Chasing Ghosts in Financial Machine Learning
![The system demonstrates an ability to generate returns exceeding a simple buy-and-hold strategy when tested against historical market data, yet this performance diminishes to indistinguishable levels when evaluated on synthetic data simulating market volatility modeled by a [latex]GARCH(1,1)[/latex] process, suggesting the model’s efficacy is heavily reliant on specific, non-stationary characteristics of the observed data.](https://arxiv.org/html/2604.15531v1/x7.png)
New research reveals how easily machine learning models can appear profitable due to hidden biases and flawed evaluation, not genuine predictive power.
Science
Beyond the Forecast: AI and Machine Learning Sharpen Subseasonal Predictions
A new probabilistic bias correction framework dramatically improves the accuracy of short-term weather forecasts generated by both traditional and artificial intelligence models.
Beyond the Code: Power, Politics, and the Future of AI
A new analysis argues that focusing solely on AI’s technical risks obscures the crucial economic and political forces driving its development and deployment.
Forging Crypto Futures: Generating Realistic Price Data with AI
Researchers are leveraging generative models to create synthetic cryptocurrency price data that mirrors the complexities of real-world market behavior.
Sensing Safety: An Intelligent System for Proactive Elder Care
A new edge-cloud architecture leverages real-time sensor data and risk assessment to provide rapid emergency response and enhanced independence for seniors.
Decoding AI Decisions: A New Path to Trustworthy Explanations
Researchers are developing a framework to rigorously verify why AI systems make specific predictions and identify the minimal features driving those choices.
Smarter Seismic Insights: Bridging Expert Knowledge and Machine Learning
![The distributions of extracted score features, conditioned on class labels within simulated environments, demonstrate how a feature summary - constructed from [latex]\ell\ell[/latex]-based scores - characterizes variations in these settings.](https://arxiv.org/html/2604.14809v1/x4.png)
A new framework leverages domain expertise to improve the accuracy and interpretability of seismic event classification, even with incomplete data.
When Logic Breaks Down: Understanding AI Reasoning Errors
New research sheds light on the specific moments large language models stumble during complex reasoning tasks, revealing patterns of failure before they become critical.
false Alarms in Early Warning Systems: A Critical Look at Trend Tests
![The study demonstrates that a modified Mann-Kendall test, applied to time series of lag-1 autocorrelation derived from simulations of a fold normal form ([latex] r = -1 [/latex]) with multiplicative noise, reliably detects trends at the nominal 5% significance level across time series of length [latex] N = 100 [/latex] using rolling windows of relative size α.](https://arxiv.org/html/2604.15230v1/images_annexes/FigB1_Mult_sigma.png)
New research reveals that commonly used statistical methods for predicting abrupt shifts in complex systems are often unreliable due to hidden biases.
Sound Off: Hijacking AI with Hidden Audio Commands
New research reveals that Large Audio-Language Models can be subtly manipulated by imperceptible audio prompts, raising significant security concerns.