AI.info

Latest Research — Page 132

Browse Latest Research on AI.info.

• Gated KalmaNet: A Fading Memory Layer Through Test-Time Ridge Regression
• Fidelity Is Not Enough: Dispatch-Level Instrumentation for Agentic Datasheet Extraction
• Lloyd's $K$-Means Clustering Algorithm Is Frank-Wolfe in Disguise
• scPilot: Large Language Model Reasoning Toward Automated Single-Cell Analysis and Discovery
• Revisiting Logit Distributions for Reliable Out-of-Distribution Detection
• Video Reasoning without Training
• Health-ORSC-Bench: A Benchmark for Measuring Over-Refusal and Safety Completion in Health Context
• Whatever Remains Must Be True: Filtering Drives Reasoning in LLMs, Shaping Diversity
• Skullptor: High Fidelity 3D Head Reconstruction in Seconds with Multi-View Normal Prediction
• Rare Event Analysis of Large Language Models
• VVS: Accelerating Speculative Decoding for Visual Autoregressive Generation via Partial Verification Skipping
• ZACH-ViT: A Zero-Token Vision Transformer with ShuffleStrides Data Augmentation for Robust Lung Ultrasound Classification
• MAPLE-Guard: Memory-Aware Link Enforcement Against Memory-Link Poisoning in Multi-Agent Systems
• WebChain: A Large-Scale Human-Annotated Dataset of Real-World Web Interaction Traces
• FluxMem: Adaptive Hierarchical Memory for Streaming Video Understanding
• A Mechanistic Perspective and Circuit-Guided Difficulty Metric for Unlearning
• BATON: A Multimodal Benchmark for Bidirectional Automation Transition Observation in Naturalistic Driving
• Multi-Stage Evolutionary Model Merging with Meta Data Driven Curriculum Learning for Sentiment-Specialized Large Language Modeling
• Conformal Sparsification for Bandwidth-Efficient Edge-Cloud Speculative Decoding
• NoLan: Mitigating Object Hallucinations in Large Vision-Language Models via Dynamic Suppression of Language Priors
• RAG-3DSG: Enhancing 3D Scene Graphs with Re-Shot Guided Retrieval-Augmented Generation
• One-Step Generative Policies with Q-Learning: A Reformulation of MeanFlow
• Windsock is Dancing: Adaptive Multimodal Retrieval-Augmented Generation
• Real-Time Robot Execution with Masked Action Chunking
• Topology and Geometry of the Learning Space of ReLU Networks: Connectivity and Singularities
• CogVis: Must Open-Vocabulary Change Detection Perceive the Scene Anew for Every Query?
• The N-Body Problem: Parallel Execution from Single-Person Egocentric Video
• Reinforced Efficient Reasoning via Semantically Diverse Exploration
• Agent-G$^2$: Gaussian Guidance for Agentic Reinforcement Learning
• ARCHER: Agentic Rule and Compliance Harness for Executable Regulations
• OpenDecoder: Open Large Language Model Decoding to Incorporate Document Quality in RAG
• One Life to Learn: Inferring Symbolic World Models for Stochastic Environments from Unguided Exploration
• STORM: Segment, Track, and Object Re-Localization from a Single Image
• Connecting Domains and Contrasting Samples: A Ladder for Domain Generalization
• CFG-Ctrl: Control-Based Classifier-Free Diffusion Guidance
• Unplugging a Seemingly Sentient Machine Is the Rational Choice -- A Metaphysical Perspective
• CARES: Context-Aware Resolution Selector for VLMs
• Expectation, Backlash, Recovery, and Excitement: How Model Releases Shape Reddit Perceptions of Conversational AI Systems
• Transformer-Based Inpainting for Real-Time 3D Streaming in Sparse Multi-Camera Setups
• When to Invoke: Refining LLM Fairness with Toxicity Assessment