- Smooth Min
- Bias function
- Deep Ensembles: A Loss Landscape Perspective
- Javascript SVG renderer
- A Differentiable Programming System to Bridge Machine Learning and Scientific Computing
- Neural Turing Machines
- Differentiable Programming Manifesto
- Differentiable parallel approximate sorting networks
- Differentiable Ranks and Sortingusing Optimal Transport
- Fast Differentiable Sorting and Ranking
Encoding: Graph neural networks. Edges are gates and nodes are variables. Optimization: SGD-like
Encoding: Optimization: SGD-like
- End-to-End Differentiable Proving
- Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets
- Learning to Transduce with Unbounded Memory
- Learning to learn by gradient descent by gradient descent