Exploiting Adaptive Contextual Masking for Aspect-Based Sentiment Analysis

2024-02-21

Pacific-Asia Conference on Knowledge Discovery and Data Mining

AI Summary

Existing ABSA methods employ static attention masks, limiting their ability to handle complex contexts with multiple co-occurring aspects and conflicting sentiments, thereby constraining the accuracy of aspect term extraction (ATE) and aspect sentiment classification (ASC). To address this, we propose a semantics-driven adaptive context masking mechanism: instead of fixed masks, we design a learnable, context-aware mask generation module that dynamically selects salient tokens based on input semantics in real time; further, we develop a Transformer-based joint ATE+ASC framework. This mechanism enables fine-grained, context-dependent dynamic attention modulation—the first of its kind in ABSA. Evaluated on four benchmark datasets, our approach achieves average improvements of +2.3% in F1-score and +1.9% in accuracy over strong baselines. Qualitative analysis confirms its capacity to precisely focus on aspect-relevant textual spans.

Technology Category

• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Cognitive Modeling & Cognitive Systems: Affective Computing
• Multiagent Systems: Agreement, Argumentation & Negotiation

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search
• Web Mining and Content Analysis: Sentiment analysis and opinion mining

Abstract

Aspect-Based Sentiment Analysis (ABSA) is a fine-grained linguistics problem that entails the extraction of multifaceted aspects, opinions, and sentiments from the given text. Both standalone and compound ABSA tasks have been extensively used in the literature to examine the nuanced information present in online reviews and social media posts. Current ABSA methods often rely on static hyperparameters for attention-masking mechanisms, which can struggle with context adaptation and may overlook the unique relevance of words in varied situations. This leads to challenges in accurately analyzing complex sentences containing multiple aspects with differing sentiments. In this work, we present adaptive masking methods that remove irrelevant tokens based on context to assist in Aspect Term Extraction and Aspect Sentiment Classification subtasks of ABSA. We show with our experiments that the proposed methods outperform the baseline methods in terms of accuracy and F1 scores on four benchmark online review datasets. Further, we show that the proposed methods can be extended with multiple adaptations and demonstrate a qualitative analysis of the proposed approach using sample text for aspect term extraction.

Problem

• Improving aspect term extraction in sentiment analysis
• Enhancing context adaptation in attention-masking mechanisms
• Accurately analyzing sentences with multiple aspects and sentiments

Innovation

• Adaptive masking for context-aware sentiment analysis
• Dynamic token removal for aspect extraction
• Improved accuracy in multi-aspect sentences