Gated Feedback Recurrent Neural Networks

TL;DR

This paper introduces the gated-feedback RNN (GF-RNN), a new architecture that enhances deep recurrent networks by adaptively controlling inter-layer signals, leading to improved performance on language modeling and program evaluation tasks.

Contribution

The paper presents a novel GF-RNN architecture that allows adaptive gating of signals between layers, improving deep RNN performance over traditional stacked models.

Findings

GF-RNN outperforms conventional stacked RNNs in language modeling and code evaluation.

Adaptive gating enables layers to operate at different timescales.

Top-down layer interactions improve model flexibility and accuracy.

Abstract

In this work, we propose a novel recurrent neural network (RNN) architecture. The proposed RNN, gated-feedback RNN (GF-RNN), extends the existing approach of stacking multiple recurrent layers by allowing and controlling signals flowing from upper recurrent layers to lower layers using a global gating unit for each pair of layers. The recurrent signals exchanged between layers are gated adaptively based on the previous hidden states and the current input. We evaluated the proposed GF-RNN with different types of recurrent units, such as tanh, long short-term memory and gated recurrent units, on the tasks of character-level language modeling and Python program evaluation. Our empirical evaluation of different RNN units, revealed that in both tasks, the GF-RNN outperforms the conventional approaches to build deep stacked RNNs. We suggest that the improvement arises because the GF-RNN can adaptively assign different layers to different timescales and layer-to-layer interactions (including the top-down ones which are not usually present in a stacked RNN) by learning to gate these interactions.