Learning long-term dependencies for action recognition with a biologically-inspired deep network

TL;DR

This paper introduces shuttleNet, a biologically-inspired deep network with loop connections and attention, designed to better learn long-term dependencies for action recognition, outperforming existing RNN variants.

Contribution

The paper proposes shuttleNet, a novel RNN architecture with loop connections and attention, inspired by biological neural systems, to improve long-term sequence learning.

Findings

ShuttleNet outperforms state-of-the-art methods on UCF101 and HMDB51 datasets.

Loop connections and attention enhance long-term dependency learning.

Embedding shuttleNet into CNN-RNN frameworks improves action recognition accuracy.

Abstract

Despite a lot of research efforts devoted in recent years, how to efficiently learn long-term dependencies from sequences still remains a pretty challenging task. As one of the key models for sequence learning, recurrent neural network (RNN) and its variants such as long short term memory (LSTM) and gated recurrent unit (GRU) are still not powerful enough in practice. One possible reason is that they have only feedforward connections, which is different from the biological neural system that is typically composed of both feedforward and feedback connections. To address this problem, this paper proposes a biologically-inspired deep network, called shuttleNet\footnote{Our code is available at \url{https://github.com/shiyemin/shuttlenet}}. Technologically, the shuttleNet consists of several processors, each of which is a GRU while associated with multiple groups of cells and states. Unlike traditional RNNs, all processors inside shuttleNet are loop connected to mimic the brain's feedforward and feedback connections, in which they are shared across multiple pathways in the loop connection. Attention mechanism is then employed to select the best information flow pathway. Extensive experiments conducted on two benchmark datasets (i.e UCF101 and HMDB51) show that we can beat state-of-the-art methods by simply embedding shuttleNet into a CNN-RNN framework.