The need for high-quality, large-scale, goaloriented dialogue datasets continues to grow as virtual assistants become increasingly widespread. However, publicly available datasets useful for this area are limited either in their size, linguistic diversity, domain coverage, or annotation granularity. In this paper, we present strategies toward curating and annotating large scale goal oriented dialogue data. We introduce the MultiDoGO dataset to overcome these limitations. With a total of over 81K dialogues harvested across six domains, MultiDoGO is over 8 times the size of MultiWOZ, the other largest comparable dialogue dataset currently available to the public. Over 54K of these harvested conversations are annotated for intent classes and slot labels. We adopt a Wizard-of-Oz approach wherein a crowd-sourced worker (the "customer") is paired with a trained annotator (the "agent"). The data curation process was controlled via biases to ensure a diversity in dialogue flows following variable dialogue policies. We provide distinct class label tags for agents vs. customer utterances, along with applicable slot labels. We also compare and contrast our strategies on annotation granularity, i.e. turn vs. sentence level. Furthermore, we compare and contrast annotations curated by leveraging professional annotators vs the crowd. We believe our strategies for eliciting and annotating such a dialogue dataset scales across modalities and domains and potentially languages in the future. To demonstrate the efficacy of our devised strategies we establish neural baselines for classification on the agent and customer utterances as well as slot labeling for each domain.
a b s t r a c tThe games considered are mixtures of Searching and Cops and Robber. The cops have partial information provided via witnesses who report ''sightings'' of the robber. The witnesses are able to provide information about the robber's position but not the direction in which he is moving. The robber has perfect information. In the case when sightings occur at regular intervals, we present a recognition theorem for graphs on which a single cop suffices to guarantee a win. In a special case, this recognition theorem provides a characterization.
We introduce a new variant of the game of Cops and Robbers played on graphs, where the robber is invisible unless outside the neighbor set of a cop. The hyperopic cop number is the corresponding analogue of the cop number, and we investigate bounds and other properties of this parameter. We characterize the cop-win graphs for this variant, along with graphs with the largest possible hyperopic cop number. We analyze the cases of graphs with diameter 2 or at least 3, focusing on when the hyperopic cop number is at most one greater than the cop number. We show that for planar graphs, as with the usual cop number, the hyperopic cop number is at most 3. The hyperopic cop number is considered for countable graphs, and it is shown that for connected chains of graphs, the hyperopic cop density can be any real number in [0, 1/2].
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.