Temporal Attention-Gated Model for Robust Sequence Classification

TL;DR

The paper introduces the Temporal Attention-Gated Model (TAGM), a novel approach combining attention mechanisms and gated recurrent networks to improve robustness and interpretability in sequence classification tasks involving noisy or unsegmented data.

Contribution

It extends attention models with a gated recurrent network to handle noisy sequences, providing both improved accuracy and interpretability.

Findings

Enhanced prediction accuracy on three diverse tasks

Provides meaningful interpretability through attention weights

Effective in handling noisy, unsegmented sequences

Abstract

Typical techniques for sequence classification are designed for well-segmented sequences which have been edited to remove noisy or irrelevant parts. Therefore, such methods cannot be easily applied on noisy sequences expected in real-world applications. In this paper, we present the Temporal Attention-Gated Model (TAGM) which integrates ideas from attention models and gated recurrent networks to better deal with noisy or unsegmented sequences. Specifically, we extend the concept of attention model to measure the relevance of each observation (time step) of a sequence. We then use a novel gated recurrent network to learn the hidden representation for the final prediction. An important advantage of our approach is interpretability since the temporal attention weights provide a meaningful value for the salience of each time step in the sequence. We demonstrate the merits of our TAGM approach, both for prediction accuracy and interpretability, on three different tasks: spoken digit recognition, text-based sentiment analysis and visual event recognition.