AI.info

Latest Research — Page 51

Browse Latest Research on AI.info.

• SPG: Sandwiched Policy Gradient for Masked Diffusion Language Models
• MiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement
• Efficient Discriminative Joint Encoders for Large Scale Vision-Language Reranking
• EduMod-LLM: A Modular Approach for Designing Flexible and Transparent Educational Assistants
• Reinforcement Learning with Backtracking Feedback
• Harnessing Textual Refusal Directions for Multimodal Safety
• System1-Arena: Which Decisions Can Be Left to System 1?
• DiA-gnostic VLVAE: Disentangled Alignment-Constrained Vision Language Variational AutoEncoder for Robust Radiology Reporting with Missing Modalities
• Clinician-in-the-Loop Smart Home System to Detect Urinary Tract Infection Flare-Ups via Uncertainty-Aware Decision Support
• LLM Swiss Round: Aggregating Multi-Benchmark Performance via Competitive Swiss-System Dynamics
• CellPath-Bench: A Multidimensional Benchmark for Whole-Slide Cellular Representations in Pathology Foundation Models
• Improving Decision Trees through the Lens of Parameterized Local Search
• DriveLaW:Unifying Planning and Video Generation in a Latent Driving World
• Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond)
• RoleMotion: A Large-Scale Dataset towards Robust Scene-Specific Role-Playing Motion Synthesis with Fine-grained Descriptions
• Sparse Orthogonal Regression Technique: A Spectral Framework for Equation Discovery, Approximation, and Integration
• Continual Learning Mechanisms Compose for Long-Horizon Memorization
• OmniXRef: A Unified Framework for Multimodal Multi-Reference Video Editing
• Is Your Long-Horizon Benchmark Really Deep? Measuring and Synthesizing Depth in Terminal Tasks
• Differential 6-DOF Pose Estimation with Provable First-Order Immunity to Camera Calibration Errors
• Self-Evaluating LLMs for Multi-Step Tasks: Stepwise Confidence Estimation for Failure Detection
• Instant Personalized Large Language Model Adaptation via Hypernetwork
• Opt-ICL at LeWiDi-2025: Maximizing In-Context Signal from Rater Examples via Meta-Learning
• A*-Thought-V2: Efficient Latent Reasoning via Geometric Dynamics of LLM
• Think Less, Locally: Local Preference Learning for Reasoning Compression
• GloCTM: Cross-Lingual Topic Modeling via a Global Context Space
• DINOde: Continuous Vision-Text Alignment for Open-Vocabulary Semantic Segmentation
• Where-OPD: Spatially Guided On-Policy Self-Distillation of MLLMs with Synthetic Scenes
• All Changes May Have Invariant Principles: Improving Ever-Shifting Harmful Meme Detection via Design Concept Reproduction
• Discovering Differences in Strategic Behavior Between Humans and LLMs
• Memory as Weights: Internalizing Long-Term History for Streaming Videos
• Sparse Internal Control of Language Models
• Pistachio: Towards Synthetic, Balanced, and Long-Form Video Anomaly Benchmarks
• Persistent Planar Memory for Video World Models
• GeoDiff: Geometry-Guided Diffusion for Metric Depth Estimation
• GeoBenchLLM: A Comprehensive Benchmark for Evaluating LLMs on Geo-Related Tasks
• Does an Agent's History Tell You When Compaction Will Hurt? A Modest, Bounded Effect on the TRACE Paired-Replay Corpus
• Stress-Testing Emotional Support Models: Moving from Homogeneous to Diverse Help Seekers
• Text2Loc++: Generalizing 3D Point Cloud Localization from Natural Language
• Unsupervised Ensemble Learning Through Deep Energy-based Models