Can we make AI sustainable?

The last decade has seen exciting advances in the capabilities of AI systems, but these advances have come with the expense of ever increasing power consumption and cost as models grow larger and larger. In this session, MIT-IBM Watson AI Lab director David Cox will discuss work currently underway at IBM Research on reducing the footprint of AI models, reducing costs and allowing AI to be usefully deployed “at the edge”. He will also cover efforts to use AI to design more efficient AI models and to dynamically process less data, as well as research on entirely new computing architectures that enable low-power AI.