Plenary Session — Invited Papers
Chair: Jan van der Spiegel, University of Pennsylvania, Philadelphia, PA
ISSCC Conference Chair
Associate Chair: Eugenio Cantatore, Eindhoven University of Technology,
Eindhoven, The Netherlands
ISSCC International Technical Program Chair
FORMAL OPENING OF THE CONFERENCE 8:30 AM
1.1 Deep Learning Hardware: Past, Present, and Future 8:45 AM
Yann LeCun, Facebook AI Research & New York University, New York, NY
Deep learning has caused revolutions in computer understanding of images, audio, and text,
enabling new applications such as information search and filtering, autonomous driving,
radiology screening, real-time language translation, and virtual assistants. But almost all these
successes largely use supervised learning, which requires human-annotated data, or
reinforcement learning, which requires too many trials to be practical in most real-world
situations. In contrast, animals and humans seem to learn vast amounts of background
knowledge about the world through mere observation and occasional actions in a self-
supervised manner. Making progress in self-supervised learning is the main challenge of AI
for the next decade. Success may result in machines with some level of common sense. But
they will be built around deep learning architectures that are considerably larger than current
ones, requiring vastly more powerful hardware than what we have today.
1.2 Intelligence on Silicon: 9:20 AM
From Deep Neural Network Accelerators to Brain-Mimicking AI-SoCs
Hoi-Jun Yoo, KAIST, Daejeon, Korea
Deep learning is influencing not only the technology itself but also our everyday lives.
Formerly, most AI functionalities and applications were centralized on datacenters. However,
the primary platform for AI has recently shifted to mobile devices. With the increasing demand
on mobile AI, conventional hardware solutions face their ordeal because of their low energy
efficiency on such power hungry applications. For the past few years, dedicated DNN
accelerators inference have been under the spotlight. However, with the rising emphasis on
privacy and personalization, ability to learn on mobile platform is becoming the second hurdle
for “on-device AI.” Going hand in hand, the brain mimicking is also a highlighted field in AI.
Fundamentals of neuromorphic architecture start from the synapses and the neurons, which
are realized with custom memories including non-volatile memories. The brain mimicking AI
is not restricted to the naïve neuron implementation, but it extends further to mimic the
behavioral characteristics such as “connectome” or “visual attention.” In conclusion, mobile
learning and brain mimicking will be the two horses driving the carriage called AI, thus
opening up new requirements for the next generation deep learning SoCs.
ISSCC, SSCS, IEEE AWARD PRESENTATIONS 9:55 AM
BREAK 10:20 AM
SESSION 1 Monday February 18
th
, 8:30 AM
12