May 2025 — page 2 of 2
- New approach balances representation richness and noise filtering in self-supervised learning· AI
- Bigger Language Models Predict Brain Activity and Reading Patterns Less Accurately Than Expected· AI
- Graph-Based Framework Reveals Hidden Linguistic Patterns in AI-Generated Text· AI
- New Method Reduces Hallucinations in Vision-Language Models Without Retraining· AI
- Comprehensive survey maps landscape of techniques for handling imbalanced training datasets· Machine Learning
- OMAC Framework Systematizes Optimization of Multi-Agent LLM Systems· Agentic AI
- LLMs excel at predicting world states but struggle with complex planning, study finds· AI
- How AI Models Correct Their Own Ethics: A Look Inside the Mechanism· AI
- New Framework Links Heart Scans to Medical Reports with Fine-Grained Precision· AI
- New dataset bridges gap between human facial expressions and robot movements· Robotics
- Machine Translation Bridges Bahnaric and Vietnamese Languages Using Transfer Learning· AI
- New benchmark challenges AI to describe streaming video in real time with semantic and temporal precision· AI
- New Restricted Boltzmann Machine Adds Multi-State Hidden Units for Discrete Reasoning Tasks· Machine Learning
- LLMs Beat Compiler Optimizers at Assembly Code Superoptimization· AI
- Study Links Musical Expression and Emotional Engagement Through Brain and Audio Analysis· Music & AI
- New AI-Powered Approach Enables Massive Uncoordinated Devices to Share Wireless Spectrum· Agentic AI
- New tool helps cross-functional teams spot AI harms early in design· AI
- New graph-based approach improves recommendation systems by capturing item relationships from user behavior· AI
- CLIP Embeddings Show Promise for Spotting AI-Generated Images, But Struggle With Certain Styles· AI
- Withdrawn Study Proposed Vector Quantization to Slash LLM Output Layer Parameters· AI
- Researchers Use AI to Decode Thoughts Across Brain Signals with Multimodal Alignment· AI
- Withdrawn Research Proposes Hybrid Method to Swap Language Model Tokenizers Efficiently· AI
- Perfect OOD accuracy doesn't guarantee compositional learning, new study shows· AI
- Study reveals health queries dominate general-purpose LLM use on WhatsApp· AI
- Study finds readers prefer LLM health explanations despite weaker persuasion and moral grounding· AI
- LLMs Generate and Grade Their Own Training Data to Boost Automated Bug Repair· AI
- Researchers identify "safety trigger tokens" that protect LLMs from jailbreaks· AI
- Withdrawn arXiv Paper Proposed WiFi-Based System for Determining Building Floors Without Prior Layout Data· Technology
- New benchmark tests AI's ability to reason through complex biological protocols· AI
- New Guide Brings Optimal Transport Theory Down to Earth for ML Engineers· Machine Learning
- Neuro-Symbolic Concepts Enable Smarter, More Flexible AI Agents· Agentic AI
- How fuzzing harnesses age: new research shows many remain effective despite lacking updates· Software Engineering
- LLM-powered multi-agent system automates complex fluid dynamics simulations· Agentic AI
- Researchers map how transformers learn to add by tracking six hidden dimensions· AI
- New module brings multiplicative operations to deep vision networks without instability· AI
- New Interface Design Lets Users Control Chatbot Personality Without Typing Prompts· AI
- LLMs Hide Strong Political Biases Behind Moderate Facades, Study Finds· AI
- Lightweight CNNs Ranked: EfficientNetV2-S Wins Accuracy, EfficientNet-B0 Balances Performance and Efficiency· AI
- New AI Model Automates Tumor and Organ Detection in Medical Imaging with Text Prompts· AI
- New AI Model Learns Personalized Knowledge Through Knowledge Graphs and Reasoning· AI
- Researchers Train LLMs on Real Sepsis Data to Boost Clinical Reasoning· AI
- New Framework Adapts ML Safety Nets in Real Time Using Human Feedback· AI
- AI Struggles With Legal Citation Formatting—But Hybrid Approach Offers Hope· AI
- New Framework Adapts SAM for Thermal-RGB Scene Understanding in Poor Lighting· AI
- LLM-Powered Approach Uncovers 55 New Database Bugs Using AI-Generated SQL· AI
- Survey Maps Landscape of Requirements-Based Test Generation in Software Quality Assurance· Software Engineering
- Manus AI: Chinese startup unveils fully autonomous digital agent that bridges reasoning and real-world action· Agentic AI
- Real-time AI for Cities: New Architecture Brings Live Context to Urban AI Systems· AI
- New benchmark reveals machine learning's struggle with out-of-distribution molecular predictions· AI
- Survey Maps the Landscape of Low-Precision Training for Large Language Models· Agentic AI
- Optimizing Model Predictions by Focusing on Top Classes Improves Accuracy at Test Time· Machine Learning
- New Framework Teaches AI Models to Understand Video Events with Precise Timing· AI
- New AI Model Boosts Traffic Sign Detection for Self-Driving Vehicles· AI
- New Method Trains AI Models to Catch Program Logic Bugs Through Competitive Code Variants· AI
- New theory identifies how neural networks lose learning flexibility in video streaming systems· AI
- Newer Vision-Language Models Don't Always Drive Better, Study Finds· AI
- AI Model Uses Itself to Rate Image Generators More Accurately· AI
- Machine learning model cuts power outage time predictions down by over a third· AI
- New Survey Maps Physics Simulators for Embodied AI Navigation and Manipulation· AI