ACL Findings2026
Fast Retrieval and Slow Reasoning for Explainable Multimodal Sentiment Analysis
Aoqiang Zhu, Min Hu, Yan Xing
Abstract
Most existing Multimodal Sentiment Analysis (MSA) methods rely on holistic fusion, treating all modalities and temporal segments equally. Such strategies often introduce redundant information and obscure the decision process, limiting both robustness and interpretability. Inspired by dual-process theory, we propose FRSR (Fast Retrieval and Slow Reasoning), an interpretable framework that decomposes multimodal sentiment modeling into two cooperative pathways. The Fast Pathway acts as a lightweight evidence selector, using contextaware convolution and auxiliary supervision to retrieve a sparse set of Top-K sentimentrelevant cues from noisy multimodal inputs. Based on these cues, the Slow Pathway performs deeper cross-modal reasoning through learnable reasoning tokens, enabling hierarchical sentiment inference. By separating salient evidence retrieval from multimodal reasoning, FRSR improves interpretability while reducing computational cost. Experiments on three benchmark datasets show that FRSR achieves competitive performance, higher efficiency, stronger robustness to noise, and clearer decision transparency than existing holistic fusion methods.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- Learning Modality-Specific Representations with Self-Supervised Multi-Task Learning for Multimodal Sentiment Analysis cites MISA: Modality-Invariant and -Specific Representations for Multimodal Sentiment Analysis
- Learning Modality-Specific Representations with Self-Supervised Multi-Task Learning for Multimodal Sentiment Analysis cites CH-SIMS: A Chinese Multimodal Sentiment Analysis Dataset with Fine-grained Annotation of Modality
Builds on17
- MISA: Modality-Invariant and -Specific Representations for Multimodal Sentiment AnalysisDevamanyu Hazarika, Roger Zimmermann, Soujanya PoriaACM MM 2020 · 1,037 citations
- Learning Modality-Specific Representations with Self-Supervised Multi-Task Learning for Multimodal Sentiment AnalysisWenmeng Yu, Hua Xu, Ziqi Yuan, Jiele WuAAAI 2021 · 737 citations
- CH-SIMS: A Chinese Multimodal Sentiment Analysis Dataset with Fine-grained Annotation of ModalityWenmeng Yu, Hua Xu, Fanyang Meng, Yilin Zhu et al.ACL 2020 · 376 citations
- Transformer-based Feature Reconstruction Network for Robust Multimodal Sentiment AnalysisZiqi Yuan, Wei Li, Hua Xu, Wenmeng YuACM MM 2021 · 186 citations
- Provable Dynamic Fusion for Low-Quality Multimodal DataQingyang Zhang, Haitao Wu, Changqing Zhang, Qinghua Hu et al.ICML 2023 · 143 citations
Related papers
- Factorize, Reconstruct, Enhance: A Unified Framework for Multimodal Sentiment AnalysisZhilu Yang, Mingcheng LiCVPR 2026
- FINE: Factorized Multimodal Sentiment Analysis via Mutual INformation EstimationYadong Liu, Shangfei WangAAAI 2026
- DDSE: A Decoupled Dual-Stream Enhanced Framework for Multimodal Sentiment Analysis with Text-Centric SSMShenjie Jiang, Zhuoyu Wang, Xuecheng Wu, Hongru Ji et al.ACM MM 2025 · 4 citations
- ConsMSA: Semantic Distribution Consistency Learning for Multimodal Sentiment AnalysisPan Wang, Lipeng Ke, Huajun Ying, Pritish Mohapatra et al.ICML 2026
- Resource-Limited Joint Multimodal Sentiment Reasoning and Classification via Chain-of-Thought Enhancement and DistillationHaonan Shangguan, Xiaocui Yang, Shi Feng, Daling Wang et al.ACL Findings 2026 · 2 citations
