From Prediction to Justification: Aligning Sentiment Reasoning with Human Rationale via Reinforcement Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhang, Shihao, Wang, Ziwei, Zhou, Jie, Wu, Yulan, Chen, Qin, Lei, Zhikai, Yu, Liyang, Dou, Liang, He, Liang
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918447462481920
author Zhang, Shihao
Wang, Ziwei
Zhou, Jie
Wu, Yulan
Chen, Qin
Lei, Zhikai
Yu, Liyang
Dou, Liang
He, Liang
author_facet Zhang, Shihao
Wang, Ziwei
Zhou, Jie
Wu, Yulan
Chen, Qin
Lei, Zhikai
Yu, Liyang
Dou, Liang
He, Liang
contents While Aspect-based Sentiment Analysis (ABSA) systems have achieved high accuracy in identifying sentiment polarities, they often operate as "black boxes," lacking the explicit reasoning capabilities characteristic of human affective cognition. Humans do not merely categorize sentiment; they construct causal explanations for their judgments. To bridge this gap, we propose ABSA-R1, a large language model framework designed to mimic this ``reason-before-predict" cognitive process. By leveraging reinforcement learning (RL), ABSA-R1 learns to articulate the why behind the what, generating natural language justifications that ground its sentiment predictions. We introduce a Cognition-Aligned Reward Model (formerly sentiment-aware reward model) that enforces consistency between the generated reasoning path and the final emotional label. Furthermore, inspired by metacognitive monitoring, we implement a performance-driven rejection sampling strategy that selectively targets hard cases where the model's internal reasoning is uncertain or inconsistent. Experimental results on four benchmarks demonstrate that equipping models with this explicit reasoning capability not only enhances interpretability but also yields superior performance in sentiment classification and triplet extraction compared to non-reasoning baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2604_13398
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle From Prediction to Justification: Aligning Sentiment Reasoning with Human Rationale via Reinforcement Learning
Zhang, Shihao
Wang, Ziwei
Zhou, Jie
Wu, Yulan
Chen, Qin
Lei, Zhikai
Yu, Liyang
Dou, Liang
He, Liang
Computation and Language
Artificial Intelligence
While Aspect-based Sentiment Analysis (ABSA) systems have achieved high accuracy in identifying sentiment polarities, they often operate as "black boxes," lacking the explicit reasoning capabilities characteristic of human affective cognition. Humans do not merely categorize sentiment; they construct causal explanations for their judgments. To bridge this gap, we propose ABSA-R1, a large language model framework designed to mimic this ``reason-before-predict" cognitive process. By leveraging reinforcement learning (RL), ABSA-R1 learns to articulate the why behind the what, generating natural language justifications that ground its sentiment predictions. We introduce a Cognition-Aligned Reward Model (formerly sentiment-aware reward model) that enforces consistency between the generated reasoning path and the final emotional label. Furthermore, inspired by metacognitive monitoring, we implement a performance-driven rejection sampling strategy that selectively targets hard cases where the model's internal reasoning is uncertain or inconsistent. Experimental results on four benchmarks demonstrate that equipping models with this explicit reasoning capability not only enhances interpretability but also yields superior performance in sentiment classification and triplet extraction compared to non-reasoning baselines.
title From Prediction to Justification: Aligning Sentiment Reasoning with Human Rationale via Reinforcement Learning
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2604.13398