AI Takes the Plunge: Monitoring Inland Water Quality with an Intelligent Agent
A new agentic AI system, NAIAD, is poised to transform how we understand and manage the health of our lakes, rivers, and streams.
Science
Predicting the Future of Networks: A Spatiotemporal Graph Approach
![DynaSTy accepts an initial state as input and iteratively refines a trajectory through a process of dynamic system modeling and trajectory optimization, ultimately producing an optimized trajectory as output that satisfies predefined constraints and objectives, effectively bridging the gap between initial conditions and desired system behavior as described by [latex] \dot{x} = f(x, u) [/latex].](https://arxiv.org/html/2601.05391v1/problem_st.png)
Researchers have developed a new framework, DynaSTy, for accurately forecasting node attributes in dynamic graphs by leveraging both spatial relationships and temporal evolution.
Robots Can’t See What They Say: The Hidden Dangers of AI Decision-Making
Despite impressive AI benchmarks, new research reveals that robots powered by large language models struggle with basic spatial reasoning, creating potentially hazardous scenarios in real-world applications.
Untangling Turbulence: How AI Predicts Particle Breakup
![The structure exhibits a scaling relationship where the aggregate’s gyration radius, [latex]R\sim eq a\,N\_{M}^{1/D\_{F}}[/latex], is determined by the radius of the constituent primary particles, [latex]a[/latex], the number of those particles, [latex]N\_{M}=384[/latex], and the fractal dimension of the cluster, [latex]D\_{F}=1.9[/latex], demonstrating how complex aggregates organize around these fundamental parameters.](https://arxiv.org/html/2601.05667v1/aggregate_DF1.9.png)
Researchers are leveraging the power of graph neural networks to accurately model the disintegration of small particle clusters within chaotic fluid flows.
Beyond Linear Predictions: A New Simulator for Dynamic Networks
![The study demonstrates that spectral models, when assessed across multiple rollout steps via [latex]L2L\_2[/latex] and [latex]R2R^2[/latex] metrics, consistently outperform spatial models in predicting positional accuracy and Poisson’s ratio; this advantage is particularly evident in the norm-NLSF model, which achieves the highest [latex]R2R^2[/latex] value at rollout step 100, suggesting its superior capacity for accurate material property estimation during iterative simulations.](https://arxiv.org/html/2601.05860v1/figures/perf.png)
Researchers have developed a novel spectral graph neural network simulator capable of more accurately and reliably modeling the complex, nonlinear dynamics of disordered elastic networks.
Predicting Claims Costs with Neural Networks
New research demonstrates how machine learning can improve the accuracy of loss reserving by leveraging both initial estimates and actual payment data.
The Rise of Autonomous Defenders: AI Agents and the Future of Cybersecurity
As artificial intelligence evolves beyond passive tools, a new generation of autonomous AI agents is emerging to reshape the landscape of cybersecurity, offering both powerful defenses and novel attack surfaces.
Lost in Translation: Bias in Global Financial Misinformation Detection
A new benchmark reveals how easily large language models can be swayed by contextual framing when identifying false financial claims across different languages.
Reconstructing the Heart with AI: A Single Model for Diverse Scans
Researchers have developed a powerful new artificial intelligence model capable of accurately reconstructing cardiac MRI images across a wide range of imaging conditions.
Local Voices, Smarter AI: Bridging the Knowledge Gap
A new approach empowers communities to enrich large language models with local narratives, improving accuracy and addressing information inequities.