CLMar 7, 2022

What Did You Say? Task-Oriented Dialog Datasets Are Not Conversational!?

arXiv:2203.03431v13 citationsh-index: 30
AI Analysis

This highlights a critical flaw in widely used datasets for virtual assistant development, potentially hindering progress in conversational AI.

The paper identified that major task-oriented dialog datasets (MultiWOZ, SGD, SMCalFlow) are largely non-conversational, as dialog state tracking can often be done using only the current user utterance, ignoring history, with less than 4% of MultiWOZ's turns and 10% of SGD's turns being conversational.

High-quality datasets for task-oriented dialog are crucial for the development of virtual assistants. Yet three of the most relevant large scale dialog datasets suffer from one common flaw: the dialog state update can be tracked, to a great extent, by a model that only considers the current user utterance, ignoring the dialog history. In this work, we outline a taxonomy of conversational and contextual effects, which we use to examine MultiWOZ, SGD and SMCalFlow, among the most recent and widely used task-oriented dialog datasets. We analyze the datasets in a model-independent fashion and corroborate these findings experimentally using a strong text-to-text baseline (T5). We find that less than 4% of MultiWOZ's turns and 10% of SGD's turns are conversational, while SMCalFlow is not conversational at all in its current release: its dialog state tracking task can be reduced to single exchange semantic parsing. We conclude by outlining desiderata for truly conversational dialog datasets.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes