AI Moves From Lab to Front Line of Outbreak Response, WHO Forum Hears
A WHO forum report describes how AI tools for outbreak forecasting, genomic surveillance and decision support are advancing, while experts ...
Tag: Large Language Models
AI Models Trained on Old Rumour Data Falter on Irish Rumours, New Dataset Shows
A new study finds that rumour stance classification models trained on RumourEval data degrade considerably on a new Irish rumour ...
AI-Structured Cancer CT Reports Win Over Readers in Prospective Survey Study
A prospective questionnaire study found readers rated LLM-structured oncologic CT reports higher than free-text reports on 11 of 12 items, ...
Retrieval-augmented LoRA-tuned language model writes structured software test cases
A FLAN-T5 model combining retrieval augmentation and parameter-efficient fine-tuning achieved high-quality structured test-case generation with measurable latency costs.
AI Agents Debate to Uncover Users’ Core Values for Better Recommendations
A multi-agent framework called ZOOM uses summarization and supervised LLM debate to extract stable user values from interaction histories, improving ...
Survey maps how retrieval-augmented generation is reshaping biomedical AI
A new survey of 94 studies charts the rise of retrieval-augmented generation in biomedicine and identifies trade-offs among reasoning depth, ...
Generative AI Shapes the Full Lifecycle of Health Misinformation, Review Finds
A new narrative review argues that generative AI simultaneously produces, detects, corrects, and amplifies the consequences of health misinformation across ...
AI Guardrails for Factories: Architecture Keeps LLMs From Unsafe Plant Actions
A four-layer multi-agent architecture confines LLM reasoning to an observation-only role and validates every proposal against plant engineering diagrams before ...
AI Agent Team Automates Earthquake-Resistant Concrete Design
A new multi-agent AI system autonomously designs earthquake-resilient fiber-reinforced concrete, cutting expert design cycles by about 75 percent while keeping ...
New LAMPP Model Aims to Guide LLM Use in Programming Education
A perspective paper proposes the LAMPP model, a theory-based framework with four design principles for integrating large language models into ...
Generative AI’s Text-Only Design Excludes Deaf Signers, Conceptual Paper Argues
A conceptual paper argues that generative AI's text-based interaction paradigm structurally disadvantages deaf signers and proposes a visuo-spatial model for ...
Foundation models promise a diabetes AI revolution—without the evidence to match
A critical review finds that foundation models are transforming diabetes AI across retinal imaging, glucose monitoring, and clinical text, yet ...
AI Model Impression-R1 Matches Radiologists in Multicenter Impression Writing Test
A domain-specialized reasoning model called Impression-R1 matched or outperformed general-purpose AI and participating radiologists on radiological impression generation across 11,884 ...
New tool audits how AI chatbots answer teen mental health questions
Researchers developed AMHART, a clinician-informed tool that found chatbot responses covered more guideline-informed criteria for anxiety than depression and for ...
AI drafts and grades image-based neurosurgery exam questions with expert oversight
A single-center study found multimodal AI models could draft image-dependent neurosurgery exam questions and grade resident responses with about 90% ...
AI Takes Aim at Beam Losses and Downtime at Europe’s Spallation Source
The European Spallation Source is developing a strategy and pilot projects to apply artificial intelligence and machine learning across its ...
Bibliometric Map Traces Computational Social Science From Big Data to Generative AI
A bibliometric analysis of 861 Scopus-indexed articles traces how computational social science evolved from agent-based modelling to generative artificial intelligence ...
Unified Building Data Platform Tested Against AI Assistants in Real University Deployment
A federated GraphQL platform unified five building data systems at a Danish university, letting an AI model answer real facility ...
Machine-translated radiology reports fluent yet error-prone, expert review essential
A retrospective study of 310 German radiology reports translated with DeepL found most translations rated good or excellent, yet clinically ...
Scientists propose strict guardrails for using generative AI in assessing pain in older adults with cognitive impairment
A new Perspective article proposes a clinician-governed framework limiting generative AI to organizing and reviewing pain-assessment information for older adults ...
