An Integrated Dialogue Analysis Model for Determining Speech Acts and Discourse Structures

Choi, Won Seug; Kim, Harksoo; Seo, Jungyun

doi:10.1093/ietisy/e88-d.1.150

Cited by 9 publications

(9 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lee [3] designed a speech-act classification method using the hidden Markov model (HMM) to estimate speech-act probabilities and this method was improved upon by using smoothed class probabilities from a decision tree. Choi [4] proposed a maximum entropy model (MEM) to determine the speech-act of current utterances using previous utterances as contextual information. Song [5] recommended a support vector machine (SVM) model to preferentially analyze classes with lower distribution when training among a set of classes.…”

Section: Related Workmentioning

confidence: 99%

Speech-Act Classification Using a Convolutional Neural Network Based on POS Tag and Dependency-Relation Bigram Embedding

Yoo

Seo

2017

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYIn this paper, we propose a deep learning based model for classifying speech-acts using a convolutional neural network (CNN). The model uses some bigram features including parts-of-speech (POS) tags and dependency-relation bigrams, which represent syntactic structural information in utterances. Previous classification approaches using CNN have commonly exploited word embeddings using morpheme unigrams. However, the proposed model first extracts two different bigram features that well reflect the syntactic structure of utterances and then represents them as a vector representation using a word embedding technique. As a result, the proposed model using bigram embeddings achieves an accuracy of 89.05%. Furthermore, the accuracy of this model is relatively 2.8% higher than that of competitive models in previous studies.

show abstract

Section: Related Workmentioning

confidence: 99%

Speech-Act Classification Using a Convolutional Neural Network Based on POS Tag and Dependency-Relation Bigram Embedding

Yoo

Seo

2017

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

show abstract

“…Recently, statistical speech act classification using a tagged dialogue corpus has be en proposed in order to solve such problems (Kim et al, 2004;Lee and Seo, 2002;Choi et al, 2005). Most previous works on speech act classification have used two feature types: sentential features and contextual features.…”

Section: Related Workmentioning

confidence: 99%

Hierarchical speech-act classification for discourse analysis

Kang

Seo

2013

Pattern Recognition Letters

View full text Add to dashboard Cite

The analysis of a speech act is important for dialogue understanding systems because the speech act of an utterance is closely associated with the user's intention in the utterance. This paper proposes a speech act classification model that effectively uses a two-layer hierarchical structure generated from the adjacency pair information of speech acts. The proposed model has two advantages when adding hierarchical information to speech act classification; the improved accuracy of the speech act classification and the reduced running time in the testing phase. As a result, it achieves higher performance than other models that do not use the hierarchical structure and has faster running time because Support Vector Machine classifiers can efficiently be arranged on the two-layer hierarchical structure.

show abstract

“…The second component is a mandatory element of the Interchange Format representation, and sometimes an utterance may be assigned a speech act with no concepts or arguments (e.g. utterance (2) in Table 1). The third component, a concept sequence, is a set of domain-dependent concepts and may contain zero or more concepts.…”

Section: Introductionmentioning

confidence: 99%

“…In the Interchange Format, while Table 1 An example of utterances along with their corresponding Interchange Format representations (1) Hello. User: greeting (greeting = hello) (2) May I help you? System: opening (3) Tell me the tomorrow schedule.…”

Section: Introductionmentioning

confidence: 99%

Domain action classification using a maximum entropy model in a schedule management domain

2008

View full text Add to dashboard Cite

To generate correct reactions, a dialogue system should identify domain actions indicated by users' utterances because speaker intentions can be captured by the domain actions. In this paper, we propose a domain action classification model to determine speech acts (general intentions) and concept sequences (semantic focuses) at the same time in a schedule management domain. To avoid biased learning problems, the proposed model uses low-level linguistic features and filters out uninformative features using statistic. Then, the proposed model determines domain actions using a maximum entropy model. In the experiment, the proposed model showed better performances than previous works in speech act classification. In addition, the proposed model showed high performances in concept sequence classification. Based on these experimental results, we believe that the proposed model will be more helpful to a dialogue system than previous speech act classification models because it can return speech acts and concept sequences at the same time on the same framework.

show abstract

An Integrated Dialogue Analysis Model for Determining Speech Acts and Discourse Structures

Cited by 9 publications

References 7 publications

Speech-Act Classification Using a Convolutional Neural Network Based on POS Tag and Dependency-Relation Bigram Embedding

Speech-Act Classification Using a Convolutional Neural Network Based on POS Tag and Dependency-Relation Bigram Embedding

Hierarchical speech-act classification for discourse analysis

Domain action classification using a maximum entropy model in a schedule management domain

Contact Info

Product

Resources

About