Deep Learning
Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules
Niklas Gebauer, Michael Gastegger, Kristof Schütt
Neural Information Processing Systems http://nips.cc/
One ticket to win them all: generalizing lottery ticket initializations across datasets and optimizers
Ari Morcos, Haonan Yu, Michela Paganini, Yuandong Tian
Unfortunately, finding these "winning ticket" initializations is computationally expensive.
FOF: Learning Fourier Occupancy Field for Monocular Real-time Human Reconstruction
The advent of deep learning has led to significant progress in monocular human reconstruction.
TinyTL: Reduce Memory, Not Parameters for Efficient On-Device Learning
Existing work solves this problem by reducing the number of trainable parameters. However, this doesn't directly translate to memory saving since the major bottleneck is the activations, not parameters.
Multi-objects Generation with Amortized Structural Regularization
Taufik Xu, Chongxuan LI, Jun Zhu, Bo Zhang
Neural Information Processing Systems http://nips.cc/
Contact-aware Human Motion Forecasting Wei Mao
During training, we explicitly encourage consistency between the global motion and the local poses via a prior defined using the contact maps and future poses. Our approach outperforms the state-of-the-art human motion forecasting and human synthesis methods on both synthetic and real datasets.
Recurrent Convolutional Neural Networks Learn Succinct Learning Algorithms
Neural networks (NNs) struggle to efficiently solve certain problems, such as learning parities, even when there are simple learning algorithms for those problems. Can NNs discover learning algorithms on their own?