Spline Frequency Estimation with Neural Networks
This research explores a novel neural network approach for accurately and efficiently determining the optimal frequency parameter in hyperbolic polynomial splines.
Science
When AI Goes Bad: Who Pays the Price?
As artificial intelligence becomes increasingly capable, the question of legal responsibility for AI-driven crime demands urgent attention.
Seeing Isn’t Always Believing: AI’s Blind Spots in Self-Driving Cars
A new approach systematically evaluates the potential safety and security risks arising from the inherent limitations of deep learning-based perception systems in autonomous vehicles.
Seeing the Big Picture: AI and Climate Change Communication
New research assesses how effectively artificial intelligence can analyze visual content on social media to understand public discourse around climate change.
Shattered Networks: Reconstructing Graphs from Fragmented Data
New research demonstrates how easily graph structures can be revealed even when using privacy-preserving spectral embeddings, and introduces tools to both benchmark this leakage and rebuild fragmented networks.
Echoes in the Data: Uncovering Hidden Market Signals
New research reveals that analyzing past text data with frozen language models can reveal economically relevant information missed by current market valuations.
Beyond Incident Counts: Mapping the Rise of AI Harms
A new framework classifies the evolving patterns of AI-related incidents to move beyond simple tracking and toward proactive risk mitigation.
Mapping Flood Risk with AI: A Faster, Data-Driven Approach
Researchers have developed an artificial intelligence framework that significantly speeds up flood hazard mapping by learning from complex hydraulic simulations.
When Algorithms Don’t Know What They Don’t Know: Fairness in Uncertain Systems
![Efforts to minimize false negatives, while initially effective, demonstrate a tendency toward instability and overshoot across decision rounds-a phenomenon exacerbated by interaction proxy bias, which causes diverging trajectories and underscores the inherent limitations of addressing uncertainty when foundational proxies are structurally compromised, as reflected in the observed [latex]\Delta\text{FNR}[/latex] fluctuations.](https://arxiv.org/html/2604.21711v1/x12.png)
New research explores how acknowledging and quantifying uncertainty in sequential decision-making-particularly when data is biased-can lead to more equitable and effective AI systems.
Opening the AI Black Box: A Path to Verifiable Safety
New research proposes a rigorous statistical framework for certifying the safety of artificial intelligence systems, moving beyond abstract risk assessments.