Dialogue State Tracking with a Language Model using Schema-Driven Prompting
By Chia-Hsuan Lee, Hao Cheng, Mari Ostendorf. 2021
148 citations recorded by the source corpus.
Task-oriented conversational systems often use dialogue state tracking to represent the user’s intentions, which involves filling in values of pre-defined slots. Many approaches have been proposed, often using task-specific architectures with special-purpose classifiers. Recently, good results have been obtained using more general architectures based on pretrained language models. Here, we introduce a new variation of the language modeling approach that uses schema-driven prompting to provide task-aware history encoding that is used for both categorical and non-categorical slots. We further improve performance by augmenting the prompting with schema descriptions, a naturally occurring source of in-domain knowledge. Our purely generative system achieves state-of-the-art performance on MultiWOZ 2.2 and achieves competitive performance on two other benchmarks: MultiWOZ 2.1 and M2M. The data and code will be available at https://github.com/chiahsuan156/DST-as-Prompting.
Subfields: Artificial Intelligence, Large Language Models, Dialogue & Conversational AI
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Caiming Xiong
- Luke S. Zettlemoyer
- Percy Liang
- Graham Neubig
- Richard Socher
- Noah A. Smith
- Mohit Bansal
- Jianfeng Gao
In Large Language Models
- Noah A. Smith
- Caiming Xiong
- Luke S. Zettlemoyer
- Yue Zhang
- Dragomir R. Radev
- Michihiro Yasunaga
- Weijia Shi
- Lingpeng Kong
In Dialogue & Conversational AI
Citations in the corpus
- UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models (2022, methodology): “∗ T5-base with the independent output scheme (Lee et al., 2021) achieves 56.66 on MWoZ2.”
- Dynamic Dialogue Policy for Continual Reinforcement Learning (2022, background): “…assume that the agent has access to an external dictionary providing a natural language description descrf of each of these, e.g. “area of the hotel” or “number of hotel database results”, which is common in dialogue state tracking (Rastogi et al., 2020; van Niekerk et al., 2021; Lee et al., 2021).”
- Do Prompts Solve NLP Tasks Using Natural Language? (2022, background): “In contrast, schema prompts (Lee et al., 2021; Paolini et al., 2021) replace a natural language sentence with a structured schema, which makes the prompt more succinct and codelike.”
- PromptCap: Prompt-Guided Task-Aware Image Captioning (2022, background): “Prompting has been successful for a variety of natural language tasks [37], including but not limited to classification tasks [42, 56], semantic parsing [44, 74], knowledge generation [35, 57], and dialogue systems [20, 28].”