Generative Conversational Networks

TL;DR

This paper introduces Generative Conversational Networks, a framework where conversational agents generate their own training data through reinforcement learning, improving performance on language tasks with limited data and computation.

Contribution

It presents a novel approach combining generative data creation and reinforcement learning for conversational agents, enhancing performance on various language tasks.

Findings

Achieves 35% improvement in intent detection

Achieves 21% improvement in slot tagging

Performs well with limited data and computation

Abstract

Inspired by recent work in meta-learning and generative teaching networks, we propose a framework called Generative Conversational Networks, in which conversational agents learn to generate their own labelled training data (given some seed data) and then train themselves from that data to perform a given task. We use reinforcement learning to optimize the data generation process where the reward signal is the agent's performance on the task. The task can be any language-related task, from intent detection to full task-oriented conversations. In this work, we show that our approach is able to generalise from seed data and performs well in limited data and limited computation settings, with significant gains for intent detection and slot tagging across multiple datasets: ATIS, TOD, SNIPS, and Restaurants8k. We show an average improvement of 35% in intent detection and 21% in slot tagging over a baseline model trained from the seed data. We also conduct an analysis of the novelty of the generated data and provide generated examples for intent detection, slot tagging, and non-goal oriented conversations.