Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.20998 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913148212084736 |
|---|---|
| author | Xia, Yan Pan, Zhuangzhuang Kamsin, Amirrudin Chan, Chee Seng |
| author_facet | Xia, Yan Pan, Zhuangzhuang Kamsin, Amirrudin Chan, Chee Seng |
| contents | Aspect-Term Sentiment Analysis (ATSA) in multi-aspect sentences faces a fundamental tradeoff between efficiency and expressiveness. Existing models either re-encode the sentence for each aspect or rely on static use of deep representations, leading to redundant computation and limited adaptivity. We argue that Transformer depth is a costly, queryable resource, and propose DABS, a single-pass inference framework that encodes each sentence once to construct a reusable, depth-ordered substrate. Each aspect then queries this shared representation to selectively read relevant tokens and abstraction levels, without re-encoding. This decouples shared sentence encoding from lightweight, aspect-conditioned readout. Experiments on four ATSA benchmarks show that DABS achieves competitive performance while reducing end-to-end computation by up to 60% in multi-aspect settings (M >= 2). Further analyses indicate that adaptive depth querying is most beneficial for linguistically complex cases such as negation and contrast. Code is publicly available at https://github.com/panzhzh/acl-dabs |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_20998 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Single-Pass, Depth-Selective Reading for Multi-Aspect Sentiment Analysis Xia, Yan Pan, Zhuangzhuang Kamsin, Amirrudin Chan, Chee Seng Computation and Language Artificial Intelligence Aspect-Term Sentiment Analysis (ATSA) in multi-aspect sentences faces a fundamental tradeoff between efficiency and expressiveness. Existing models either re-encode the sentence for each aspect or rely on static use of deep representations, leading to redundant computation and limited adaptivity. We argue that Transformer depth is a costly, queryable resource, and propose DABS, a single-pass inference framework that encodes each sentence once to construct a reusable, depth-ordered substrate. Each aspect then queries this shared representation to selectively read relevant tokens and abstraction levels, without re-encoding. This decouples shared sentence encoding from lightweight, aspect-conditioned readout. Experiments on four ATSA benchmarks show that DABS achieves competitive performance while reducing end-to-end computation by up to 60% in multi-aspect settings (M >= 2). Further analyses indicate that adaptive depth querying is most beneficial for linguistically complex cases such as negation and contrast. Code is publicly available at https://github.com/panzhzh/acl-dabs |
| title | Single-Pass, Depth-Selective Reading for Multi-Aspect Sentiment Analysis |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2605.20998 |