Why does machine learning work so well, and what are the theoretical constraints on what it can learn?
A Universal Law of Robustness via Isoperimetry
Sébastien Bubeck, Mark Sellke
Training Neural Networks is ER-complete
Mikkel Abrahamsen, L. Kleist, Tillmann Miltzow
Understanding deep learning requires rethinking generalization
Chiyuan Zhang, Samy Bengio, Moritz Hardt
The Lack of A Priori Distinctions Between Learning Algorithms
Draft : Deep Learning in Neural Networks : An Overview
On the Number of Linear Regions of Deep Neural Networks
Guido Montúfar, Razvan Pascanu, Kyunghyun Cho
An exact mapping between the Variational Renormalization Group and Deep Learning
Pankaj Mehta, D. Schwab
Deep learning via Hessian-free optimization
Why Does Deep and Cheap Learning Work So Well?
Henry W. Lin, Max Tegmark
1 Efficient BackProp
Yann LeCun, L. Bottou, G. Orr
Neural Networks and the Bias/Variance Dilemma
Stuart Geman, Elie Bienenstock, Rene Doursat
A Theoretically Grounded Application of Dropout in Recurrent Neural Networks
Y. Gal, Zoubin Ghahramani
The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks
Jonathan Frankle, Michael Carbin
What's hidden in the hidden layers?
D. Touretzky, D. Pomerleau
The Description Length of Deep Learning models
Léonard Blier, Y. Ollivier
Provable Bounds for Learning Some Deep Representations
Sanjeev Arora, Aditya Bhaskara, Rong Ge
Bottom-up Deep Learning using the Hebbian Principle
Aseem Wadhwa, Upamanyu Madhow
Group theoretical methods in machine learning
Deep Learning and Quantum Entanglement: Fundamental Connections with Implications to Network Design
Yoav Levine, David Yakira, Nadav Cohen
Welcome to your learning pathway, "Theoretical concerns in machine learning".
You can add new resources by clicking the button. The resource will be added at the end of the pathway, or anywhere in the middle.
To add a scientific paper (research article), paste its title, then hit Return. The paper will be identified using Semantic Scholar so that you don't have to type in the year, authors, journal or abstract.
To add a web page, paste its URL and Learning Pathways will fetch the page title for you. You may need to edit the title if it is not correct.
To add other resources, enter their details manually.
Click on a resource to see details about it, or to edit it. Using the buttons at the bottom of the page, resources can be re-ordered and section headings can be added.