Adam Zewe

MIT researchers demonstrate that simpler physics-based models can generate more accurate climate predictions than state-of-the-art deep-learning models in certain scenarios, and that common benchmarking techniques can be distorted by natural data variations.

“Author of "Simpler models can outperform deep learning at climate prediction" in mit.edu”

“Byline author reporting on MIT researchers' HardFlow algorithm that forces AI to obey hard constraints while finding optimal solutions.”