![The system dissects hurricane storm surge forecasting by constructing a spatio-temporal graph-nodes representing gauge stations and edges quantifying their correlations-to predict localized water level offsets [latex]\hat{o}_{i}(t)[/latex] and refine physics-based ADCIRC models, effectively learning to correct inherent biases in surge prediction.](https://arxiv.org/html/2604.20688v1/x2.png)
A new graph neural network model, StormNet, leverages connections between coastal monitoring stations to significantly reduce biases and improve the accuracy of storm surge predictions.
![The heat-exchanger model leverages prior probability densities-established for the changepoint time τ, fouling strength [latex]\beta_f[/latex], leak rate [latex]\beta_l[/latex], and fouling-event arrival rate λ-to constrain parameter estimation within the scenarios detailed in Table 1, acknowledging the inherent uncertainty in predicting system degradation.](https://arxiv.org/html/2604.20735v1/x3.png)
New research demonstrates a computationally efficient method for monitoring heat exchanger health, paving the way for real-time diagnostics and scalable predictive maintenance programs.
New research shows that artificial intelligence can assess a user’s risk of depression by analyzing their language on social media platforms.
![The study demonstrates that cosine similarity of latent feature vectors-calculated both with a limited set of the most activated channels and the full channel set-effectively captures forecast relationships, exhibiting distinct patterns when analyzed across regions-specifically, one mirroring the analysis region of Figure 1 and another centered at [latex]50^{\circ}N, 48^{\circ}W[/latex] with a [latex]5.81^{\circ}[/latex] radius-thereby highlighting the spatial dependence of forecast correlations.](https://arxiv.org/html/2604.20467v1/x3.png)
Researchers have developed a new visualization tool to explore the inner workings of artificial intelligence systems used for weather prediction.
New research suggests the rise of intelligent agents in finance isn’t simply about replacing workers, but fundamentally changing how financial work is organized and performed.
New research explores the limits of artificial intelligence in harnessing collective knowledge, even in structured environments like prediction markets.
New research reveals that artificial intelligence consistently outperforms people in identifying fraudulent claims and resists being swayed by biased investor pressure.
New research uses computational modeling to reveal how flawed governance structures can drive economic systems toward predictable breakdowns, while decentralized designs foster lasting growth.
A novel framework adapts public health surveillance methods to move beyond simple incident counts and understand the underlying dynamics of AI-related risks.
![The simulation, parameterized with [latex]\theta = 0.3[/latex], demonstrates that a shock to a specific Brazilian asset ([latex]VIVT3.SA[/latex]) triggers a localized cascade of defaults within the Brazilian financial subnetwork, while developed-market assets remain insulated from contagion, underscoring the inherent vulnerabilities of emerging economies.](https://arxiv.org/html/2604.19796v1/figure_7.png)
New research reveals that the way global equity markets are connected, combined with the potential for extreme losses, is a primary driver of systemic risk, rather than a broad, uniform threat.