DanceHA: A Multi-Agent Framework for Document-Level Aspect-Based Sentiment Analysis
arXiv:2603.16546v1 Announce Type: new Abstract: Aspect-Based Sentiment Intensity Analysis (ABSIA) has garnered increasing attention, though research largely focuses on domain-specific, sentence-level settings. In contrast, document-level ABSIA--particularly in addressing complex tasks like extracting Aspect-Category-Opinion-Sentiment-Intensity (ACOSI) tuples--remains underexplored. In this work, we introduce DanceHA, a multi-agent framework designed for open-ended, document-level ABSIA with informal writing styles. DanceHA has two main components: Dance, which employs a divide-and-conquer strategy to decompose the long-context ABSIA task into smaller, manageable sub-tasks for collaboration among specialized agents; and HA, Human-AI collaboration for annotation. We release Inf-ABSIA, a multi-domain document-level ABSIA dataset featuring fine-grained and high-accuracy labels from DanceHA. Extensive experiments demonstrate the effectiveness of our agentic framework and show that the mult
arXiv:2603.16546v1 Announce Type: new Abstract: Aspect-Based Sentiment Intensity Analysis (ABSIA) has garnered increasing attention, though research largely focuses on domain-specific, sentence-level settings. In contrast, document-level ABSIA--particularly in addressing complex tasks like extracting Aspect-Category-Opinion-Sentiment-Intensity (ACOSI) tuples--remains underexplored. In this work, we introduce DanceHA, a multi-agent framework designed for open-ended, document-level ABSIA with informal writing styles. DanceHA has two main components: Dance, which employs a divide-and-conquer strategy to decompose the long-context ABSIA task into smaller, manageable sub-tasks for collaboration among specialized agents; and HA, Human-AI collaboration for annotation. We release Inf-ABSIA, a multi-domain document-level ABSIA dataset featuring fine-grained and high-accuracy labels from DanceHA. Extensive experiments demonstrate the effectiveness of our agentic framework and show that the multi-agent knowledge in DanceHA can be effectively transferred into student models. Our results highlight the importance of the overlooked informal styles in ABSIA, as they often intensify opinions tied to specific aspects.
Executive Summary
The article introduces DanceHA, a multi-agent framework for document-level aspect-based sentiment analysis. DanceHA consists of two main components: Dance, which decomposes the task into smaller sub-tasks, and HA, which enables human-AI collaboration for annotation. The framework is evaluated on the newly released Inf-ABSIA dataset, demonstrating its effectiveness in extracting fine-grained aspect-category-opinion-sentiment-intensity tuples. The results highlight the importance of considering informal writing styles in aspect-based sentiment analysis.
Key Points
- ▸ Introduction of DanceHA, a multi-agent framework for document-level ABSIA
- ▸ Decomposition of the task into smaller sub-tasks using a divide-and-conquer strategy
- ▸ Release of the Inf-ABSIA dataset featuring fine-grained and high-accuracy labels
Merits
Effective Framework
DanceHA demonstrates high effectiveness in extracting ACOSI tuples, outperforming existing methods
Human-AI Collaboration
The HA component enables effective human-AI collaboration for annotation, improving the accuracy of the framework
Demerits
Limited Domain Adaptability
The framework's performance may be limited to the specific domains and datasets it was trained on
Computational Complexity
The multi-agent architecture may increase computational complexity, potentially limiting its scalability
Expert Commentary
The introduction of DanceHA marks a significant advancement in document-level aspect-based sentiment analysis. The framework's ability to decompose the task into smaller sub-tasks and enable human-AI collaboration for annotation demonstrates a nuanced understanding of the complexities involved in ABSIA. However, further research is needed to address the potential limitations of the framework, such as its adaptability to different domains and datasets. Nevertheless, DanceHA has the potential to significantly improve the accuracy and effectiveness of sentiment analysis in various applications.
Recommendations
- ✓ Further evaluation of DanceHA on diverse datasets to assess its domain adaptability
- ✓ Investigation of methods to reduce the computational complexity of the multi-agent architecture