LineConGraphs: Line Conversation Graphs for Speaker-Independent Emotion Recognition in Conversations

Published

November 1, 2025

Author(s)

Sarala Padi, Ram Sriram, Craig Greenberg, Gokul S Krishnan, Balaraman Ravindran, Dinesh Manocha

Abstract

Emotion Recognition in Conversations (ERC) is a critical aspect of affective computing, and it has many practical applications in healthcare, education, chatbots, and social media platforms. Earlier approaches to ERC analysis involved modeling speaker and contextual information using graph neural network architectures. However, these methods have limitations as the speaker's identity is unavailable, or it is not practical to build and deploy speaker-independent models for real-time applications. Additionally, long-distance context can confuse recognizing the emotion of an utterance in a conversation. To overcome these limitations, we propose novel line conversation graph convolutional network (LineConGCN) and graph attention (LineConGAT) models for ERC analysis. These models are speaker-independent and developed on the line conversation graphs (LineConGraphs). The context in LineConGraphs is limited to one previous and future utterance, and speaker information is not part of the graph. We evaluate the performance of our proposed models on two benchmark datasets, IEMOCAP and MELD, and show that our LineConGAT model outperforms the state-of-the-art methods with an F1-score of 64.58 and 76.50. Moreover, we demonstrate that embedding sentiment shift information into line conversation graphs further enhances the ERC performance.

Citation

IEEE Transactions on Affective Computing

Pub Type

Journals

Download Paper

Local Download

Keywords

Line Conversation Graphs, Emotion Recognition in Conversations, Graph Neural Networks, Graph Convolutional Networks, Graph Attention Networks, IEMOCAP, MELD, Sentiment Analysis, Emotion Recognition

Citation

Padi, S. , Sriram, R. , Greenberg, C. , S Krishnan, G. , Ravindran, B. and Manocha, D. (2025), LineConGraphs: Line Conversation Graphs for Speaker-Independent Emotion Recognition in Conversations, IEEE Transactions on Affective Computing, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=956726 (Accessed January 9, 2026)