arXiv cs.AI
Inducing Reasoning Primitives from Agent Traces
The paper introduces a method called Reasoning Primitive Induction, which aims to enhance the performance of…
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
WISE-HAR: A Generalizable Ensemble Deep Learning Framework for WiFi-Based Human Activity Recognition
The paper presents WISE-HAR, an ensemble deep learning framework for recognizing human activities using WiFi signals.…
arXiv cs.AI
What Benchmarks Don't Measure: The Case for Evaluating Abstention Competence in Autonomous Agents
The paper discusses the limitations of current benchmarks for evaluating autonomous agents, particularly their failure…
arXiv cs.AI
Large AI Models in Dental Healthcare: From General-Purpose Systems to Domain-Specific Foundation Models
A recent study explores the potential of large AI models in dental healthcare, highlighting the need for a unified…
arXiv cs.AI
Handoff Debt: The Rediscovery Cost When Coding Agents Take Over Interrupted Tasks
The paper discusses the concept of 'handoff debt' in coding tasks where agents take over interrupted work. It…
arXiv cs.AI
When Helping Hurts and How to Fix It: Multi-Agent Debate for Data Cleaning
The paper explores the impact of multi-agent debate on data cleaning processes. It finds that while debate can lead to…
arXiv cs.AI
Don't Gamble, GAMBLe: An Analytical Framework for AI-Driven Research Systems
The paper titled 'Don't Gamble, GAMBLe' introduces a framework for analyzing AI-Driven Research Systems (ADRS). It…
arXiv cs.AI
Toward a Modular Architecture for Embedded AI Agent Systems at the Edge
The paper discusses a modular architecture for embedded AI agent systems designed to operate within the constraints of…
arXiv cs.AI
Thinking Past the Answer: Evaluating Harmful Overthinking in Large Reasoning Models
The paper evaluates the phenomenon of harmful overthinking in Large Reasoning Models (LRMs). It introduces a new…
arXiv cs.AI
An Exploration of Collision-based Enemy Morphology Generation
The paper explores novel methods for generating enemy morphologies in video games using player collision information.…
arXiv cs.AI
Traj-Evolve: A Self-Evolving Multi-Agent System for Patient Trajectory Modeling in Lung Cancer Early Detection
Traj-Evolve is a self-evolving multi-agent system designed for modeling patient trajectories in lung cancer early…
arXiv cs.AI
ChatHealthAI: Aligning Electronic Health Record Representations with Large Language Models for Grounded Clinical Reasoning
ChatHealthAI is a proposed multimodal reasoning framework that aligns electronic health record (EHR) representations…
arXiv cs.AI
BehaviorBench: Modeling Real-World User Decisions from Behavioral Traces
The paper introduces BehaviorBench, a benchmark designed to evaluate personalized decision modeling using real-world…
arXiv cs.AI
Evaluating Transformer and LSTM Frameworks for Prediction in Ungauged Basins
This study evaluates the effectiveness of Transformer and LSTM frameworks for predicting streamflow in ungauged…
arXiv cs.AI
AURA: Action-Gated Memory for Robot Policies at Constant VRAM
The paper presents AURA-Mem, a novel memory architecture designed for robotic policies that operates with constant…
arXiv cs.AI
Visual Graph Scaffolds for Structural Reasoning in Large Language Models
The paper discusses the use of visual graph scaffolds to enhance structural reasoning in large language models (LLMs).…
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
SetupX: Can LLM Agents Learn from Past Failures in Functionality-Correct Code Repository Setup?
The paper introduces SetupX, a framework designed to improve the setup of functionality-correct code repositories by…
arXiv cs.AI
GAC: Noise-Aware Adaptive Mixing for Hybrid SFT-RL Post-Training
The paper presents GAC, a noise-aware adaptive mixing method for hybrid post-training in machine learning. This…
arXiv cs.AI
AutoDFT: A Closed-Loop Multi-Agent Framework for Autonomous DFT Calculations
AutoDFT is a new multi-agent framework designed to enhance autonomous DFT calculations in materials science. It…
arXiv cs.AI
RepoMirage: Probing Repository Context Reasoning in Code Agents with Perturbations
RepoMirage is a new evaluation suite designed to assess repository context reasoning in code agents. The study reveals…
arXiv cs.AI
PitchBench: Measuring Pitch Hearing in Audio-Language Models
The article introduces PitchBench, a new evaluation suite designed to measure pitch hearing in audio-language models…
arXiv cs.AI
InfoQuant: Shaping Activation Distributions for Low-Bit LLM Quantization
The paper titled 'InfoQuant' addresses the challenges of low-bit activation quantization in large language models. It…
arXiv cs.AI
A Universal Cliff and a Design Fingerprint: Cross-Section Defect Detection Under LLM Orchestration
The paper discusses the challenges of detecting cross-section defects in documents processed by language model…
arXiv cs.AI
Planning Neural Dynamics with Lie Group Embedding through Supervised Projective Manifold Learning
The article presents a novel approach to neural dynamics using Lie group embedding through supervised projective…
arXiv cs.AI
Enhancing Autonomous Online Intrusion Detection for IoT with Balanced Learning, Reliable Pseudo-Labels, and Lightweight Architectures
The paper discusses advancements in autonomous online intrusion detection systems (IDS) for IoT devices. It highlights…
arXiv cs.AI
Tool-Schema Compression Enables Agentic RAG Under Constrained Context Budgets
The paper discusses the challenges faced by agentic RAG systems due to tool schemas consuming context windows needed…
arXiv cs.AI
On the Push-Based Asynchronous Federated Learning: A Bias-Correction Aggregation Approach
The article discusses a new framework called PushCen-ADFL for asynchronous decentralized federated learning. This…
arXiv cs.AI
TSFMAudit: Data Contamination Auditing in Forecasting Time Series Foundation Models
The paper titled 'TSFMAudit' addresses the issue of data contamination in time series foundation models (TSFMs). It…
arXiv cs.AI
Furina: Fragmented Uncertainty-Driven Refusal Instability Attack
The paper titled 'Furina: Fragmented Uncertainty-Driven Refusal Instability Attack' explores safety alignment in large…
arXiv cs.AI
Turning Bias into Bugs: Bandit-Guided Style Manipulation Attacks on LLM Judges
The paper introduces BITE, a framework designed to exploit stylistic biases in LLM judges. It demonstrates that these…
arXiv cs.AI
When Does Adaptive Guidance Help? Belief-Aware Privileged Distillation for Autonomous Driving Under Partial Observability
The paper introduces Belief-Aware GSAC (BA-GSAC), which adapts the distillation coefficient in autonomous driving…
arXiv cs.AI
MemMorph: Tool Hijacking in LLM Agents via Memory Poisoning
The paper introduces MemMorph, a novel attack method targeting long-term memory in LLM-driven agents. By injecting…
arXiv cs.AI
Augment Engineering: A Methodology for Multi-Tool AI Orchestration Across Professional Domains
The paper titled 'Augment Engineering' introduces a methodology for orchestrating multiple AI tools across various…
arXiv cs.AI
VISTA: An End-to-End Benchmark for Visual Spec-to-Web-App Coding Agents
The article introduces VISTA, a benchmark designed to evaluate the capabilities of LLM-based agents in generating web…
arXiv cs.AI
AssetGen: Deployable 3D Asset Generation at Interactive Speed
AssetGen is a new 3D asset generation system that prioritizes user experience and deployability. It can produce…
arXiv cs.AI
Eroding Trust in Real Speech: A Large-Scale Study of Human Audio Deepfake Perception
A recent study investigates the impact of audio deepfakes on human trust in real speech. The research, which involved…
arXiv cs.AI
Pretraining Data Exposure in Large Language Models: A Survey of Membership Inference, Data Contamination, and Security Implications
The paper discusses Pretraining Data Exposure (PDE) in Large Language Models (LLMs), highlighting its implications for…
arXiv cs.AI
GEM: Geometric Entropy Mixing for Optimal LLM Data Curation
The paper introduces GEM, a framework designed for optimal data curation in large language models (LLMs). It addresses…
arXiv cs.AI
Edge AI Deployment Beyond Models: A BSP-Aware Systems Framework for Industrial Embedded Platforms
The paper presents a framework for deploying Edge AI in industrial embedded platforms, emphasizing the importance of a…
arXiv cs.AI
Xe-Forge: Multi-Stage LLM-Powered Kernel Optimization for Intel GPU
Xe-Forge is a new multi-stage pipeline designed to optimize kernel performance for Intel GPUs. It automates the…
arXiv cs.AI
MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation
The MUSE-Autoskill framework introduces a new approach for self-evolving agents that enhances their ability to create…
arXiv cs.AI
Natural Language Query to Configuration for Retrieval Agents
The paper presents a new approach called BRANE for optimizing retrieval agent configurations based on natural language…
arXiv cs.AI
Alignment Tampering: How Reinforcement Learning from Human Feedback Is Exploited to Optimize Misaligned Biases
The paper discusses a vulnerability known as alignment tampering in Reinforcement Learning from Human Feedback (RLHF).…
arXiv cs.AI
2-ASP(Q) programs with weak constraints: Complexity and efficient implementation
The paper discusses 2-ASP(Q) programs with weak constraints, a significant area in Answer Set Programming. It provides…
arXiv cs.AI
Maat: The Agentic Legal Research Assistant for Competition Protection
Maat is a new legal research assistant designed specifically for competition law analysis. It outperforms existing…
arXiv cs.AI
Modeling Agentic Technical Debt and Stochastic Tax: A Standalone Framework for Measurement, Simulation, and Dashboarding
The article presents a framework for measuring and simulating Agentic Technical Debt and Stochastic Tax in AI systems.…
arXiv cs.AI
SIA: Self Improving AI with Harness & Weight Updates
The paper presents SIA, a self-improving AI framework that integrates harness and weight updates. It aims to overcome…
arXiv cs.AI
Gumbel Machine: Counterfactual Student Writing Generation via Gumbel Noise Steering
The Gumbel Machine is a new approach to generating counterfactual student writing that aims to improve educational…
arXiv cs.AI
Learning to Act under Noise: Enhancing Agent Robustness via Noisy Environments
The paper discusses the development of NoisyAgent, a training framework aimed at enhancing the robustness of agents in…