Prototypical Recurrent Unit

TL;DR

This paper introduces the prototypical recurrent unit (PRU), a simplified yet effective alternative to LSTM/GRU, enabling easier analysis of recurrent networks while maintaining comparable performance and exploring their memorization capabilities.

Contribution

The paper proposes the PRU, a minimalistic recurrent unit that captures key features of LSTM/GRU, facilitating theoretical analysis and understanding of recurrent neural networks.

Findings

PRU achieves performance comparable to LSTM and GRU.

Memorization ability depends on information quantity and state space size.

PRU provides a simpler model for studying long-term memory in recurrent networks.

Abstract

Despite the great successes of deep learning, the effectiveness of deep neural networks has not been understood at any theoretical depth. This work is motivated by the thrust of developing a deeper understanding of recurrent neural networks, particularly LSTM/GRU-like networks. As the highly complex structure of the recurrent unit in LSTM and GRU networks makes them difficult to analyze, our methodology in this research theme is to construct an alternative recurrent unit that is as simple as possible and yet also captures the key components of LSTM/GRU recurrent units. Such a unit can then be used for the study of recurrent networks and its structural simplicity may allow easier analysis. Towards that goal, we take a system-theoretic perspective to design a new recurrent unit, which we call the prototypical recurrent unit (PRU). Not only having minimal complexity, PRU is demonstrated experimentally to have comparable performance to GRU and LSTM unit. This establishes PRU networks as a prototype for future study of LSTM/GRU-like recurrent networks. This paper also studies the memorization abilities of LSTM, GRU and PRU networks, motivated by the folk belief that such networks possess long-term memory. For this purpose, we design a simple and controllable task, called ``memorization problem'', where the networks are trained to memorize certain targeted information. We show that the memorization performance of all three networks depends on the amount of targeted information, the amount of ``interfering" information, and the state space dimension of the recurrent unit. Experiments are also performed for another controllable task, the adding problem, and similar conclusions are obtained.