A living collection of deep learning problems
Switch branches/tags
Nothing to show
Clone or download
gdb Merge pull request #33 from hyperdo/master
Add natural gradient Q-learning solution
Latest commit b02d21f Feb 21, 2018
Failed to load latest commit information.
_requests_for_research Update link to paper Feb 22, 2018
.gitignore remove temp file Jul 1, 2016
Makefile Initial release Jun 8, 2016
README.md Initial release Jun 8, 2016
_config.yml Initial release Jun 8, 2016
index.html Initial release Jun 8, 2016


Requests for Research

It's easy to get started in deep learning, with many resources to learn the latest techniques. But it's harder to know what problems are worth working on.

This repository contains a living collection of important and fun problems to help new people enter the field, and for enthusiastic practitioners to hone their skills. Many will require inventing new ideas.

If you've solved a problem

Please write up the problem in a Gist or paper, and open a pull request linking it in a "solutions" section for the relevant problem. (Alternatively, let us know about it in community chat.)

The best solutions will contain both code and an explanation of your methodology. Please also feel free to report things you tried that didn't work, or anything else helpful to someone trying to learn how to do their own deep learning research.

We'll accept multiple solutions to each problem, so long as each solution is materially different.

This repository

This respository hosts the source for the requests for research. Feel free to open a pull request. Especially encouraged are:

  • Suggestions for new problems
  • Suggestions for improvements to existing problems
  • Links to your solution.

Running this repo locally

Install or upgrade jekyll via gem install jekyll. You can run this repo locally via:

jekyll serve -w

Your content will then be available at It won't be styled, but that should be enough to get started.