Conditional 3D generation is undergoing a significant advancement, enabl...
Knowledge distillation conducts an effective model compression method wh...
Training a 3D scene understanding model requires complicated human
annot...
Mainstream 3D representation learning approaches are built upon contrast...
The success of deep learning heavily relies on large-scale data with
com...
The success of deep learning is usually accompanied by the growth in neu...
Recent progress in image-to-image translation has witnessed the success ...
The remarkable breakthroughs in point cloud representation learning have...
Quantized neural networks typically require smaller memory footprints an...
Most feedforward convolutional neural networks spend roughly the same ef...