AI.info

Latest Research — Page 105

Browse Latest Research on AI.info.

• Multi-modal On-Device Learning for Monocular Depth Estimation on Ultra-low-power MCUs
• HiFi-Portrait: Zero-shot Identity-preserved Portrait Generation with High-fidelity Multi-face Fusion
• Complete or Sparse: A Tale of Two Identifiabilities
• SONIC: Spectral Optimization of Noise for Inpainting with Consistency
• Few-Step Generation via Data-Space Iteration
• The 2026 PNPL Competition: Word Classification and Efficient Cross-Subject Generalisation in LibriBrain100
• Large-Scale Terminal Agentic Trajectory Generation from Dockerized Environments
• Learning Bug Context for PyTorch-to-JAX Translation with LLMs
• Self-Generated Feedback Destabilizes Test-Time Training: A Causal Decomposition of Long-Horizon Adaptation
• Domain Expansion: A Latent Space Construction Framework for Multi-Task Learning
• Exposing DeepFakes via Hyperspectral Domain Mapping
• Generating Sketches in a Hierarchical Auto-Regressive Process for Flexible Sketch Drawing Manipulation at Stroke-Level
• How Many Tokens Do 3D Point Cloud Transformer Architectures Really Need?
• Text images processing system using artificial intelligence models
• IdeaTaste: Do LLMs Share the Research Taste of Human Researchers?
• UniMate: One Unified Model to Animate Diverse Skeletons
• Benchmarking Document Parsers on Mathematical Formula Extraction from PDFs
• ELIQ: A Label-Free Framework for Quality Assessment of Evolving AI-Generated Images
• Bee: A High-Quality Corpus and Full-Stack Suite to Unlock Advanced Fully Open MLLMs
• Roleplaying with Structure: Synthetic Therapist-Client Conversation Generation from Questionnaires
• Opera: A Verbal Critic Framework for Long-horizon Coding Agents
• Right Tool, Right Job: Native-Language Evaluation, Tokenizer Sensitivity, and Methodological Findings from a French-Only BabyLM
• Mitigating Mismatch within Reference-based Preference Optimization
• BigCell: Generating Gigapixel Whole-Slide Images
• Adapting Noise to Data: Generative Flows from 1D Processes
• Dropouts in Confidence: Moral Uncertainty in Human-LLM Alignment
• Adaptive Active Learning for Regression via Reinforcement Learning
• Hybrid Gaussian Splatting for Novel Urban View Synthesis
• TurkColBERT: A Benchmark of Dense and Late-Interaction Models for Turkish Information Retrieval
• SPARTA: Scalable and Principled Benchmark of Tree-Structured Multi-hop QA over Text and Tables
• XR Design Framework for Early Childhood Education
• LongVT: Incentivizing "Thinking with Long Videos" via Native Tool Calling
• HD$^2$-SSC: High-Dimension High-Density Semantic Scene Completion for Autonomous Driving
• Agentic Reasoning for Large Language Models
• Diffusion Meta-Prompting and Steering for Generalizable Foundation Model Adaptation
• Lingua-SafetyBench: A Benchmark for Safety Evaluation of Multilingual Vision-Language Models
• BurstDeflicker: A Benchmark Dataset for Flicker Removal in Dynamic Scenes
• PandaPose: 3D Human Pose Lifting from a Single Image via Propagating 2D Pose Prior to 3D Anchor Space
• RCScore: Quantifying Response Consistency in Large Language Models
• Memorize-and-Generate: Towards Long-Term Consistency in Real-Time Video Generation