Why it’s critical to move beyond overly aggregated machine-learning metrics
New research detects hidden evidence of mistaken correlations — and provides a method to improve accuracy.
Institute for Medical Engineering and Science (IMES)
MIT scientists investigate memorization risk in the age of clinical AI
New research demonstrates how AI models can be tested to ensure they don’t cause harm by revealing anonymized patient health data.
Researchers discover a shortcoming that makes LLMs less reliable
Large language models can learn to mistakenly link certain sentence patterns with specific topics — and may then repeat these patterns instead of reasoning.
Charting the future of AI, from safer answers to faster thinking
MIT PhD students who interned with the MIT-IBM Watson AI Lab Summer Program are pushing AI tools to be more flexible, efficient, and grounded in truth.
Using generative AI, researchers design compounds that can kill drug-resistant bacteria
The team used two different AI approaches to design novel antibiotics, including one that showed promise against MRSA.
New AI system uncovers hidden cell subtypes, boosts precision medicine
CellLENS reveals hidden patterns in cell behavior within tissues, offering deeper insights into cell heterogeneity — vital for advancing cancer immunotherapy.
Changing the conversation in health care
The Language/AI Incubator, an MIT Human Insight Collaborative project, is investigating how AI can improve communications among patients and practitioners.
MIT and Mass General Brigham launch joint seed program to accelerate innovations in health
The MIT-MGB Seed Program, launched with support from Analog Devices Inc., will fund joint research projects that advance technology and clinical research.
LLMs factor in unrelated information when recommending medical treatments
Researchers find nonclinical information in patient messages — like typos, extra white space, and colorful language — reduces the accuracy of an AI model.
3 Questions: How to help students recognize potential bias in their AI datasets
Courses on developing AI models for health care need to focus more on identifying and addressing bias, says Leo Anthony Celi.