Controllable and Reliable Knowledge-Intensive Task-Oriented Conversational Agents with Declarative Genie Worksheets

TL;DR

The paper introduces Genie, a programmable framework for creating reliable, knowledge-intensive task-oriented conversational agents that outperform state-of-the-art methods in complex real-world tasks by ensuring grounded responses and controllable policies.

Contribution

Genie provides a novel declarative specification and algorithmic runtime system that enhances the reliability and controllability of LLM-based conversational agents for complex tasks.

Findings

Genie agents outperform SOTA methods on complex logic dialogue datasets.

Genie agents with GPT-4 Turbo achieve goal completion rates of 82.8%.

User study shows significant improvement in real-world task success.

Abstract

Large Language Models can carry out human-like conversations in diverse settings, responding to user requests for tasks and knowledge. However, existing conversational agents implemented with LLMs often struggle with hallucination, following instructions with conditional logic, and integrating knowledge from different sources. These shortcomings compromise the agents' effectiveness, rendering them unsuitable for deployment. To address these challenges, we introduce Genie, a programmable framework for creating knowledge-intensive task-oriented conversational agents. Genie can handle involved interactions and answer complex queries. Unlike LLMs, it delivers reliable, grounded responses through advanced dialogue state management and supports controllable agent policies via its declarative specification -- Genie Worksheet. This is achieved through an algorithmic runtime system that implements the developer-supplied policy, limiting LLMs to (1) parse user input using a succinct conversational history, and (2) generate responses according to supplied context. Agents built with Genie outperform SOTA methods on complex logic dialogue datasets. We conducted a user study with 62 participants on three real-life applications: restaurant reservations with Yelp, as well as ticket submission and course enrollment for university students. Genie agents with GPT-4 Turbo outperformed the GPT-4 Turbo agents with function calling, improving goal completion rates from 21.8% to 82.8% across three real-world tasks.