Science
Machine learning as a scientific instrument: protein structure prediction, materials discovery pipelines, drug design models, climate simulation, and the papers demonstrating AI as a tool for accelerating discovery across scientific disciplines.
AI Overlooks a Key Way People Show Emotion Online
Researchers find that repeated letters and punctuation in social media posts are crucial for sentiment analysis, but many AI models miss their meaning, leading to a new method to improve understanding.
AI Models Learn to Focus on What Matters
A new training-free method helps AI systems identify and prioritize relevant visual and textual evidence, improving accuracy in complex question-answering tasks without any model modifications.
AI Uncovers Hidden Symmetries in Quantum Systems
A new method uses spectral data to reveal hidden symmetries in quantum many-body systems, enabling precise identification of symmetry groups without prior knowledge.
A New Way to Verify Complex Software Piece by Piece
Researchers have developed a method that breaks down software verification into manageable components, enabling reliable checks of large systems without overwhelming complexity.
New AI Methods Tackle 3D Shape Matching Challenges
Researchers categorize advances in aligning deformed 3D shapes into spectral, combinatorial, and deformation-based approaches, highlighting progress in zero-shot and partial matching but noting limitations with real-world data.
AI Optimizer Outperforms Standard Methods in Key Tests
A new algorithm called Sven uses a clever mathematical trick to train neural networks faster and more accurately, especially for regression tasks, while keeping computational costs low.
AI Predicts Uncertainty in Sparse Sensor Data
A new method adds noise to sensor inputs, enabling AI to generate reliable confidence intervals for reconstructing complex systems like climate and brain activity—without extra computational cost.
New Algorithm Speeds Up Complex Counting Problems
A breakthrough in parallel computing allows scientists to estimate complex mathematical quantities faster than ever before, with applications in physics and data analysis.
AI Learns to Focus on What Matters in Time Series Data
A new method helps AI models predict future trends by intelligently compressing historical data, reducing noise and computational costs while improving accuracy.
AI Transforms Library Search from Keywords to Knowledge
A new AI-powered system uses knowledge graphs to help researchers find relevant papers faster and reduce information overload by up to 90%, modernizing digital libraries for the age of big data.
AI Solves Physics Equations Without Retraining
A new method uses diffusion models to solve complex partial differential equations by incorporating physical laws during inference, achieving high accuracy and generalization across unseen parameters in seconds.
Networks That Grow Like Living Organisms
A new AI-driven architecture allows networks to evolve their own communication protocols in real-time, treating failures as fuel for adaptation rather than causes of collapse.
AI Models How Cooperation Persists Without Contracts
A new computational framework reveals how reciprocity sustains cooperation in platform ecosystems and multi-agent systems—explaining everything from App Store dynamics to AI coordination.
AI Learns to Match Optimizers to Complex Problems
A new framework uses reinforcement learning to dynamically select the best algorithm for each part of a large-scale optimization problem, dramatically improving performance and efficiency in real-world applications like satellite design.
AI Learns to See Images Like Humans Do
A new method fine-tunes image similarity metrics with human feedback, making AI tools more reliable for creative and practical tasks where personal judgment matters.
AI Learns to Think Like a Logical Machine
A new neural network design can simulate complex logical reasoning with exponential efficiency, bridging deep learning and formal automata theory for more interpretable AI systems.
A Simpler Way to Measure Confidence in Safety Arguments
Researchers develop a new method to quantify confidence in safety assurance cases, helping engineers balance risk and cost without false precision.
Quantum Clock Reveals When Particles Arrive
A new approach to an old physics puzzle shows how time emerges from quantum correlations, challenging traditional interpretations of quantum mechanics.
Quantum Beats Classical in New Communication Task
Researchers discover a surprising quantum advantage in identifying messages through noisy channels, with implications for secure data transmission and quantum contextuality.
New Math Tool Solves Decades-Old AI Problem
A novel matrix inverse preserves units across transformations, enabling more reliable robotics, control systems, and machine learning without arbitrary assumptions.
AI's Hidden Environmental Cost in Materials Science
A new perspective reveals how machine learning in chemistry and materials research is driving up energy consumption—and proposes strategies to make AI-driven discovery more sustainable without sacrificing accuracy.
AI Safety Gates Fail, But a Simple Check Succeeds
A new study shows that AI systems cannot reliably self-improve using traditional safety filters, but a verification method achieves perfect safety across scales, including large language models.
AI's Hidden Climate Cost Revealed in New Study
Climate-focused AI chatbots can consume significantly more energy than generic models, with some designs using over 10 times the power without clear quality gains, challenging their sustainability claims.
AI Still Fails at Women's Health Advice
A new benchmark reveals that even the best AI models struggle with women's health questions, often providing unsafe or incomplete guidance, with equity gaps as a universal blind spot.