Summary of Large Language Models As Zero-shot Dialogue State Tracker Through Function Calling, by Zekun Li et al.
Large Language Models as Zero-shot Dialogue State Tracker through Function Calling
by Zekun Li, Zhiyu Zoey Chen, Mike Ross, Patrick Huber, Seungwhan Moon, Zhaojiang Lin, Xin Luna Dong, Adithya Sagar, Xifeng Yan, Paul A. Crook
First submitted to arxiv on: 16 Feb 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 The proposed FnCTOD approach utilizes function calling to improve zero-shot dialogue state tracking (DST) in large language models (LLMs). This method enables adaptation to diverse domains without extensive data collection or model tuning. The experimental results demonstrate exceptional performance, surpassing the previous state-of-the-art achieved by ChatGPT and improving its average joint goal accuracy (JGA) by 5.6%. The individual model results for GPT-3.5 and GPT-4 are boosted by 4.8% and 14%, respectively. Additionally, fine-tuning on a small collection of diverse task-oriented dialogues enables modestly sized models to achieve DST performance comparable to ChatGPT while maintaining their chat capabilities. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary FnCTOD is a new way to make large language models better at understanding what’s going on in conversations that are working towards specific goals. It does this by using “function calling” which helps the model keep track of the conversation’s state without needing lots of extra data or fine-tuning. The results show that this approach works really well, beating the previous best result and improving how well the models can understand what’s going on in conversations. |
Keywords
» Artificial intelligence » Fine tuning » Gpt » Tracking » Zero shot
