The Shifting Roots of Speech AI Errors
New research reveals how the causes of speech recognition ‘hallucinations’ fundamentally change as model size increases.
Science
When Agents Fail: Uncovering the Roots of Error in AI Systems
A new empirical study dives deep into the common failure modes of modern agentic frameworks, pinpointing the architectural weaknesses and bug patterns that lead to unpredictable behavior.
Banking on Context: A New Foundation Model for Financial Understanding
![A user’s record is represented as an ordered event history and profile state, where each field is decomposed into a semantic type, associated values, and a temporal coordinate; keys and values are embedded from a shared lookup table, and value tokens receive positional embeddings within each field, allowing a Profile State Encoder to map the profile state-with time since life-long events encoded via RoPE-into a [USR] embedding, while an Event Encoder independently maps event tokens into a [EVT] embedding augmented with calendar features, and a History Encoder contextualizes the resulting sequence with time to the last event-also encoded via RoPE-to produce a comprehensive representation of the user record.](https://arxiv.org/html/2604.08649v1/x4.png)
Revolut’s PRAGMA model introduces a novel approach to understanding user financial histories by leveraging masked modeling and heterogeneous data sources.
From Digital Waste to Planetary Insight
A new approach, dubbed ‘Scrapyard AI’, repurposes discarded artificial intelligence models to monitor the environmental consequences of resource extraction.
Powering the AI Boom: Smart Grids for Data Center Demand
As artificial intelligence infrastructure expands, ensuring reliable and efficient transmission network capacity is crucial for avoiding outages and escalating costs.
Sequences and Connections: Rethinking Event Modeling
New research demonstrates that incorporating relationships between users and events significantly boosts the performance of sequence prediction models.
Seeing Trouble Ahead: A New Approach to Accident Anticipation
Researchers have developed a computationally efficient system that predicts potential accidents by analyzing broad video features, moving beyond traditional object detection methods.
Sensing the Unseen: Optimizing Tsunami Warning Networks
A new framework leverages high-performance computing and Bayesian methods to strategically position offshore sensors for faster, more accurate tsunami detection.
Calming the Chaos: AI-Powered Disaster Response
![Despite inherent model prediction errors [latex]RMSE=0.19[/latex] for Hurricane Harvey and [latex]0.29[/latex] for Irma, a learned actor-critic policy demonstrably reduced mean fear by approximately 70% in Harvey and 50% in Irma-despite a higher initial fear level in the latter-while simultaneously maintaining or improving power availability and physical health, suggesting effective intervention even when extrapolating to novel but related disaster scenarios.](https://arxiv.org/html/2604.08802v1/artifacts_irma/plots/states.png)
A new framework leverages artificial intelligence to coordinate critical resources during disasters, minimizing public fear and maximizing the effectiveness of emergency services.
Beyond the Crash: Building Resilient Investment Strategies
![The system distills market momentum by first identifying sector leaders based on cumulative growth [latex]R_{i}[/latex], then refining this selection via a Volatility-Adjusted Momentum (VAM) score to establish an ‘Anchor Triad’ - the top three most robustly trending sectors - thereby constructing a portfolio grounded in prevailing structural forces.](https://arxiv.org/html/2604.09060v1/signal_gen_module.png)
A new framework aims to mitigate downside risk and consistently generate alpha by intelligently combining momentum, risk parity, and robust optimization techniques.