Statistical Framework with Knowledge Base Integration for Robust Speech Understanding of the Tunisian Dialect

Graja, Marwa; Jaoua, Maher; Belguith, Lamia Hadrich

doi:10.1109/taslp.2015.2464687

Cited by 16 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(4) BiLSTM Soft ATT: A standard BiLSTM with soft attention mechanism, which has been successfully applied in utterance classification task [9,10]. (5) KChat: a base model of knowledge-based chat detection, which follows previous works [12][13][14][15] to concatenate the knowledge feature with the contextual representation in the attentive BiLSTM. (6) KChat Gate: a base model of knowledge-gated chat detection, which has been proposed to control the information of external knowledge to flow into the UDC model [15].…”

Section: Experiments Settingsmentioning

confidence: 99%

“…Several studies have been proposed to integrate knowledge information to learn a knowledge-based sentence representation on text classification [12], speech understanding [13], and utterance domain classification (UDC) [14,15]. In this paper, we follow this line of research and propose a base model of knowledge-based chat detection (KChat).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Knowledge-Based Chat Detection with False Mention Discrimination

Liu

Huang

Zhou

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Chat detection is critical for recently emerged personal intelligent assistants (PIA), which can be seen as a hybrid of domain-specific task-oriented spoken dialogue systems and open-domain non-task-oriented ones. Recent advances have attempted to utilize external domain knowledge to enhance utterance semantics understanding and can contribute to chat detection. However, it also inevitably introduces false mention (i.e., token spans being misidentified as entity mentions) in Chat utterances, causing performance to degrade. To deal with this issue, this paper proposes a new model for knowledge-based chat detection with false mention discrimination (FMD-KChat). A two-stage pipeline is adopted, which contains an additional neural network-based classifier in the first stage for distinguishing the false mentions and a feature fusion gate in the chat detection stage for combining the contextual representation with the external knowledge feature based on the false mention discrimination probability. Experiments on the SMP-ECDT benchmark corpus show the well performance of the proposed model.

show abstract

Section: Experiments Settingsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Knowledge-Based Chat Detection with False Mention Discrimination

Liu

Huang

Zhou

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…Other researchers worked on domain ontologies for the TD processing in dialogue systems (Graja et al, 2011a(Graja et al, , 2011bKaroui et al, 2013aKaroui et al, , 2013bGraja et al, 2015). Graja et al (2011aGraja et al ( , 2011b) used ontologies to cover the lexicon used in train stations.…”

Section: Related Workmentioning

confidence: 99%

Contributions to the automatic processing of the user-generated Tunisian dialect on the social web

et al. 2020

View full text Add to dashboard Cite

With the growing use of social media in the Arab world, Arabic dialects are rapidly spreading on the web, leading to a growing interest from NLP researchers. These dialects are however, still under-resourced languages which is a major obstacle to their study and processing. In this paper, we focus on the automatic processing of the user-generated Tunisian dialect (TD) on the social web and propose an approach that aids to automatically generate TD language resources. This approach exploits the large amounts of textual productions on the social web to extract and generate dialectal content. It is based on two main NLP components, namely the TD identification and the TD transliteration. A machine learning approach using conditional random fields is proposed for implementing these two components and reached an accuracy of 87.45 for the TD identification and 90.49 for the automatic generation of dialectal contents by transliteration.

show abstract

“…Like other Arabic dialects, the Tunisian dialect respects in most cases the regular grammar of MSA or that of Classical Arabic [20]. Yet it has some syntactic particularities.…”

Section: Syntactic Levelmentioning

confidence: 99%