GODEL: Large-Scale Pre-Training for Goal-Directed Dialog

TL;DR

GODEL is a large pre-trained dialog model that excels in various tasks by incorporating grounded pre-training, improving response usefulness and adaptability over previous models.

Contribution

The paper introduces GODEL, a novel grounded pre-training approach for dialog models, enhancing performance across multiple dialog tasks and introducing a new extrinsic evaluation methodology.

Findings

GODEL outperforms state-of-the-art dialog models in few-shot settings.

Grounded pre-training improves response usefulness and task adaptability.

Extrinsic evaluation correlates better with human judgments.

Abstract

We introduce GODEL (Grounded Open Dialogue Language Model), a large pre-trained language model for dialog. In contrast with earlier models such as DialoGPT, GODEL leverages a new phase of grounded pre-training designed to better support adapting GODEL to a wide range of downstream dialog tasks that require information external to the current conversation (e.g., a database or document) to produce good responses. Experiments against an array of benchmarks that encompass task-oriented dialog, conversational QA, and grounded open-domain dialog show that GODEL outperforms state-of-the-art pre-trained dialog models in few-shot fine-tuning setups, in terms of both human and automatic evaluation. A novel feature of our evaluation methodology is the introduction of a notion of utility that assesses the usefulness of responses (extrinsic evaluation) in addition to their communicative features (intrinsic evaluation). We show that extrinsic evaluation offers improved inter-annotator agreement and correlation with automated metrics. Code and data processing scripts are publicly available.