We propose a multi task learning-based neural model for resolving bridging references tackling two key challenges. The first challenge is the lack of large corpora annotated with bridging references. To address this, we use multi-task learning to help bridging reference resolution with coreference resolution. We show that substantial improvements of up to 8 p.p. can be achieved on full bridging resolution with this architecture. The second challenge is the different definitions of bridging used in different corpora, meaning that hand-coded systems or systems using special features designed for one corpus do not work well with other corpora. Our neural model only uses a small number of corpus independent features, thus can be applied to different corpora. Evaluations with very different bridging corpora (ARRAU, ISNOTES, BASHI and SCICORP) suggest that our architecture works equally well on all corpora, and achieves the SoTA results on full bridging resolution for all corpora, outperforming the best reported results by up to 36.3 p.p..
Disambiguating Reference in Visually Grounded Dialogues through Joint Modeling of Textual and Multimodal Semantic Structures
Nara Institute of Science and Technology（奈良科学技術研究所）；Guardian Robot Project（守護機器人計劃）；RIKEN（理化学研究所）；Kyoto University（京都大學）；Institute of Science Tokyo（東京科學研究院）
Most previous studies on bridging anaphora resolution (Poesio et al., 2004; Hou et al., 2013b; Hou, 2018a) use the pairwise model to tackle the problem and assume that the gold mention information is given. In this paper, we cast bridging anaphora resolution as question answering based on context. This allows us to find the antecedent for a given anaphor without knowing any gold mention information (except the anaphor itself). We present a question answering framework (BARQA) for this task, which leverages the power of transfer learning. Furthermore, we propose a novel method to generate a lar…
IBM Research Ireland（IBM爱尔兰研究公司（IBM Research Ireland））
Most current models of word representations(e.g.,GloVe) have successfully captured fine-grained semantics. However, semantic similarity exhibited in these word embeddings is not suitable for resolving bridging anaphora, which requires the knowledge of associative similarity (i.e., relatedness) instead of semantic similarity information between synonyms or hypernyms. We create word embeddings (embeddings_PP) to capture such relatedness by exploring the syntactic structure of noun phrases. We demonstrate that using embeddings_PP alone achieves around 30% of accuracy for bridging anaphora resolut…
Previous work on bridging anaphora resolution (Poesio et al., 2004; Hou et al., 2013b) use syntactic preposition patterns to calculate word relatedness. However, such patterns only consider NPs' head nouns and hence do not fully capture the semantics of NPs. Recently, Hou (2018) created word embeddings (embeddings_PP) to capture associative similarity (ie, relatedness) between nouns by exploring the syntactic structure of noun phrases. But embeddings_PP only contains word representations for nouns. In this paper, we create new word vectors by combining embeddings_PP with GloVe. This new word e…
Reference resolution on extended texts (several thousand references) cannot be evaluated manually. An evaluation algorithm has been proposed for the MUC tests, using equivalence classes for the coreference relation. However, we show here that this algorithm is too indulgent, yielding good scores even for poor resolution strategies. We elaborate on the same formalism to propose two new evaluation algorithms, comparing them first with the MUC algorithm and giving then results on a variety of examples. A third algorithm using only distributional comparison of equivalence classes is finally descri…
J-CRe3: A Japanese Conversation Dataset for Real-world Reference Resolution
Kyoto University（京都大学）；RIKEN（理化学研究所）；Tokyo University of Science（东京理科大学）；Nara Institute of Science and Technology（奈良科学技术研究所）；National Institute of Informatics（情报通信研究机构）
University of Illinois Chicago（伊利诺伊大学芝加哥分校）；University of Cincinnati（辛辛那提大学）
Reference resolution, which aims to identify entities being referred to by a speaker, is more complex in real world settings: new referents may be created by processes the agents engage in and/or be salient only because they belong to the shared physical setting. Our focus is on resolving references to visualizations on a large screen display in multimodal dialogue; crucially, reference resolution is directly involved in the process of creating new visualizations. We describe our annotations for user references to visualizations appearing on a large screen via language and hand gesture and als…
Anaphora resolution is envisaged in this paper as part of the reference resolution process. A general open architecture is proposed, which can be particularized and configured in order to simulate some classic anaphora resolution methods. With the aim of improving pronoun resolution, the system takes advantage of elementary cues about characters of the text, which are represented through a particular data structure. In its most robust configuration, the system uses only a general lexicon, a local morpho-syntactic parser and a dictionary of synonyms. A short comparative corpus analysis shows th…
A model for reference use in communication is proposed, from a representationist point of view. Both the sender and the receiver of a message handle representations of their common environment, including mental representations of objects. Reference resolution by a computer is viewed as the construction of object representations using referring expressions from the discourse, whereas often only coreference links between such expressions are looked for. Differences between these two approaches are discussed. The model has been implemented with elementary rules, and tested on complex narrative te…
This paper describes a standalone, publicly-available implementation of the Resolution of Anaphora Procedure (RAP) given by Lappin and Leass (1994). The RAP algorithm resolves third person pronouns, lexical anaphors, and identifies pleonastic pronouns. Our implementation, JavaRAP, fills a current need in anaphora resolution research by providing a reference implementation that can be benchmarked against current algorithms. The implementation uses the standard, publicly available Charniak (2000) parser as input, and generates a list of anaphora-antecedent pairs as output. Alternately, an in-pla…
Following the principles of Cognitive Grammar, we concentrate on a model for reference resolution that attempts to overcome the difficulties previous approaches, based on the fundamental assumption that all reference (independent on the type of the referring expression) is accomplished via access to and restructuring of domains of reference rather than by direct linkage to the entities themselves. The model accounts for entities not explicitly mentioned but understood in a discourse, and enables exploitation of discursive and perceptual context to limit the set of potential referents for a giv…
University of Warwick（华威大学）；Alan Turing Institute（艾伦·图灵研究所）；Filament AI（Filament AI公司）；Bar-Ilan University（巴伊兰大学）；Aberystwyth University（阿伯里斯特威斯大学）
Cross-document co-reference resolution (CDCR) is the task of identifying and linking mentions to entities and concepts across many text documents. Current state-of-the-art models for this task assume that all documents are of the same type (e.g. news articles) or fall under the same theme. However, it is also desirable to perform CDCR across different domains (type or theme). A particular use case we focus on in this paper is the resolution of entities mentioned across scientific work and newspaper articles that discuss them. Identifying the same entities and corresponding concepts in both sci…
DEVCOM Army Research Laboratory（DEVCOM陆军研究实验室）；U.S. Naval Research Laboratory（美国海军研究实验室）；Tufts University（塔夫茨大学）
Dialogue agents that interact with humans in situated environments need to manage referential ambiguity across multiple modalities and ask for help as needed. However, it is not clear what kinds of questions such agents should ask nor how the answers to such questions can be used to resolve ambiguity. To address this, we analyzed dialogue data from an interactive study in which participants controlled a virtual robot tasked with organizing a set of tools while engaging in dialogue with a live, remote experimenter. We discovered a number of novel results, including the distribution of question …
Reference resolution is one of the important tasks in natural language processing. In this paper, the author first determines the referents and their locations of "dousha", literally meaning "the same company", which appear in Japanese newspaper articles. Secondly, three heuristic methods, two of which use semantic information in text such as company names and their patterns, are proposed and tested on how accurately they identify the correct referents. The proposed methods based on semantic patterns show high accuracy for reference resolution of "dousha" (more than 90\%). This suggests that s…
This paper presents the results of an empirical investigation of temporal reference resolution in scheduling dialogs. The algorithm adopted is primarily a linear-recency based approach that does not include a model of global focus. A fully automatic system has been developed and evaluated on unseen test data with good results. This paper presents the results of an intercoder reliability study, a model of temporal reference resolution that supports linear recency and has very good coverage, the results of the system evaluated on unseen test data, and a detailed analysis of the dialogs assessing…
In this paper we examine the benefit of performing named entity recognition (NER) and co-reference resolution to an English and a Greek corpus used for text segmentation. The aim here is to examine whether the combination of text segmentation and information extraction can be beneficial for the identification of the various topics that appear in a document. NER was performed manually in the English corpus and was compared with the output produced by publicly available annotation tools while, an already existing tool was used for the Greek corpus. Produced annotations from both corpora were man…
Soochow University（苏州大学）；University of Edinburgh（爱丁堡大学）；Lenovo Research AI Lab（联想研究院人工智能实验室）
Ellipsis and co-reference are common and ubiquitous especially in multi-turn dialogues. In this paper, we treat the resolution of ellipsis and co-reference in dialogue as a problem of generating omitted or referred expressions from the dialogue context. We therefore propose a unified end-to-end Generative Ellipsis and CO-reference Resolution model (GECOR) in the context of dialogue. The model can generate a new pragmatically complete user utterance by alternating the generation and copy mode for each user utterance. A multi-task learning framework is further proposed to integrate the GECOR int…
I see what you mean: Co-Speech Gestures for Reference Resolution in Multimodal Dialogue
Multimodal Language Department, Max Planck Institute for Psycholinguistics（马克斯·普朗克心理语言学研究所多模态语言部门）；Donders Institute for Brain, Cognition and Behaviour, Radboud University（拉德堡德大学大脑、认知与行为研究所）；Department of Advanced Computing Sciences, Maastricht University（马斯特里赫特大学高级计算科学系）；Institute for Logic, Language and Computation, University of Amsterdam（阿姆斯特丹大学逻辑、语言与计算研究所）
