InstructERC: Reforming Emotion Recognition in Conversation with a Multi-task Retrieval-based LLMs Framework
Core Concepts
InstructERC transforms emotion recognition tasks from discriminative to generative frameworks using Large Language Models (LLMs), achieving state-of-the-art results.
Abstract
The article introduces InstructERC, a novel approach for emotion recognition in conversation.
It proposes a retrieval template module and emotional alignment tasks to improve performance.
InstructERC outperforms previous models on three benchmarks.
Unified dataset experiments show robustness and effectiveness of InstructERC.
Data scaling analysis demonstrates the impact of different data mixing strategies on model performance.
InstructERC
Stats
"Our LLM-based plugin framework significantly outperforms all previous models and achieves comprehensive SOTA on three commonly used ERC datasets."
"Extensive analysis of parameter-efficient and data-scaling experiments provides empirical guidance for applying it in practical scenarios."
Quotes
"The question is not whether intelligent machines can have emotions, but whether machines without emotions can achieve intelligence." - Minsky, 1988