Summary of Redefining Proactivity For Information Seeking Dialogue, by Jing Yang Lee et al.
Redefining Proactivity for Information Seeking Dialogue
by Jing Yang Lee, Seokhwan Kim, Kartik Mehta, Jiun-Yu Kao, Yu-Hsiang Lin, Arpit Gupta
First submitted to arxiv on: 20 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents a new definition of proactive dialogue for Information-Seeking Dialogue (ISD) agents. The existing definitions focus on reactive behavior, whereas this study introduces a proactivity-focused approach that enhances each generated response with new information related to the initial query. A dataset of 2,000 single-turn conversations is constructed and automatic metrics are introduced to evaluate response proactiveness, achieving high correlation with human annotation. Additionally, two innovative Chain-of-Thought (CoT) prompts, the 3-step CoT and the 3-in-1 CoT prompts, outperform standard prompts by up to 90% in the zero-shot setting. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study helps ISD agents have better conversations with users. Right now, these agents are good at answering simple questions, but they don’t know how to keep talking and make the conversation more interesting. The researchers propose a new way of thinking about proactive dialogue that makes each response more engaging and helps the conversation go on longer. They also create a special dataset and tools to measure how well the responses do this. This leads to some really good results, like being able to have conversations up to 90% better than before. |
Keywords
» Artificial intelligence » Zero shot
