Energy-Efficient AI | Vivienne Sze | TEDxMIT

Today, most of the processing for Artificial Intelligence (AI) happens in the cloud (i.e., data centers); however, there are many compelling reasons to perform the processing locally on the device (e.g., smartphones or robots) including reducing the dependence on communication infrastructure, preserving data privacy, and reducing reaction time.

One of the key limitations of local processing is energy consumption. Researchers are working on various techniques to enable energy-efficient AI, and how energy-efficient AI extends the reach of AI beyond the cloud to enable a wide range of applications from robotics to health care.
Vivienne Sze received the B.A.Sc. (Hons) degree in electrical engineering from the University of Toronto, Toronto, ON, Canada, in 2004, and the S.M. and Ph.D. degree in electrical engineering from the Massachusetts Institute of Technology (MIT), Cambridge, MA, in 2006 and 2010 respectively. She received the Jin-Au Kong Outstanding Doctoral Thesis Prize for her Ph.D. thesis in electrical engineering at MIT in 2011.

She is an Associate Professor in the Electrical Engineering and Computer Science Department at MIT. Her research interests include energy efficient algorithms and architectures for portable multimedia applications. From September 2010 to July 2013, she was a Member of Technical Staff in the Systems and Applications R&D Center at Texas Instruments (TI), Dallas, TX, where she designed low-power algorithms and architectures for video coding. She also represented TI in the JCT-VC committee of ITU-T and ISO/IEC standards body during the development of High Efficiency Video Coding (HEVC), which received a Primetime Emmy Engineering Award. She co-edited a book entitled High Efficiency Video Coding (HEVC) - Algorithms and Architecture (Springer, 2014).

[linglingfan]

It seems that there is no end for energy efficient hardware. You can always go more efficiently.