# Learning distant cause and effect using only local and immediate credit assignment

## TL;DR

This paper introduces a biologically-plausible recurrent neural network with sparse coding that effectively learns distant cause-and-effect relationships, predicts sequences, and aids navigation with less memory than traditional models.

## Contribution

The proposed network uses local, immediate credit assignment and sparse coding to encode sequences, enabling learning of long-range dependencies with reduced memory requirements.

## Key findings

- Successfully associates distant causes and effects in stochastic processes.
- Predicts higher-order sequences with partial observability.
- Reduces memory usage compared to LSTM, GRU, and autoregressive models.

## Abstract

We present a recurrent neural network memory that uses sparse coding to create a combinatoric encoding of sequential inputs. Using several examples, we show that the network can associate distant causes and effects in a discrete stochastic process, predict partially-observable higher-order sequences, and enable a DQN agent to navigate a maze by giving it memory. The network uses only biologically-plausible, local and immediate credit assignment. Memory requirements are typically one order of magnitude less than existing LSTM, GRU and autoregressive feed-forward sequence learning models. The most significant limitation of the memory is generalization to unseen input sequences. We explore this limitation by measuring next-word prediction perplexity on the Penn Treebank dataset.