Learning Associative Inference Using Fast Weight Memory

This repository contains the offical code for the paper Learning Associative Inference Using Fast Weight Memory published at ICLR 2021. The repo is split into three parts with three independent code bases: catbAbI, language modelling on PTB and WT2, and our meta-RL toy environment with randomly sampled POMDPs. Each folder contains a readme with further details and instructions.

Citation

@inproceedings{
schlag2021learning,
title={Learning Associative Inference Using Fast Weight Memory},
author={Imanol Schlag and Tsendsuren Munkhdalai and J{\"u}rgen Schmidhuber},
booktitle={International Conference on Learning Representations},
year={2021},
url={https://openreview.net/forum?id=TuK6agbdt27}
}