We present a novel approach to dialogue state tracking and referring expression resolution tasks. Successful contextual understanding of multi-turn spoken dialogues requires resolving referring expressions across turns and tracking the entities relevant to the conversation across turns. Tracking conversational state is particularly challenging in a multi-domain scenario when there exist multiple spoken language understanding (SLU) sub-systems, and each SLU sub-system operates on its domain-specific meaning representation. While previous approaches have addressed the disparate schema issue by learning candidate transformations of the meaning representation, in this paper, we instead model the reference resolution as a dialogue context-aware user query reformulation task – the dialog state is serialized to a sequence of natural language tokens representing the conversation. We develop our model for query reformulation using a pointer-generator network and a novel multi-task learning setup. In our experiments, we show a significant improvement in absolute F1 on an internal as well as a, soon to be released, public benchmark respectively.

In a spoken dialogue system, dialogue state tracker (DST) components track the state of the conversation by updating a distribution of values associated with each of the slots being tracked for the current user turn, using the interactions until then. Much of the previous work has relied on modeling the natural order of the conversation, using distance based offsets as an approximation of time. In this work, we hypothesize that leveraging the wall-clock temporal difference between turns is crucial for finer-grained control of dialogue scenarios. We develop a novel approach that applies a time mask, based on the wall-clock time difference, to the associated slot embeddings and empirically demonstrate that our proposed approach outperforms existing approaches that leverage distance offsets, on both an internal benchmark dataset as well as DSTC2.

Ce papier présente une méthode de recherche des phrases évaluatives dans les articles de presse économique et financière à partir de marques et d’indices stéréotypés, propres au style journalistique, apparaissant de manière concomitante à l’expression d’évaluation(s) dans les phrases. Ces marques et indices ont été dégagés par le biais d’une annotation manuelle. Ils ont ensuite été implémentés, en vue d’une phase-test d’annotation automatique, sous forme de grammaires DCG/GULP permettant, par filtrage, de matcher les phrases les contenant. Les résultats de notre première tentative d’annotation automatique sont présentés dans cet article. Enfin les perspectives offertes par cette méthode relativement peu coûteuse en ressources (à base d’indices non intrinsèquement évaluatifs) font l’objet d’une discussion.