Governing AI’s Expansion: A New Framework for Open Institutions
As artificial intelligence systems grow in autonomy and scale, a formal model is needed to manage their boundaries and ensure responsible expansion.
Science
Seeing the World Like a Robot: 3D Object Perception from a Single Image
New research demonstrates how vision-language models can enable robots to accurately estimate the 3D position of objects using only standard RGB images.
Beyond Search: A Multimodal Agent for Deep Research
Researchers introduce a new agentic system that combines visual and textual data to dramatically improve the quality and efficiency of online investigation.
Decoding Swarm Behavior: How Groups Emerge From the Chaos
A new framework accurately infers hidden group structures within complex, overlapping multi-agent swarms, unlocking better understanding and prediction of collective motion.
Untangling Influence in the Age of AI Agents
![The proposed framework estimates human treatment effects by constructing subpopulations stratified by expected human composition and treatment exposure, then fitting an experimental state evolution model to aggregate outcome trajectories to project counterfactuals under full treatment and control-yielding a difference that quantifies the human total treatment effect [latex]q^{S}=1[/latex].](https://arxiv.org/html/2603.01339v1/2603.01339v1/figures/uai_fig1.png)
New research offers a framework for understanding how interventions affect users on online platforms increasingly populated by artificial intelligence.
Seeing, Speaking, and Stepping: Autonomous Control for Soft Robots
Researchers have developed a new end-to-end framework that allows magnetically controlled soft robots to navigate complex environments using vision and language instructions.
Simulating Science: When AI Takes the Reins
![The study elucidates a formal description of a simulated fluid displacement problem, encompassing foundational assumptions, governing equations-including constitutive laws and the introduction of fractional calculus-and a nuanced analysis of mobility ratio effects on favorable versus unfavorable flow regimes, ultimately emphasizing the significance of the quarter-five-spot configuration as a benchmark for understanding multiphase flow dynamics in porous media, expressed mathematically as [latex] \frac{d P}{d x} = -\frac{\mu}{\kappa} v [/latex].](https://arxiv.org/html/2603.00214v1/2603.00214v1/x3.png)
A new approach combines artificial intelligence with physics-based simulation to build and validate scientific models through active experimentation.
Predicting People, Navigating Safely: A Real-Time Approach
New research details a system for rapidly forecasting human movement, allowing robots to navigate crowded spaces with increased agility and safety.
Bridging the Gap: AI That Plans and Executes Scientific Experiments Safely
![BioProAgent operates on the premise that robust action emerges not from centralized control, but from a layered ecosystem of cognition and rectification, where contextual understanding-grounded in symbolic representation Φ-informs a neural planner [latex]\pi\_{\theta}[/latex] operating within a Design-Verify-Rectify finite state machine [latex]\Delta(\sigma)[/latex], all secured by hierarchical verification protocols [latex]\mathcal{K}\_{s},\mathcal{K}\_{p}[/latex] that deterministically enforce physical safety.](https://arxiv.org/html/2603.00876v1/2603.00876v1/x2.png)
Researchers have developed a new framework that combines the power of large language models with deterministic reasoning to enable trustworthy autonomous experimentation.
Beyond Black Boxes: Building Trust in Social Robots
As foundation models power increasingly sophisticated social robots, ensuring these systems can explain their actions in a way that is both ethical and tailored to individual users is paramount.