Reproduce “Emotional Chatting Machine: Emotional Conversation Generation with Internal and External Memory” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hao Zhou, Minlie Huang, Tianyang Zhang, Xiaoyan Zhu, Bing Liu
The Emotional Chatting Machine (ECM) introduces a framework for generating emotionally appropriate and content-relevant responses in large-scale, open-domain conversations. This system, developed by researchers at Tsinghua University and the University of Illinois at Chicago, integrates emotion category embeddings with dynamic internal and external memory mechanisms to achieve a 0.773 emotion accuracy and strong human preference over baseline models.
Key insights
- Effectively integrating both implicit emotional states (Internal Memory) and explicit emotion word selection (External Memory) into a generation model is crucial for nuanced and controlled emotional response generation.
- Various appropriate emotional responses can stem from a single input post, highlighting the need for flexible emotion conditioning in conversational AI.
- Emotional context can be successfully integrated into large-scale, data-driven neural conversation models, enabling more human-like and empathetic interactions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.