![Market volatility, measured by a six-factor norm, consistently spikes during periods of documented stress-clustering into a distinct “crisis” regime [latex]\text{(red)}[/latex]-while moderate fluctuations define an “elevated” regime [latex]\text{(yellow)}[/latex], suggesting that fear and response to major events predictably shape market behavior over the 1990-2024 period.](https://arxiv.org/html/2601.10732v1/x1.png)
New research reveals a shifting predictive power between key equity factors, demonstrating that the Value premium can foreshadow movements in the Size premium during periods of market crisis.
A new artificial intelligence model analyzes electrocardiograms to predict a surprisingly wide range of diseases, moving beyond traditional cardiac diagnostics.
Researchers have developed an adaptable, data-driven pipeline to improve the accuracy and efficiency of operational wildfire forecasting.
A new approach dissects financial time series into distinct temporal components, revealing underlying dynamics and improving risk assessment.
New research explores how artificial intelligence can be used to design and enforce rules that prevent unfair practices in complex economic simulations.
A new analysis quantifies the impact of artificial intelligence on the US economy, revealing a complex picture of investment, production, and value creation.
A new deep learning framework leverages multi-modal brain imaging and patient data to enhance the accuracy of Parkinson’s and Alzheimer’s disease diagnosis.
![The study demonstrates how the dynamics of neuronal firing rates-specifically, the mean firing rate and membrane potential-shift across regimes determined by the correlation of noise, with theoretical predictions aligning with simulations of quadratic integrate-and-fire neurons, and revealing a dependence on both the correlation coefficient and noise intensity as described by the deterministic component of [latex]Eq. (9)[/latex].](https://arxiv.org/html/2601.10032v1/x2.png)
New research reveals that correlated noise in networks of spiking neurons can surprisingly stabilize activity and induce unique states of metastability.
A new approach leverages mixtures of transparent local models to create interpretable machine learning systems with guaranteed performance bounds.
A new benchmark reveals that current audio AI struggles with the complexities of everyday sound, often performing worse with common noise reduction techniques.