AI Takes on Science: A New Era of Workflow?
New research demonstrates that artificial intelligence can meaningfully accelerate scientific discovery when combined with robust validation procedures.
Science
The AI Tetrahedron: A New Framework for Materials Discovery
A novel research paradigm integrating artificial intelligence and materials science is emerging, promising to accelerate the design and discovery of advanced materials.
Smarter Sensing: Prioritizing Perception for Seamless Human-Robot Interaction
A new framework intelligently schedules perception modules, focusing computational resources on the most relevant information to enhance efficiency and reduce latency in collaborative robotics.
Beyond Benchmarks: Building a Lab to Truly Measure AI Minds
![The system architecture establishes a framework for [latex] n [/latex] interconnected modules, enabling scalable and modular implementation of complex functionalities.](https://arxiv.org/html/2603.13126v1/figure1.png)
Researchers have created a cloud-based platform, PsyCogMetrics™ AI Lab, designed to move beyond simple performance scores and rigorously evaluate the cognitive capabilities of large language models.
Navigating the Unknown: Robots Learn to Adapt and Explore
A new approach fine-tunes pre-trained navigation policies using reinforcement learning, allowing robots to safely and reliably explore environments they’ve never seen before.
Predictive Models Unlock Physical System Insights
New research demonstrates that learning to predict future states within a latent space yields superior representations of physical systems compared to traditional reconstruction or autoregressive methods.
Robots Learn to Dismantle Themselves for Better Movement
New research shows robots can strategically sacrifice their own parts to adapt to challenging terrains and improve locomotion performance.
Sensing with Understanding: AI for Wearable Health
![Time series decomposition enables the creation of an extended model [latex]M^\hat{M}[/latex] by transforming an original time series [latex]\mathbf{x}[/latex] into component representations [latex]{\mathbf{C}\_{\mathbf{x}}}[/latex] via a forward function FF, and then leveraging the inverse transformation [latex]F^{-1}(\cdot)[/latex] in combination with the original time series network MM, all without requiring model retraining.](https://arxiv.org/html/2603.12880v1/x1.png)
New research explores how to build artificial intelligence systems that not only predict health metrics from wearable sensors, but also clearly explain why they made those predictions.
Bending the Rules: Robotic Control of Flexible Objects in Tight Spaces
Researchers have developed a new approach to controlling deformable objects – like cables or hoses – alongside rigid structures, enabling more complex manipulation in cluttered environments.
Reasoning Agents Aren’t Always Logical
New research reveals that even the most advanced AI reasoning systems can be surprisingly fragile when faced with subtly altered inputs.