Rethinking Computation: Substrates for Machine Intelligence

Deep learning has had a major impact in the last 3 years. Imperfect interactions with machines, such as speech, natural language, or image processing have been made robust by deep learning and deep learning holds promise in finding usage structure in large datasets. However, the training process is lengthy and has proven to be difficult to scale due to constraints of existing compute architectures. Beyond the algorithms, deep learning is a fundamentally new way to express computation. In this talk, I will outline some of these challenges and how fundamental changes to the organization of computation and communication can lead to large advances in capabilities.