Wednesday, August 19, 2026 · Vol. I
Top AI stories from across the web, summarized and cited. Updated automatically throughout the day.
FedPref is a novel federated learning framework designed to extract structured information from free-text radiology reports, enabling collaborative AI model training among healthcare institutions without sharing sensitive patient data. Radiology reports, while rich in clinical detail, present a significant challenge…
An AI API contract specifies the full terms of interaction with a large language model, encompassing not only the chosen model but also critical details like requested reasoning…
via arXiv (cs.AI)
Aegis is a novel runtime governance system designed to manage and secure the actions of agentic AI systems. These increasingly capable AI models, which can interact with the real…
via arXiv (cs.AI)
Clinical trial programming refers to the critical process of transforming raw study data into analysis-ready datasets, adhering to strict CDISC standards for regulatory…
via arXiv (cs.AI)
A new study investigates how large language models (LLMs) perform in medical diagnosis, specifically evaluating both their accuracy and their *confidence* in those diagnoses. As…
via arXiv (cs.AI)
AI efficiency assessment refers to the measurement of computational resources required to train and run artificial intelligence models, a critical concern given the industry's…
via arXiv (cs.AI)
Flock Safety is a police-tech giant known for its vast network of over 120,000 automatic license plate readers (ALPRs) across the U.S. While presented as a crime-solving tool,…
via MIT Technology Review
AI companion toys for neurodivergent children are robotic or digital devices designed to provide connection, social skills practice, and therapeutic support through interactive…
via MIT Technology Review
Mixture-of-Experts (MoE) architectures refer to a class of large language models (LLMs) that enhance efficiency by activating only a subset of their parameters for any given…
via arXiv (cs.AI)
The evaluation of large language model (LLM) agents often relies on other LLMs acting as judges, particularly when obtaining real-world performance metrics is costly, slow, or…
via arXiv (cs.AI)
Language models, increasingly integrated into scientific workflows as "co-scientists," are powerful AI systems designed to understand, generate, and process human language. A…
via arXiv (cs.AI)
Autonomous reasoning refers to an AI system's ability to independently draw conclusions, make inferences, and solve problems based on its understanding of information. This…
via arXiv (cs.AI)
Top-k Mixture-of-Experts (MoE) routing refers to a crucial component in large language models that dynamically selects a small subset of specialized neural networks (experts) to…
via arXiv (cs.AI)