Reading list

Apart from the papers listed below, presenter may also discuss few other relevent papers. Please contact Viraj to add/remove the paper in this list.

Polynomial networks, Factorization machines
- On the Computational Efficiency of Training Neural Networks, Livni et al., NIPS 2014 (Details)
- Polynomial networks and factorization machines, Blondel et al., ICML 2016(Details)
- Multi-Output factorization machines, Blondel et al, preprint 2017
Invertibility of neural nets
- Learning RELUs via gradient descent, Soltanolkotabi, 2017. (Details)
- Towards understanding invertibility of CNNs, Gilbert et al, 2017.(Details)
- Identity matters in deep learning, Hardt and Ma, 2017.
Recurrent models and sparse coding
- Recurrent inference machines, Welling et al.
- Understanding Sparse Coding via Matrix Factorization, Bruna 2017
- Convolutional NNs and convolutional sparse coding, Elad et al.
- Learning to learn by gradient descent by gradient descent, Andrychowicz et al. 2016
Statistical aspects of neural net learning
- Why and When Can Deep – but Not Shallow – Networks Avoid the Curse of Dimensionality: a Review, Poggio et al., 2016.
- L1-regularized Neural Networks are Improperly Learnable in Polynomial Time, Wainwright et al., 2017.
- Understanding neural networks requires rethinking generalization, Zhang et al, ICLR 2017.
Generative models
- Compressed sensing using generative models, Price et al., ICML 2017.
- Generation and Equilibrium in GANs, Arora et al, 2017
- Wasserstein GANs, Arjovsky, 2017.
- Stabilizing GANs via random projections, Neyshabur et al, 2017
- Global guarantees for Enforcing Deep Generative Priors, Hand et al, 2017.
- Towards Understanding the Dynamics of Generative Adversarial Networks, Li et al, 2017.