Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FaiMA: Feature-aware In-context Learning for Multi-domain Aspect-based Sentiment Analysis

Songhua Yang, Xinke Jiang, Hanjie Zhao, Wenxuan Zeng, Hongde Liu, Yuxiang Jia

TL;DR

This work proposes a novel framework, Feature-aware In-context Learning for Multi-domain ABSA (FaiMA), to utilize in-context learning (ICL) as a feature-aware mechanism that facilitates adaptive learning in multi-domain ABSA tasks.

Abstract

Multi-domain aspect-based sentiment analysis (ABSA) seeks to capture fine-grained sentiment across diverse domains. While existing research narrowly focuses on single-domain applications constrained by methodological limitations and data scarcity, the reality is that sentiment naturally traverses multiple domains. Although large language models (LLMs) offer a promising solution for ABSA, it is difficult to integrate effectively with established techniques, including graph-based models and linguistics, because modifying their internal architecture is not easy. To alleviate this problem, we propose a novel framework, Feature-aware In-context Learning for Multi-domain ABSA (FaiMA). The core insight of FaiMA is to utilize in-context learning (ICL) as a feature-aware mechanism that facilitates adaptive learning in multi-domain ABSA tasks. Specifically, we employ a multi-head graph attention network as a text encoder optimized by heuristic rules for linguistic, domain, and sentiment features. Through contrastive learning, we optimize sentence representations by focusing on these diverse features. Additionally, we construct an efficient indexing mechanism, allowing FaiMA to stably retrieve highly relevant examples across multiple dimensions for any given input. To evaluate the efficacy of FaiMA, we build the first multi-domain ABSA benchmark dataset. Extensive experimental results demonstrate that FaiMA achieves significant performance improvements in multiple domains compared to baselines, increasing F1 by 2.07% on average. Source code and data sets are anonymously available at https://github.com/SupritYoung/FaiMA.

FaiMA: Feature-aware In-context Learning for Multi-domain Aspect-based Sentiment Analysis

TL;DR

This work proposes a novel framework, Feature-aware In-context Learning for Multi-domain ABSA (FaiMA), to utilize in-context learning (ICL) as a feature-aware mechanism that facilitates adaptive learning in multi-domain ABSA tasks.

Abstract

Multi-domain aspect-based sentiment analysis (ABSA) seeks to capture fine-grained sentiment across diverse domains. While existing research narrowly focuses on single-domain applications constrained by methodological limitations and data scarcity, the reality is that sentiment naturally traverses multiple domains. Although large language models (LLMs) offer a promising solution for ABSA, it is difficult to integrate effectively with established techniques, including graph-based models and linguistics, because modifying their internal architecture is not easy. To alleviate this problem, we propose a novel framework, Feature-aware In-context Learning for Multi-domain ABSA (FaiMA). The core insight of FaiMA is to utilize in-context learning (ICL) as a feature-aware mechanism that facilitates adaptive learning in multi-domain ABSA tasks. Specifically, we employ a multi-head graph attention network as a text encoder optimized by heuristic rules for linguistic, domain, and sentiment features. Through contrastive learning, we optimize sentence representations by focusing on these diverse features. Additionally, we construct an efficient indexing mechanism, allowing FaiMA to stably retrieve highly relevant examples across multiple dimensions for any given input. To evaluate the efficacy of FaiMA, we build the first multi-domain ABSA benchmark dataset. Extensive experimental results demonstrate that FaiMA achieves significant performance improvements in multiple domains compared to baselines, increasing F1 by 2.07% on average. Source code and data sets are anonymously available at https://github.com/SupritYoung/FaiMA.
Paper Structure (22 sections, 14 equations, 3 figures, 4 tables)

This paper contains 22 sections, 14 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Problem Definition
  5. Multi-head Graph Attention Network Encoder
  6. Feature Selection and Heuristic Rules
  7. Multi-head Graph Attention Network
  8. Contrastive Learning
  9. Feature-aware Example Retrieval
  10. Experiments
  11. Dataset
  12. Baselines
  13. Experimental Settings
  14. Main Results
  15. Ablation Study
  16. ...and 7 more sections

Figures (3)

  • Figure 1: An example of feature-aware in-context learning for ABSA. By selecting one relevant example on each of the three features, sufficient reference is provided for LLM.
  • Figure 2: The overall architecture of FaiMA: MGATE training part and example retrieval part. MGATE training involves three steps: heuristic rules for positive/negative pairs generation, multi-head graph attention network to embed sentences upon three features, and contrastive learning. The diagram to the far right illustrates an ICL process that reliably fetches three domain-relevant and global average samples for any input sentence.
  • Figure 3: The retrieval success rate of the three relevant feature examples retrieved for each domain.