Convolutional Neural Networks, AI, and the Physical Limits of Computing Online
It is commonly held that recent advances in “artificial intelligence” have largely been possible due to recent increases in computing power. Moreover, it has long been an engineering assumption, emblematized by Gordon E. Moore's “law,” that this power will continue to improve over time. This talk interrogates the first assumption by way of the increasing untenability of the second: rather than a resurgence enabled by improvements in computation, the recent prominence of AI, it argues, is due to machine learning's potential for exploiting a material computational basis that is in the process of losing its dynamism.
|Presenter:||Andrew Lison, Assistant Professor, Department of Media Study|
This workshop is one in a series presented by the UB Digital Scholarship Studio and Network.
Click here for more events offered by DSSN.
- Thursday, November 19, 2020
- 3:30pm - 5:00pm
- Time Zone:
- Eastern Time - US & Canada (change)
- This is an online event. Event URL will be sent via registration email.