AI.info

Latest Research — Page 101

Browse Latest Research on AI.info.

• AMaPO: Adaptive Margin-attached Preference Optimization for Language Model Alignment
• Ground Truth First: A Longitudinal Evaluation Instrument for Agent Memory, and the Tenure Crossover in Memory-Architecture Rankings
• Leaner Transformers Can Easily Learn to Cluster
• TailSteer: Trajectory-Tail Steering for Drift-Resistant Multi-turn Image Editing
• RGBD20K: A Large-Scale Benchmark for RGB-D Semantic Segmentation
• When Refusals Fail: Unstable Safety Mechanisms in Long-Context LLM Agents
• AIFloodSense: A Global Aerial Imagery Dataset for Semantic Segmentation and Understanding of Flooded Environments
• A systematic review of relation extraction task since the emergence of Transformers
• AReA: Adaptive Receptive-Field Attention for Efficient High-Resolution Diffusion Transformers
• CRAFTIIF: Cross-Resolution Analytic Four-Type Interpretable Isolation Forest for Multivariate Time Series Anomaly Detection
• Neural Texture Splatting: Expressive 3D Gaussian Splatting for View Synthesis, Geometry, and Dynamic Reconstruction
• An evaluation of SVBRDF Prediction from Generative Image Models for Appearance Modeling of 3D Scenes
• In-Flight KV Cache with Clean Anchors for Faster Autoregressive Video Diffusion
• Probing RLVR training instability through the lens of objective-level hacking
• SAM 3D Body: Robust Full-Body Human Mesh Recovery
• Ladders in Chaos: When, How, (and Perhaps Why) Does Test-Time Scaling Improve LLM Machine Translation
• Finding Culture-Sensitive Neurons in Vision-Language Models
• MerLin: A Discovery Engine for Photonic and Hybrid Quantum Machine Learning
• LEVIO: Lightweight Embedded Visual Inertial Odometry for Resource-Constrained Devices
• CrossVid: A Comprehensive Benchmark for Evaluating Cross-Video Reasoning in Multimodal Large Language Models
• REFLEX: Self-Refining Explainable Fact-Checking via Verdict-Anchored Style Control
• VideoGPA: Distilling Geometry Priors for 3D-Consistent Video Generation
• Tackling Resource-Constrained and Data-Heterogeneity in Federated Learning with Double-Weight Sparse Pack
• UltraHR-100K: Enhancing UHR Image Synthesis with A Large-Scale High-Quality Dataset
• ActionMesh: Animated 3D Mesh Generation with Temporal 3D Diffusion
• VQ-Seg: Vector-Quantized Token Perturbation for Semi-Supervised Medical Image Segmentation
• Socially Grounded Agentic AI: Coordinating Plural Perspectives through Social Theory
• M3SR: Multi-Scale Multi-Perceptual Mamba for Efficient Spectral Reconstruction
• SceneCOT: Eliciting Grounded Chain-of-Thought Reasoning in 3D Scenes
• GEO-Bench-2: From Performance to Capability, Rethinking Evaluation in Geospatial AI
• StepShield: When, Not Whether to Intervene on Rogue Agents
• Agensh: Scaling Organizational Intelligence to 1,024 Agents
• Beyond Future Prediction: Denoising as Generative Adaptation for Robot Control
• Thinking-in-Bets at Test Time Reduces Sycophantic Revisions in LLM Agents
• Identifying and Mitigating Bottlenecks in Role-Playing Agents: A Systematic Study of Disentangling Character Profile Axes
• Showing versus Doing: Teaching by Demonstration
• Seismology modeling agent: A smart assistant for geophysical researchers
• Vote-in-Context: Turning VLMs into Zero-Shot Rank Fusers
• Particle Monte Carlo methods for Lattice Field Theory
• Understanding the Generalization of Stochastic Gradient Adam in Learning Neural Networks