Modern Self-Referential Weight Matrix

This is the official repository containing code for the paper:

A Modern Self-Referential Weight Matrix That Learns to Modify Itself

An earlier/shorter version of the paper (only containing the RL part) was presented at NeurIPS 2021 Deep RL Workshop. The corresponding version is available on Openreview.

General instructions

Please refer to the readme file under each directory for further instructions.

License files can be found under the corresponding directories.

In all tasks, our custom CUDA kernels will be automatically compiled. To avoid recompiling the code multiple times, we recommend to specify the path to a directory to store the compiled code via:

export TORCH_EXTENSIONS_DIR="/home/me/torch_extensions/rl"

BibTex

@article{irie2022modern,
  title={A Modern Self-Referential Weight Matrix That Learns to Modify Itself},
  author={Kazuki Irie and Imanol Schlag and R\'obert Csord\'as and J\"urgen Schmidhuber},
  journal={Preprint arXiv:2202.05780},
  year={2022}
}

If for some reason the workshop version is needed:

@inproceedings{irie2021modern,
      title={A Modern Self-Referential Weight Matrix That Learns to Modify Itself}, 
      author={Kazuki Irie and Imanol Schlag and R\'obert Csord\'as and J\"urgen Schmidhuber},
      booktitle={Workshop on Deep Reinforcement Learning, NeurIPS},
      address={Virtual only},
      year={2021}
}

Links

• This is a follow up to two of our previous works:
  • Linear Transformers are Secretly Fast Weight Programmers (ICML 2021)
  • Going Beyond Linear Transformers with Recurrent Fast Weight Programmers (NeurIPS 2021)
• Jürgen Schmidhuber's AI blog post on Fast Weight Programmers (March 26, 2021).