January 2026 — page 2 of 2
- New Framework Teaches AI Agents to Plan by Imagining Future Scenarios· Agentic AI
- Researchers Measure How Competing Narratives Polarize Online Discourse· AI, Research
- Calibration alone misses the mark: New framework reveals blind spots in how LLMs estimate confidence· AI
- New benchmark exposes how language models struggle with non-Western kinship reasoning· AI
- New Benchmark Reveals Major Gaps in LLM Temporal Reasoning, Even With Web Search· AI
- New Model Diagnoses Heart Conditions Accurately With 95% Fewer Parameters· Machine Learning
- New method makes neural network training faster by smartly choosing which images to label· AI
- LLMs Generate Interpretable Explanations for Fine-Grained Social Media Sentiment Analysis· AI
- Major multimodal AI models still struggle with basic visual tasks that toddlers solve easily· AI
- New framework verifies medical AI claims against research evidence· AI
- How LLMs Geometrically Reshape Truth Representations When Given Context· AI
- New framework personalizes AI language models while keeping user data private· AI
- New Benchmark Measures How Research AI Agents Learn From User Feedback· AI
- New Dataset Links Student Writing and Feedback to Benchmark AI Assessment Tools· AI
- New Framework Calibrates NLP Models to Respect Diverse Human Values in Subjective Tasks· Agentic AI
- New Fine-Tuning Method Hardens Code Summarization Against Poorly Written Code· AI
- Rigorous Design Principles Unlock Better Multilingual Named Entity Recognition· AI
- RingSQL bridges synthetic and real data for training SQL-generation models· AI
- New Benchmark Tackles Predicting Clinical Trial Eligibility Criteria Changes· AI
- New AI Framework Maps Ancient Chinese History into Structured Event Taxonomy· AI
- New Dataset and Framework Tackle Evolving Safety Risks in Conversational AI with Images· AI
- Researchers propose privacy-aware machine text detection using noise injection· AI
- New LLM-based method tackles entity linking without domain-specific training· AI
- Researchers Use Sparse Autoencoders to Steer Large Language Models' Reasoning Strategies· AI
- New Evaluation Framework Tests How Well AI Handles Long Legal Documents· AI
- New Method Detects LLM Hallucinations Using Single Response and Token Probabilities· AI
- Transformer models show consistent layer-by-layer positional biases that favor different word types· AI
- New clipping technique stabilizes reinforcement learning for large language models· Agentic AI
- InfiniteWeb Automatically Generates Training Environments for AI GUI Agents· AI
- EZInput Brings Automatic UI Generation to Python Research Tools· Software Engineering
- Researchers map how language models learn and forget concepts during continued training· AI
- Researchers teach AI to spot and fix moral mistakes instead of just mimicking acceptable language· AI
- Metaphors in Training Data Drive Cross-Domain Misalignment in Large Language Models· AI Safety
- LLMs Now Generate and Debug Hardware Design Tests at Scale· AI
- New benchmark tests AI's ability to reason about economics, timelines, and numbers in long documents· AI
- New benchmark pushes medical AI to pinpoint exact phrases behind clinical summaries· AI
- Targeted Forgetting: New Method Erases Sensitive AI Knowledge Without Harming Overall Performance· AI
- New Framework Teaches LLMs to Chat More Like Humans Through Targeted Alignment· Agentic AI
- AI System Learns to Write Dictionary Definitions Optimized for Language Learners· AI
- Digital Twins Boost Federated Learning for Industrial Anomaly Detection· AI
- New framework enables LLM agents to autonomously manage memory for extended reasoning tasks· Agentic AI
- New Technique Improves Security Bug Detection by Focusing on Relevant Code Features· AI
- How Information Placement and Safety Prompts Undermine Long-Context LLM Reliability· AI
- New benchmark helps developers evaluate AI-generated web applications fairly· AI
- LLMs Are Data Compressors, Not Optimal Predictors—The Road to AI Singularity Needs Symbolic Reasoning· AI
- ELF Models Simplify ECG Interpretation Without Dedicated Encoders· AI
- New Framework Gives Language Models Persistent State Management for Complex Tasks· Agentic AI
- LACONIC bridges the gap: sparse retrieval matches dense model performance while cutting memory use· AI
- New Benchmark Reveals How AI Models Handle Mixed Chinese-English Technical Language· AI
- Quantizing Large Language Models Weakens Their Self-Explanations, Study Finds· AI
- AI coding agents contribute meaningful security work but face stricter code review· Agentic AI
- New method optimizes transformer depth through structured residual edits· Agentic AI