: Same Symptoms, Unequal Urgency
A recent study investigates gender-dependent disparities in medical triage recommendations made by large language…
arXiv cs.AI
Towards Non-Monotonic Entailment in Propositional Defeasible Standpoint Logic
The paper discusses advancements in propositional defeasible standpoint logic, focusing on non-monotonic entailment.…
arXiv cs.AI
Diagnosing Knowledge Gaps in LLM Tool Use: An Agentic Benchmark for Novel API Acquisition
The paper introduces NovelAPIBench, a dynamic benchmark designed to evaluate large language models' ability to use…
arXiv cs.AI
From Answers to States: Verifiable Process-Level Evaluation of Chemical Reasoning in Large Language Models
The paper introduces ChemCoTBench-V2, a benchmark designed for evaluating chemical reasoning in large language models.…
arXiv cs.AI
EvoDrive: Pareto Evolution for Safety-Critical Autonomous Driving via Self-Improving LLM Agents
EvoDrive is a new framework designed for generating safety-critical scenarios in autonomous driving systems. It…
arXiv cs.AI
The DeepSpeak-Agentic Dataset
The DeepSpeak-Agentic dataset consists of over 37 hours of semi-structured conversations between humans and AI agents.…
arXiv cs.AI
SkillPyramid: A Hierarchical Skill Consolidation Framework for Self-Evolving Agents
The article introduces SkillPyramid, a framework designed to enhance the skill consolidation of self-evolving AI…
arXiv cs.AI
Dynamic Objective Selection with Safeguards and LLM Oversight for Financial Decision-Making
The paper introduces Dynamic Objective Selection with Safeguards (DOSS) for financial decision-making. DOSS aims to…
arXiv cs.AI
Code-on-Graph: Iterative Programmatic Reasoning via Large Language Models on Knowledge Graphs
The paper introduces Code-on-Graph (CoG), a new framework for integrating Large Language Models (LLMs) with Knowledge…
arXiv cs.AI
Unveiling the Structure of Do-Calculus Reasoning via Derivation Graphs
The paper introduces derivation graphs to enhance the understanding of do-calculus reasoning. These graphs help in…
arXiv cs.AI
When to Re-Plan: Subgoal Persistence in Hierarchical Latent Reasoning
The paper titled 'When to Re-Plan: Subgoal Persistence in Hierarchical Latent Reasoning' explores the balance between…
arXiv cs.AI
Proof-Refactor: Refactoring Generated Formal Proofs into Modular Artifacts
The paper titled 'Proof-Refactor' addresses the challenges in generating formal proofs using Large Language Models…
arXiv cs.AI
LAP: An Agent-to-Instrument Protocol for Autonomous Science
The article introduces the Lab Agent Protocol (LAP), designed to enhance the interaction between autonomous agents and…
Google News
How AI is Transforming Scientific Discovery While Keeping Humans at the Center - Stanford HAI
Comprehensive up-to-date news coverage, aggregated from sources all over the world by Google News.
arXiv cs.AI
BrickAnything: Geometry-Conditioned Buildable Brick Generation with Structure-Aware Tokenization
The article presents BrickAnything, a new framework for generating buildable brick structures from 3D shapes. This…
arXiv cs.AI
Can LLMs Introspect? A Reality Check
The paper titled 'Can LLMs Introspect? A Reality Check' questions the ability of large language models (LLMs) to…
arXiv cs.AI
Is Agent Memory a Database? Rethinking Data Foundations for Long-Term AI Agent Memory
The paper discusses the need for persistent memory in long-running AI agents. It critiques current memory systems and…
arXiv cs.AI
Personalizing Embodied Multimodal Large Language Model Agents over Long-term User Interactions
The paper presents POLAR, a framework designed for personalizing embodied multimodal large language model agents…
arXiv cs.AI
Constraint acquisition needs better benchmarks
The paper discusses the need for improved benchmarks in Constraint Acquisition (CA) research. Current benchmarks are…
arXiv cs.AI
Your Agents Are Aging Too: Agent Lifespan Engineering for Deployed Systems
The paper discusses the aging of AI agents deployed in operational systems and introduces a new benchmark called…
arXiv cs.AI
Experiments in Agentic AI for Science
The paper discusses two innovative frameworks for creating autonomous AI systems to enhance scientific workflows.…
arXiv cs.AI
Anchor: Mitigating Artifact Drift in Agent Benchmark Generation
The paper introduces Anchor, a task-generation pipeline designed to address artifact drift in AI agent benchmark…
arXiv cs.AI
OmniToM: Benchmarking Theory of Mind in LLMs via Explicit Belief Modeling
The paper introduces OmniToM, a benchmark designed to evaluate the Theory of Mind capabilities in large language…
arXiv cs.AI
JobBench: Aligning Agent Work With Human Will
The paper introduces JobBench, a new benchmark for evaluating AI agents based on human needs rather than economic…
arXiv cs.AI
Managing Uncertainty in LLM-Generated Procedural Knowledge for Virtual Laboratory Planning
The paper discusses a framework for managing uncertainty in procedural knowledge generated by large language models…
arXiv cs.AI
ScientistOne: Towards Human-Level Autonomous Research via Chain-of-Evidence
The paper titled 'ScientistOne: Towards Human-Level Autonomous Research via Chain-of-Evidence' presents a new…
arXiv cs.AI
Automatic Layer Selection for Hallucination Detection
A new study proposes an automated method for selecting layers in large language models to improve hallucination…
arXiv cs.AI
Exploiting Local Dynamics Regularity for Reusable Skills in Offline Hierarchical RL
The paper discusses advancements in Hierarchical Reinforcement Learning (HRL) by focusing on the reuse of skills…
arXiv cs.AI
Advancing Creative Physical Intelligence in Large Multimodal Models
A new paper introduces MM-CreativityBench, a benchmark designed to evaluate creative problem-solving in large…
arXiv cs.AI
From Static Context to Calibrated Interactive RL: Mitigating Distribution Shift in Multi-turn Dialogue with Aligned Simulator
The article discusses a new approach to improving dialogue agents through a method called Calibrated Interactive RL.…
arXiv cs.AI
Reasoning, Code, or Both? How Large Language Models Handle Variations in Math Questions
A recent study evaluates how Large Language Models (LLMs) perform on mathematical reasoning tasks when faced with…
arXiv cs.AI
The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence
The MiniMax-M2 series introduces a new family of Mixture-of-Experts language models. These models leverage mini…
arXiv cs.AI
Which Changes Matter? Towards Trustworthy Legal AI via Relevance-Sensitive Evaluation and Solver-Grounded Reasoning
The paper discusses the importance of distinguishing legally relevant changes in legal AI systems. It introduces a new…
arXiv cs.AI
PolyFusionAgent: A Multimodal Foundation Model and Autonomous AI Assistant for Polymer Property Prediction and Inverse Design
PolyFusionAgent is a new multimodal foundation model designed to enhance polymer property prediction and inverse…
arXiv cs.AI
MobileExplorer: Accelerating On-Device Inference for Mobile GUI Agents via Online Exploration
MobileExplorer is a new framework designed to enhance on-device inference for mobile GUI agents. It aims to reduce…
arXiv cs.AI
MedGuideX: Internalizing Decision Logic from Executable Guidelines into Large Language Models for Clinical Reasoning
The article discusses MedGuideX, a new approach to integrating clinical practice guidelines into large language models…
arXiv cs.AI
AGORA: Adapter-Grounded Observation-Action Retention for Inference-Free Prompt Compression in LLM Agents
The article discusses a new method called AGORA for improving prompt compression in large language model (LLM) agents.…
arXiv cs.AI
FAST-GOAL: Fast and Efficient Global-local Object Alignment Learning
The paper presents FAST-GOAL, a method designed to improve the performance of vision-language models like CLIP when…
arXiv cs.AI
Tail-Aware HiFloat4: W4A4 Post-Training Quantization for Wan2.2
The article discusses a new method called Tail-Aware HiFloat4 for post-training quantization in low-bit text-to-video…
arXiv cs.AI
UnityMAS-O: A General RL Optimization Framework for LLM-Based Multi-Agent Systems
The article introduces UnityMAS-O, a general reinforcement learning optimization framework designed for large language…
arXiv cs.AI
Completion vs Optimality: Policy Gradient in Long-Horizon Cumulative-Damage Problems
The paper discusses long-horizon decision problems characterized by cumulative damage and the challenges faced by…
arXiv cs.AI
MemFail: Stress-Testing Failure Modes of LLM Memory Systems
The paper introduces MemFail, a diagnostic benchmark designed to stress-test the failure modes of memory systems in…
arXiv cs.AI
Mind the Tool Failures: Achieving Synergistic Tool Gains for Medical Agents
The article discusses the challenges faced by medical AI agents when using external tools for diagnosis and treatment.…
arXiv cs.AI
Towards Feedback-to-Plan Decisions for Self-Evolving LLM Agents in CUDA Kernel Generation
The paper discusses advancements in self-evolving large language models (LLMs) for CUDA kernel generation. It…
arXiv cs.AI
It's Not the Capability: Harness Sensitivity Is Non-Monotone Across LLM Agent Tiers
A recent study challenges the assumption that higher-capability LLM models require less structural guidance. The…
arXiv cs.AI
A Dataset of Robot-Patient and Doctor-Patient Medical Dialogues for Spoken Language Processing Tasks
A new dataset named MeDial-Speech has been introduced to enhance spoken language processing in medical consultations.…