VIS Lab
VIS Lab
People
Publications
Contact
Light
Dark
Automatic
Interpretability
LieGG: Studying Learned Lie Group Generators
Symmetries built into a neural network have appeared to be very beneficial for a wide range of tasks as it saves the data to learn them. We depart from the position that when symmetries are not built into a model a priori, it is advantageous for …
Cite
×