
Why Deep Networks and Brains Learn Similar Features with Sophia Sanborn - #644
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Convolution on Curved Surfaces and Robust Features in CNNs
Doing a convolution on a graph or curved surface, like a sphere, is equivalent to 3D rotation. Generalizing the concept of a CNN to different groups and geometric domains has been successful, but it assumes knowing the structure of the data. However, many transformations are difficult to describe, making data augmentation necessary but imperfect.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.