Saved in:
Bibliographic Details
Main Authors: Wang, Yijin, Sun, Fandi, Wen, Haoyu
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.28205
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910217955966976
author Wang, Yijin
Sun, Fandi
Wen, Haoyu
author_facet Wang, Yijin
Sun, Fandi
Wen, Haoyu
contents Aspect-Based Sentiment Analysis (ABSA) faces critical challenges due to representation entanglement and false-negative collisions in real-valued embedding spaces. In this paper, we propose a novel framework featuring a Zero-Initialized Residual Complex Projection (ZRCP) and an Anti-collision Masked Angle Loss. Our approach projects textual features into a complex semantic space, utilizing the phase to isolate sentiment polarities while regularizing the amplitude to ensure structural consistency within aspect categories. To mitigate this, we introduce an anti-collision mask that preserves intra-polarity aspect cohesion while significantly expanding the discriminative margin between opposing polarities. Experimental results on the ASAP dataset demonstrate that our framework achieves a state-of-the-art Macro-F1 score of 0.8923, outperforming robust baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2603_28205
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Beyond Cosine Similarity: Zero-Initialized Residual Complex Projection for Aspect-Based Sentiment Analysis
Wang, Yijin
Sun, Fandi
Wen, Haoyu
Computation and Language
Aspect-Based Sentiment Analysis (ABSA) faces critical challenges due to representation entanglement and false-negative collisions in real-valued embedding spaces. In this paper, we propose a novel framework featuring a Zero-Initialized Residual Complex Projection (ZRCP) and an Anti-collision Masked Angle Loss. Our approach projects textual features into a complex semantic space, utilizing the phase to isolate sentiment polarities while regularizing the amplitude to ensure structural consistency within aspect categories. To mitigate this, we introduce an anti-collision mask that preserves intra-polarity aspect cohesion while significantly expanding the discriminative margin between opposing polarities. Experimental results on the ASAP dataset demonstrate that our framework achieves a state-of-the-art Macro-F1 score of 0.8923, outperforming robust baselines.
title Beyond Cosine Similarity: Zero-Initialized Residual Complex Projection for Aspect-Based Sentiment Analysis
topic Computation and Language
url https://arxiv.org/abs/2603.28205