Sequential Domain Adaptation through Elastic Weight Consolidation for Sentiment Analysis

TL;DR

This paper introduces Sequential Domain Adaptation (SDA), a framework using Elastic Weight Consolidation to improve sentiment analysis across multiple domains, outperforming complex models with simple architectures.

Contribution

The paper presents a novel, model-independent SDA framework leveraging EWC for effective sequential domain adaptation in sentiment analysis.

Findings

SDA enables simple architectures like CNNs to outperform state-of-the-art models.

Harder first Anti-Curriculum ordering of source domains enhances performance.

SDA effectively transfers knowledge across domains for sentiment analysis.

Abstract

Elastic Weight Consolidation (EWC) is a technique used in overcoming catastrophic forgetting between successive tasks trained on a neural network. We use this phenomenon of information sharing between tasks for domain adaptation. Training data for tasks such as sentiment analysis (SA) may not be fairly represented across multiple domains. Domain Adaptation (DA) aims to build algorithms that leverage information from source domains to facilitate performance on an unseen target domain. We propose a model-independent framework - Sequential Domain Adaptation (SDA). SDA draws on EWC for training on successive source domains to move towards a general domain solution, thereby solving the problem of domain adaptation. We test SDA on convolutional, recurrent, and attention-based architectures. Our experiments show that the proposed framework enables simple architectures such as CNNs to outperform complex state-of-the-art models in domain adaptation of SA. In addition, we observe that the effectiveness of a harder first Anti-Curriculum ordering of source domains leads to maximum performance.