This paper proposes a new dataset, MuDoCo, composed of authored dialogs between a fictional user and a system who are given tasks to perform within six task domains. These dialogs are given rich linguistic annotations by expert linguists for several types of reference mentions and named entity mentions, either of which can span multiple words, as well as for coreference links between mentions. The dialogs sometimes cross and blend domains, and the users exhibit complex task switching behavior such as re-initiating a previous task in the dialog by referencing the entities within it. The dataset contains a total of 8,429 dialogs with an average of 5.36 turns per dialog. We are releasing this dataset to encourage research in the field of coreference resolution, referring expression generation and identification within realistic, deep dialogs involving multiple domains. To demonstrate its utility, we also propose two baseline models for the downstream tasks: coreference resolution and referring expression generation.

Generating fluent natural language responses from structured semantic representations is a critical step in task-oriented conversational systems. Sequence-to-sequence models on flat meaning representations (MR) have been dominant in this task, for example in the E2E NLG Challenge. Previous work has shown that a tree-structured MR can improve the model for better discourse-level structuring and sentence-level planning. In this work, we propose a tree-to-sequence model that uses a tree-LSTM encoder to leverage the tree structures in the input MR, and further enhance the decoding by a structure-enhanced attention mechanism. In addition, we explore combining these enhancements with constrained decoding to improve semantic correctness. Our experiments not only show significant improvements over standard seq2seq baselines, but also is more data-efficient and generalizes better to hard scenarios.

We describe our exploratory system for the shallow surface realization task, which combines morphological inflection using character sequence-to-sequence models with a baseline linearizer that implements a tree-to-tree model using sequence-to-sequence models on serialized trees. Results for morphological inflection were competitive across languages. Due to time constraints, we could only submit complete results (including linearization) for English. Preliminary linearization results were decent, with a small benefit from reranking to prefer valid output trees, but inadequate control over the words in the output led to poor quality on longer sentences.

Generating fluent natural language responses from structured semantic representations is a critical step in task-oriented conversational systems. Avenues like the E2E NLG Challenge have encouraged the development of neural approaches, particularly sequence-to-sequence (Seq2Seq) models for this problem. The semantic representations used, however, are often underspecified, which places a higher burden on the generation model for sentence planning, and also limits the extent to which generated responses can be controlled in a live system. In this paper, we (1) propose using tree-structured semantic representations, like those used in traditional rule-based NLG systems, for better discourse-level structuring and sentence-level planning; (2) introduce a challenging dataset using this representation for the weather domain; (3) introduce a constrained decoding approach for Seq2Seq models that leverages this representation to improve semantic correctness; and (4) demonstrate promising results on our dataset and the E2E dataset.

Neural approaches to Natural Language Generation (NLG) have been promising for goal-oriented dialogue. One of the challenges of productionizing these approaches, however, is the ability to control response quality, and ensure that generated responses are acceptable. We propose the use of a generate, filter, and rank framework, in which candidate responses are first filtered to eliminate unacceptable responses, and then ranked to select the best response. While acceptability includes grammatical correctness and semantic correctness, we focus only on grammaticality classification in this paper, and show that existing datasets for grammatical error correction don’t correctly capture the distribution of errors that data-driven generators are likely to make. We release a grammatical classification and semantic correctness classification dataset for the weather domain that consists of responses generated by 3 data-driven NLG systems. We then explore two supervised learning approaches (CNNs and GBDTs) for classifying grammaticality. Our experiments show that grammaticality classification is very sensitive to the distribution of errors in the data, and that these distributions vary significantly with both the source of the response as well as the domain. We show that it’s possible to achieve high precision with reasonable recall on our dataset.