Associate Professor Jonathan Ragan-Kelley, an MIT-IBM Watson AI Lab and CSAIL researcher, optimizes how computer graphics and images are processed for the hardware of today and tomorrow. Ragan-Kelley specializes in high-performance, domain-specific programming languages and machine learning.
A computational imaging algorithm reveals that ambient light sensors, which are passive components embedded in the screens of smart devices to alter monitor brightness, pose an imaging privacy threat by exposing users’ touch interactions to hackers, MIT researchers find.
SoftZoo is an open-source platform developed at MIT CSAIL that simulates wildlife for soft robotics co-design more systematically and computationally, thus better advancing the development of relevant algorithms.
A machine-learning technique called SALIENT addresses key bottlenecks in computation with graph neural networks by optimizing usage of the hardware, particularly GPUs. This upgrade significantly reduces training and inference time on extensive datasets to keep pace with fast-moving and growing data in finance, social networks, and fraud detection in cryptocurrency.
A new algorithm makes headway in solving the problem of exploration vs. exploitation in reinforcement learning. Developed by MIT CSAIL researchers, it attempts to make decision-making more efficient.