---
title: "DELTA: Disentangled Hierarchical Interaction and Adaptive Adjustment for Emotion and Intent Understanding in Multimodal Conversations"
authors:
  - "Shenjie Jiang"
  - "Xiangfeng Liu"
  - "Xianghua Li"
  - "Xuecheng Wu"
doi: "10.24963/ijcai.2026/301"
url: "https://ijcai.org/proceedings/2026/301"
pdf: "https://ijcai.org/proceedings/2026/0301.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "2708-2716"
track: "Main Track"
keywords:
  - "Search: Applications"
  - "Data Mining: Applications"
  - "Machine Learning: Applications"
  - "Knowledge Representation and Reasoning: Applications"
---

# DELTA: Disentangled Hierarchical Interaction and Adaptive Adjustment for Emotion and Intent Understanding in Multimodal Conversations

Shenjie Jiang, Xiangfeng Liu, Xianghua Li, Xuecheng Wu

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 2708-2716. DOI: [10.24963/ijcai.2026/301](https://doi.org/10.24963/ijcai.2026/301)

## Abstract

Emotion and intent joint understanding in multimodal conversations (MC-EIU) aims to infer emotion and intent information by leveraging the semantic dependencies among multimodal data. However, prior works largely ignore redundancy interference and suffer from insufficient interaction and inter-task noise propagation due to their reliance on shallow mechanisms. To overcome these limitations, we propose a novel framework named Disentangled Hierarchical Interaction and Adaptive Adjustment (DELTA) for MC-EIU. We first design a Disentangled Feature Denoising module based on orthogonal decomposition to effectively filter out noise and redundancy from heterogeneous data. Second, we propose a Dual-Level Adaptive Adjustment mechanism to dynamically optimize learning dynamics from both modality and instance perspectives. Furthermore, we introduce a Hierarchical Task-Bottleneck Interaction(HTBI) module, which employs a set of progressively compressed bottleneck tokens as a communication hub. This design can effectively simulating the multi-round iterative interaction between emotion and intent tasks. Experiments on the benchmark MC-EIU bilingual dataset demonstrate that our framework significantly outperforms state-of-the-art baselines in both emotion and intent tasks.