With global power demand from data centers expected to more than double by 2030, the MIT Energy Initiative (MITEI) in September launched an effort that brings together MIT researchers and industry experts to explore innovative solutions for powering the data-driven future. At its annual research conference, MITEI announced the Data Center Power Forum, a targeted research effort for MITEI member companies interested in addressing the challenges of data center power demand. The Data Center Power Forum builds on lessons from MITEI’s May 2025 symposium on the energy to power the expansion of artificial intelligence (AI) and focus panels related to data centers at […]

× Predicting Ego-centric Video from human Actions (PEVA). Given past video frames and an action specifying a desired change in 3D pose, PEVA predicts the next video frame. Our results show that, given the first frame and a sequence of actions, our model can generate videos of atomic actions (a), simulate counterfactuals (b), and support long video generation (c). Recent years have brought significant advances in world models that learn to simulate future outcomes for planning and control. From intuitive physics to multi-step video prediction, these models have grown increasingly powerful and expressive. But few are designed for truly embodied […]

What exactly does word2vec learn, and how? Answering this question amounts to understanding representation learning in a minimal yet interesting language modeling task. Despite the fact that word2vec is a well-known precursor to modern language models, for many years, researchers lacked a quantitative and predictive theory describing its learning process. In our new paper, we finally provide such a theory. We prove that there are realistic, practical regimes in which the learning problem reduces to unweighted least-squares matrix factorization. We solve the gradient flow dynamics in closed form; the final learned representations are simply given by PCA. Learning dynamics of […]