Natural Language Decomposition and Interpretation of Complex Utterances
This work addresses the problem of reducing the need for extensive supervised data in designing natural language interfaces for conversational assistants, representing an incremental improvement over existing methods.
The paper tackles the challenge of interpreting complex user utterances in natural language interfaces by introducing a hierarchical decomposition approach that uses a pre-trained language model to break down complex requests into simpler steps, achieving performance that outperforms standard few-shot prompting methods with minimal complex training data.
Designing natural language interfaces has historically required collecting supervised data to translate user requests into carefully designed intent representations. This requires enumerating and labeling a long tail of user requests, which is challenging. At the same time, large language models (LLMs) encode knowledge about goals and plans that can help conversational assistants interpret user requests requiring numerous steps to complete. We introduce an approach to handle complex-intent-bearing utterances from a user via a process of hierarchical natural language decomposition and interpretation. Our approach uses a pre-trained language model to decompose a complex utterance into a sequence of simpler natural language steps and interprets each step using the language-to-program model designed for the interface. To test our approach, we collect and release DeCU -- a new NL-to-program benchmark to evaluate Decomposition of Complex Utterances. Experiments show that the proposed approach enables the interpretation of complex utterances with almost no complex training data, while outperforming standard few-shot prompting approaches.