Bridging the Gap: Aligning Language Models Without Sharing Data
A new framework allows independent large language models to collaborate on inference tasks while preserving the privacy of their underlying data and weights.
Science
Decoding Time Series: A New Approach to Forecasting
Researchers have developed a novel method for disentangling complex time series data, leading to improved forecasting accuracy and interpretability.
Beyond Package Lists: Uncovering Hidden Vulnerabilities in Python
A new approach to software security focuses on tracing dependencies beyond readily available package metadata to identify risks lurking in native libraries.
Decoding Engagement: How Speaker Expression Fuels Video Appeal
New research reveals how analyzing a speaker’s facial expressions, voice, and language can accurately predict audience engagement and perceived vocal attractiveness in video learning materials.
The Price of Pruning: When Slimming Neural Networks Kills Understanding
Aggressively reducing the size of neural networks can maintain performance, but new research reveals a surprising cost: a drastic loss of interpretability.
Seeing the Universe Clearly: A New Approach to Weak Gravitational Lensing
![A machine learning model estimates galaxy shapes, but its raw output requires calibration; this is achieved by analytically computing the shear response of a smoothed image and contrasting it with the model’s gradient-yielding a calibration matrix [latex]R\_{ij}=\partial e\_{i}/\partial\gamma\_{j}[/latex]-allowing for linear correction of the estimator and subsequent evaluation of residual biases quantified as multiplicative ([latex]m[/latex]) and additive ([latex]c[/latex]) parameters.](https://arxiv.org/html/2603.19046v1/x1.png)
Researchers have developed a novel machine learning framework that dramatically improves the accuracy and reliability of measuring the distortion of light caused by gravity, opening new avenues for cosmological studies.
Fortifying Critical Infrastructure with Intelligent Digital Twins
A new framework combines physics-informed machine learning with anomaly detection to create cyber-resilient digital twins for safeguarding industrial control systems.
Uncovering Hidden Rhythms in Chaotic Data
A new network-based approach reveals how events cluster together in irregular time series, offering insights into complex systems from heartbeats to turbulent flows.
Who Governs the Governors? Ethics in AI Collectives
New research reveals that the structure of AI governance, rather than the intelligence of the AI itself, is the key to preventing corruption in multi-agent systems.
Predicting Material Stress with AI: A New Hybrid Approach
Researchers have developed a novel artificial intelligence framework that accurately forecasts stress within hyperelastic materials, paving the way for more efficient simulations and designs.