MIT-led team receives up to $38 million for disease-monitoring patch
An MIT-led team has received funding from the U.S. Advanced Research Projects Agency for Health for a four-year project worth up to $38 million to develop a wearable patch that continuously monitors biomarkers of heart disease and sepsis.
MEMENTO study improves book question answering with a 20K-token context
Researchers report in a preprint that MEMENTO raised Qwen3-8B's accuracy from 36.9% to 67.8% across four book question-answering benchmarks while holding no more than 20K tokens in context. The tests included books of up to 2.5M tokens.
CheXPoT study improves AI radiology reports by removing unsupported findings
Researchers report in a preprint that CheXPoT improved the precision of AI-generated chest radiograph reports by checking findings against image evidence and removing unsupported statements.
DreamX preprint reports 300× compression of video action representations
Researchers report in a DreamX preprint that they compressed each latent video frame from 1,200 tokens to 4 while retaining action information that transfers across human and robot embodiments and environments.
GUILD preprint reports more efficient collection of robot training data
Researchers report in a GUILD preprint that guided collection of robot demonstrations improved sample efficiency and deployment performance in simulation and real robot tests, compared with passive collection strategies, multiple baselines and experienced roboticists.
RoboTryOn study reports robot task success rising from 2% to 60%
Researchers report in an arXiv preprint that robot policies trained with RoboTryOn synthetic demonstrations improved average success on novel tasks from 2% to 60%, without real robot demonstrations of those tasks.