- Improving the Gating Mechanism of RNNs
- What Can Learned Intrinsic Rewards Capture?
- A Free-Energy Principle for Representation Learning
- LEEP: A New Measure to Evaluate Transferability of Learned Representations
- TASKNORM: Rethinking Batch Normalization for Meta-Learning
- Feature Quantization Improves GAN Training
- Deep Divergence Learning
- Do RNN and LSTM have Long Memory?
- Generative Pretraining from Pixels
- Reverse-Engineering Deep ReLU Networks