AI.info

Latest Research — Page 46

Browse Latest Research on AI.info.

• Fine-R1: Make Multi-modal LLMs Excel in Fine-Grained Visual Recognition by Chain-of-Thought Reasoning
• nnActive: A Framework for Evaluation of Active Learning in 3D Biomedical Segmentation
• Ar2Can: An Architect and an Artist Leveraging a Canvas for Multi-Human Generation
• MUFASA: A Multi-Layer Framework for Slot Attention
• Geometry-Aware Single-Image 4D Synthesis via Dense Trajectory Generation
• Diffusion-Based Electromagnetic Inverse Design of Scattering Structured Media
• BeyondSWE: Can Current Code Agent Survive Beyond Single-Repo Bug Fixing?
• Towards Multi-modal Multi-turn Safety: From Agentic Interaction to Strategic Alignment
• Geometric Data Valuation via Leverage Scores
• Ensembling LLM-Induced Decision Trees for Explainable and Robust Error Detection
• Reliable Active Learning from Unreliable Labels via Neural Collapse Geometry
• Losses that Cook: Topological Optimal Transport for Structured Recipe Generation
• MeanCache: From Instantaneous to Average Velocity for Accelerating Flow Matching Inference
• Benchmarking Overton Pluralism in LLMs
• Fine-Grained Representation for Lane Topology Reasoning
• Risks and Controls for Multi-Agent Systems: an analytical framework for deployment of AI agents across organisational boundaries
• Synthetic Data for Robust Runway Detection
• Soft Quality-Diversity Optimization
• Evolution in Simulation: AI-Agent School with Dual Memory for High-Fidelity Educational Dynamics
• Conformal Thinking: Risk Control for Reasoning on a Compute Budget
• TermGPT: Multi-Level Contrastive Fine-Tuning for Terminology Adaptation in Legal and Financial Domain
• DiscoverLLM: From Executing Intents to Discovering Them
• Finding the Translation Switch: Discovering and Exploiting the Task-Initiation Features in LLMs
• ChartAttack: Testing the Vulnerability of LLMs to Malicious Prompting in Chart Generation
• BuildArena: A Physics-Aligned Interactive Benchmark of LLMs for Engineering Construction
• State-Space Models for Tabular Prior-Data Fitted Networks
• SNAP: Low-Latency Test-Time Adaptation with Sparse Updates
• When In Doubt, Abstain: The Impact of Abstention on Strategic Classification
• Co-Layout: LLM-driven Co-optimization for Interior Layout
• Deep sequence models tend to memorize geometrically; it is unclear why
• Debiasing Diffusion Priors via 3D Attention for Consistent Gaussian Splatting
• Instant Video Models: Universal Adapters for Stabilizing Image-Based Networks
• "World Knowledge" in the Weights: Reading Concept Circuits of Vision Transformers
• Tensor Methods: A Unified and Interpretable Approach for Material Design
• Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference
• StereoSpace: Depth-Free Synthesis of Stereo Geometry via End-to-End Diffusion in a Canonical Space
• SERL: Self-Examining Reinforcement Learning on Open-Domain
• RAG-Safety-Bench: Reliable Evaluation of Retrieval-Augmented LLM Safety
• Think Outside the Policy: In-Context Steered Policy Optimization
• Generative AI Agents for Controllable and Protected Content Creation