- Boundary Seeking GANs
- Learning General Purpose Distributed Sentence Representations via Large Scale Multi-task Learning
- Deep Complex Networks
- Fraternal Dropout
- Residual Connections Encourage Iterative Inference
- Twin Networks: Matching the Future for Sequence Generation
- Graph Attention Networks
- Extending the Framework of Equilibrium Propagation to General Dynamics
- FigureQA: An Annotated Figure Dataset for Visual Reasoning
- SEARNN: Training RNNs with global-local losses
- Parametric Adversarial Divergences are Good Task Losses for Generative Modeling
- Learning General Purpose Distributed Sentence Representations via Large Scale Multi-task Learning
- Deep Complex Networks
- Twin Networks: Matching the Future for Sequence Generation