Measuring Semantic Coherence of a Conversation

Vakulenko, Svitlana; Rijke, Maarten de; Cochez, Michael; Savenkov, Vadim; Polleres, Axel

doi:10.1007/978-3-030-00671-6_37

Cited by 23 publications

(12 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Following the account presented in this thesis, the clear coherence relation between the predicates in (37b) (namely, extensional overlap, with the overlapping feature being location-related information) greatly improves the copredication. 98 Lastly, Vakulenko et al (2018) note that semantic coherence is always relative to the background knowledge of participants. As such, it is possible that the present acceptability experiments constitute an unusual form of coherence relation processing, given that the experiment involved single sentences.…”

Section: Incremental Semantic Complexity: a General Cognitive Principle?mentioning

confidence: 99%

Linguistic Representation and Processing of Copredication

Murphy¹

2021

Preprint

View full text Add to dashboard Cite

This thesis addresses the lexical and psycholinguistic properties of copredication. In particular, it explores its acceptability, frequency, cross-linguistic and electrophysiological features. It proposes a general parsing bias to account for novel acceptability data, through which Complex-Simple predicate orderings are degraded across distinct nominal types relative to the reverse order. This bias, Incremental Semantic Complexity, states that the parser seeks to process linguistic representations in incremental stages of semantic complexity. English and Italian acceptability data are presented which demonstrate that predicate order preferences are based not on sense dominance but rather sense complexity. Initial evidence is presented indicating that pragmatic factors centred on coherence relations can impact copredication acceptability when such copredications host complex (but not simple) predicates. The real-time processing and electrophysiological properties of copredication are also presented, which serve to replicate and ground the acceptability dynamics presented in the thesis.

show abstract

Section: Incremental Semantic Complexity: a General Cognitive Principle?mentioning

confidence: 99%

Linguistic Representation and Processing of Copredication

Murphy¹

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Introducing automatically generated labels may propagate errors into the model extraction phase. Nevertheless, discovering patterns in raw conversational data that is automatically tagged with semantic labels is an exciting research direction [25]. In addition, the predictions of the QRFA model may be an informative signal for evaluating or training reinforcement learning-based dialogue systems [14].…”

Section: Resultsmentioning

confidence: 99%

QRFA: A Data-Driven Model of Information-Seeking Dialogues

Vakulenko

Revoredo

Ciccio

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Understanding the structure of interaction processes helps us to improve information-seeking dialogue systems. Analyzing an interaction process boils down to discovering patterns in sequences of alternating utterances exchanged between a user and an agent. Process mining techniques have been successfully applied to analyze structured event logs, discovering the underlying process models or evaluating whether the observed behavior is in conformance with the known process. In this paper, we apply process mining techniques to discover patterns in conversational transcripts and extract a new model of information-seeking dialogues, QRFA, for Query, Request, Feedback, Answer. Our results are grounded in an empirical evaluation across multiple conversational datasets from different domains, which was never attempted before. We show that the QRFA model better reflects conversation flows observed in real information-seeking conversations than models proposed previously. Moreover, QRFA allows us to identify malfunctioning in dialogue system transcripts as deviations from the expected conversation flow described by the model via conformance analysis.

show abstract

“…As to future work, we plan to explore more fashion items in our dataset, e.g., hats, glasses and shoes, etc. Also, to alleviate the problem of generating meaningless comments, studies into coherence in information retrieval [67] or dialogue systems can be explored [68,69]. And we would like to incorporate other mechanisms, such as an auto-encoder, to further improve the performance.…”

Section: Discussionmentioning

confidence: 99%

Explainable Outfit Recommendation with Joint Outfit Matching and Comment Generation

Lin

Ren

Chen

et al. 2020

IEEE Trans. Knowl. Data Eng.

Self Cite

View full text Add to dashboard Cite

Most previous work on outfit recommendation focuses on designing visual features to enhance recommendations. Existing work neglects user comments of fashion items, which have been proved to be effective in generating explanations along with better recommendation results. We propose a novel neural network framework, neural outfit recommendation (NOR), that simultaneously provides outfit recommendations and generates abstractive comments. NOR consists of two parts: outfit matching and comment generation. For outfit matching, we propose a convolutional neural network with a mutual attention mechanism to extract visual features. The visual features are then decoded into a rating score for the matching prediction. For abstractive comment generation, we propose a gated recurrent neural network with a cross-modality attention mechanism to transform visual features into a concise sentence. The two parts are jointly trained based on a multi-task learning framework in an end-to-end back-propagation paradigm. Extensive experiments conducted on an existing dataset and a collected real-world dataset show NOR achieves significant improvements over state-of-the-art baselines for outfit recommendation. Meanwhile, our generated comments achieve impressive ROUGE and BLEU scores in comparison to human-written comments. The generated comments can be regarded as explanations for the recommendation results. We release the dataset and code to facilitate future research.

show abstract

Measuring Semantic Coherence of a Conversation

Cited by 23 publications

References 25 publications

Linguistic Representation and Processing of Copredication

Linguistic Representation and Processing of Copredication

QRFA: A Data-Driven Model of Information-Seeking Dialogues

Explainable Outfit Recommendation with Joint Outfit Matching and Comment Generation

Contact Info

Product

Resources

About