A model’s ability to generalize is influenced by both the diversity of the data and the way the model is trained, researchers report.

Lincoln Laboratory leads a large-scale measurement campaign in New York City to improve air dispersion models and emergency protocols.

Deep-learning methods confidently recognize images that are nonsense, a potential problem for medical and autonomous-driving decisions.

MIT-IBM Watson AI Lab researchers aim to design concrete mixtures that use AI to shrink environmental footprint and cost, while recycling byproducts and increasing performance.

A new study finds that investments in R&D on materials and chemistry were key, while economies of scale contributed less.

New work on linear-probing hash tables from MIT CSAIL could lead to more efficient data storage and retrieval in computers.

Researchers decipher when and why immune cells fail to respond to immunotherapy, and suggest that T cells need a different kind of prodding in order to re-engage the immune response.

A visual analytics tool helps child welfare specialists understand machine learning predictions that can assist them in screening cases.

A deep model was trained on historical crash data, road maps, satellite imagery, and GPS to enable high-resolution crash maps that could lead to safer roads.

Researchers find blind and sighted readers have sharply different takes on what content is most useful to include in a chart caption.