March 2026 — page 2 of 2
- New Models Beat Lookahead Bias in Financial Forecasting with Strict Data Cutoffs· AI
- New benchmark reveals AI language models still struggle with women's health guidance· AI
- Topic Models Capture Two Distinct Forms of Meaning—and Task Performance Depends on Which· AI
- New Method Turns LLMs Into Interpretable Text Embeddings While Preserving Safety Features· AI
- AlphaGo at 10: DeepMind Traces Path from Game-Playing AI to AGI Ambitions· Agentic AI
- New Benchmark Tests AI Models' Ability to Imagine Egocentric Futures From Action Sequences· AI
- Study reveals how language model output layers suppress critical training signals· AI Research
- New lightweight system trims long context for faster AI question-answering· AI
- Researchers develop AI system to automatically extract complex drug combinations from medical literature· AI
- New Benchmark Reveals How LLM Judges Develop Systematic Biases and How to Fix Them· AI
- GenAI Tools Won't Solve Qualitative Research in Software Engineering, New Study Warns· AI
- Researchers uncover efficiency-based backdoor attacks on AI GUI agents· AI Security
- New technique helps LLMs extend context windows while preserving core capabilities· AI
- Researchers tackle language interference in multilingual speech AI through targeted distillation· AI
- Study Reveals How Reward Hacking Can Sneak Into AI Models During Training· Agentic AI
- New AI Framework Cuts Hardware Verification Errors by 70% Through Operator-Aware Rule Learning· AI
- Bridging Logic and Language: New AI Pipeline Uncovers Hidden Arguments in Everyday Reasoning· AI
- New Method Generates Custom Evaluation Criteria Tailored to Individual LLM Questions· AI
- Compact Biomedical Language Model Matches Frontier AI for Evidence Verification· AI
- New AI Detection Tool Explains Why It Flags Machine-Generated Text· AI
- First taxonomy of real faults in AI model context protocol systems· AI
- Speech LLMs Can Function Effectively With Only 60% of Decoder Layers· AI
- New Framework Automates Building and Testing Software Across Languages and Platforms· Agentic AI
- Researchers Propose Systematic Framework to Evaluate Software Model Datasets· AI
- AI Framework Bridges Gap Between LLM Code Generation and Scientific Simulation Execution· Agentic AI
- Framework stabilizes AI agents for mapping applications by externalizing knowledge and enforcing protocols· Agentic AI
- Active Learning and AI Transparency Cut Software Debt Detection Labeling Work in Half· Software Engineering
- Machine Learning Extracts Hidden Sensory Details from Reviews to Improve Product Recommendations· AI
- Code Agents Learn to Generate Harder Math Problems for AI Training· Agentic AI
- New Framework Simulates Emergency Room Triage Conversations for AI Training· AI
- New conformal method lets AI agents explore safely in high-stakes settings· Agentic AI
- Study finds widespread deception in shadow APIs claiming to offer premium LLM access· AI
- KDFlow: Fast, Efficient System for Compressing Large Language Models· Agentic AI
- Researchers Release First Large Crime-Domain Dataset for Extracting Information from Law Enforcement Reports· AI
- New LLM Framework Tackles Repository-Wide Code Fixing, Boosting Accuracy by 4%· Agentic AI
- How Energy-Efficient Is AI-Generated Code? A Systematic Review Reveals Major Gaps in Assessment· AI
- AI agent automates discovery of new catalysts for clean energy reactions· AI