Homework 4

Submission date: August 26st, 2021

Topics

• Deep reinforcement learning based on policy gradients.
• Course summary questions on various topics.
• Mini project, either:
  • Advanced GANs
  • Sentiment analysis with self-attention

Downloading

The assignment is available here (use the download button at the top).

Technical Notes for Part 1

• Part 1 does not require a GPU. We won’t need large models, and the computation bottleneck will be the generation of episodes to train on.
• The OpenAI gym library is not officially supported on windows. It should be possible to install and run the necessary environment for this exercise. However, we cannot provide you with technical support for Windows. If you have trouble installing locally, please use the course server.

FAQ

Make sure to read the getting started page, the guide for using course servers and our collaboration policy before starting the assignment.

Q: What is the _final checkpoint file?
A: You must use this to create your final submission with result video from your best-trained model. When you get results that your happy with, rename the checkpoint file by appending _final. You don’t need to submit the checkpoints/ folder (the main.py script will ignore them).

Q: Should the results/ directory be part of the submission?
A: Yes. The submission script will include it for you. This is OK. Do not put any unnecessary files in this directory apart from the results files generated by the notebooks.

Q: Should the project/ directory be part of the submission?
A: Yes. You should include here the source files necessary to implement your project. Do not put any unnecessary files in this directory apart from the source files and result files you need to load for display in the project notebook.