AI.info

Latest Research — Page 40

Browse Latest Research on AI.info.

• Transformers Stop Thinking Too Early, and a Tiny LoRA Fixes It
• BanglaShield: A Script-Robust Safety Classifier and Benchmark for Bangla Native, Romanized, and Code-Mixed Prompts
• Five Years of SciCap: What We Learned and Future Directions for Scientific Figure Captioning
• WiseEdit: Benchmarking Cognition- and Creativity-Informed Image Editing
• PLANA3R: Zero-shot Metric Planar 3D Reconstruction via Feed-Forward Planar Splatting
• Deconstructing Multi-Task Active Learning: The Paradox of Gradient Conflict and Orthogonal Decomposition
• The Controllability Trap: A Governance Framework for Military AI Agents
• Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training
• Generative Semantic Multi-Object Tracking: A Large-Scale Benchmark and an MLLM-Driven Reasoning Framework
• STARCaster: Spatio-Temporal AutoRegressive Video Diffusion for Identity- and View-Aware Talking Portraits
• Multi-modal Dynamic Proxy Learning for Personalized Multiple Clustering
• STReasoner: Empowering LLMs for Spatio-Temporal Reasoning in Time Series via Spatial-Aware Reinforcement Learning
• A Lightweight Sparse Interaction Network for Time Series Forecasting
• FrontierCS 2: A Benchmark for Computer Science R&D
• Leveraging Foundational Models and Simple Fusion for Multi-modal Physiological Signal Analysis
• AffectSim: A Controllable Interactive 3D Simulation Benchmark for Embodied Affective Perception
• Dream-RSI: Recursive Self-Improvement through Evolving Worlds
• SIEVE: Overcoming Topological Obstruction in Equivariant Self-Supervised Learning
• Understanding Syllogistic Reasoning in LLMs from Formal and Natural Language Perspectives
• Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior
• Re-coding for Uncertainties: Edge-awareness Semantic Concordance for Resilient Event-RGB Segmentation
• Towards Characterizing Knowledge Distillation of PPG Heart Rate Estimation Models
• MJev-Spatial: System One Spatial Intelligence via Native Attention Matching
• HydraLLM: Adaptive Width Slicing for Elastic Text and Vision-Language Models
• DP-AdamW: Investigating Decoupled Weight Decay and Bias Correction in Private Deep Learning
• When AI Democratizes Exploitation: LLM-Assisted Strategic Manipulation of Fair Division Algorithms
• Generative Modeling from Black-box Corruptions via Self-Consistent Stochastic Interpolants
• Instruction Capacity: Behavior and Degradation of Language Models under Controlled Load
• StructXLIP: Enhancing Vision-language Models with Multimodal Structural Cues
• Dynamic Rank Reinforcement Learning for Adaptive Low-Rank Multi-Head Self Attention in Large Language Models
• Human Mesh Modeling for Anny Body
• PreCoMem: Predictive Cognitive Memory for Self-Evolving Long-Term Dialogue Agents
• Stress Testing Chain-of-Thought Monitoring Against Covert Misalignment
• ReForm: Reflective Autoformalization with Prospective Bounded Sequence Optimization
• HalleluBERT: Let Every Token That Has Meaning Bear Its Weight
• Global Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages
• BaRISTA: Brain Scale Informed Spatiotemporal Representation of Human Intracranial Neural Activity
• MOBIUS: Big-to-Mobile Universal Instance Segmentation via Multi-modal Bottleneck Fusion and Calibrated Decoder Pruning
• Reason or Recite: Complementary Specialists with an LLM Judge for Traffic Anomaly Reasoning
• ECG-R1: Protocol-Guided and Modality-Agnostic MLLM for Reliable ECG Interpretation