Summary of Llm+reasoning+planning For Supporting Incomplete User Queries in Presence Of Apis, by Sudhir Agarwal (intuit Inc.) et al.
LLM+Reasoning+Planning for Supporting Incomplete User Queries in Presence of APIs
by Sudhir Agarwal, Anu Sreepathy, David H. Alonso, Prarit Lamba
First submitted to arxiv on: 21 May 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 proposes an innovative approach to natural language interfaces for various end-user tasks, leveraging logical reasoning and classical AI planning along with Large Language Models (LLMs). The authors aim to address the limitations of LLMs in handling incomplete user queries by accurately identifying and gathering missing information. To achieve this, they develop a framework that translates user queries into Planning Domain Definition Language (PDDL) via an intermediate representation in Answer Set Programming (ASP). This allows for the orchestration of API calls, including those required to gather missing information, using a classical AI planner. The authors demonstrate the effectiveness of their approach, achieving over 95% success rate on a dataset containing complete and incomplete single-goal and multi-goal queries. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us talk to computers in natural language by combining two powerful tools: Large Language Models (LLMs) and classical AI planning. Right now, LLMs can answer some questions, but they often struggle with missing information or don’t know which computer programs (APIs) to use. The authors created a new way to translate user queries into a special format that computers understand, using Answer Set Programming (ASP). This lets them figure out the best order of API calls needed to get an answer, even if some information is missing. By combining LLMs and AI planning, they can handle most questions correctly over 95% of the time. |
