Toward Automating Oral Presentation Scoring During Principal Certification Program Using Audio-Video Low-Level Behavior Profiles

Hsiao, Shan-Wen; Sun, Hung-Ching; Hsieh, Ming-Chuan; Tsai, Ming-Hsueh; Tsao, Yu; Lee, Chi-Chun

doi:10.1109/taffc.2017.2749569

Cited by 6 publications

(10 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The next most common category is digital competence, which is generally defined as a set of knowledge, skills, and attitudes required when using new technologies to create, communicate, and resolve problems in an efficient and effective way, and also to improve the teaching process when using technology [63]. For example, in our findings, some of the studies are focused on measuring the digital competence [37,45,53]. Similarly to these findings, a review carried out by Wilson, Ritzhaupt, and Cheng [64] found that pre-and in-service teachers' digital competence should be considered a necessary skill in their teaching activities.…”

Section: Goals and Objectives Reference Number Number Of Studiesmentioning

confidence: 95%

See 1 more Smart Citation

Artificial Intelligence and Learning Analytics in Teacher Education: A Systematic Review

Salas‐Pilco

Xiao

2022

Education Sciences

View full text Add to dashboard Cite

In recent years, artificial intelligence (AI) and learning analytics (LA) have been introduced into the field of education, where their use has great potential to enhance the teaching and learning processes. Researchers have focused on applying these technologies to teacher education, as they see the value of technology for educating. Therefore, a systematic review of the literature on AI and LA in teacher education is necessary to understand their impact in the field. Our methodology follows the PRISMA guidelines, and 30 studies related to teacher education were identified. This review analyzes and discusses the several ways in which AI and LA are being integrated in teacher education based on the studies’ goals, participants, data sources, and the tools used to enhance teaching and learning activities. The findings indicate that (a) there is a focus on studying the behaviors, perceptions, and digital competence of pre- and in-service teachers regarding the use of AI and LA in their teaching practices; (b) the main data sources are behavioral data, discourse data, and statistical data; (c) machine learning algorithms are employed in most of the studies; and (d) the ethical clearance is mentioned by few studies. The implications will be valuable for teachers and educational authorities, informing their decisions regarding the effective use of AI and LA technologies to support teacher education.

show abstract

Section: Goals and Objectives Reference Number Number Of Studiesmentioning

confidence: 95%

“…Behavior when using AI and LA [31,35,39,40,42,44,52] 7 Digital competence [34,38,45,50,53,58] 6 Perception of AI and LA [32,41,48,56,60] 5 Self-regulation and reflection [33,35,47,54,55] 5 Engagement [36,51,57,59] 4 Analysis of educational data [37,43,49] 3…”

Section: Goals and Objectives Reference Number Number Of Studiesmentioning

confidence: 99%

Artificial Intelligence and Learning Analytics in Teacher Education: A Systematic Review

Salas‐Pilco

Xiao

2022

Education Sciences

View full text Add to dashboard Cite

show abstract

“…We further encode this sequence of LLDs using a Gaussian Mixture Model (GMM) based Fisher scoring [23] to derive a fixed high-dimensional acoustic at the turn-level. This particular method has been shown to be useful in speech-related recognition tasks of emotion [24] and para-linguistic attribute [25], also in the assessment of impromptu speech [26].…”

Section: Turn-level Acoustic Featuresmentioning

confidence: 99%

An Interlocutor-Modulated Attentional LSTM for Differentiating between Subgroups of Autism Spectrum Disorder

2018

Self Cite

View full text Add to dashboard Cite

Recalling and discussing personal emotional experiences is one of the key procedures in assessing complex affect processing of individuals with Autism Spectrum Disorder (ASD). This procedure is a standard subpart of a diagnostic interview to assess ASD-the Autism Diagnostic Observation Schedule (ADOS). Previous work has demonstrated that the behavior features computed from this procedure in ADOS possess discriminative information between the three distinct ASD subgroups: Autistic Disorder (AD), High Functioning Autism (HFA), and Asperger Syndrome (AS). In this work, we propose an interlocutor-modulated attentional long short term memory network (IM-aLSTM) that models the ASD individual's acoustic features with a novel interlocutor-modulated attention mechanism. Our IM-aLSTM achieves ASD subgroup categorization accuracy of 66.5%, which is a 14% absolute improvement over baseline method on the same database. Our analyses further indicate that the attention weights are concentrated more on interaction segments where the ASD individual is being asked to recall and discuss his/her own negative emotional experiences.

show abstract

“…We compute the gradient log-likelihood function, i.e., Fisher scoring (indicating the direction of λ to better fitx l ), with respect to the first and second order statistics of the learned latent VaDE-GMM parameters to further encode a sequence of acoustic latent representationx l into a fixed-length representation (also terms as GMMbased Fisher-vector encoding [26]). The use of Fisher-vector encoding has been shown to be competitive in speech-related tasks of paralinguistic recognition [27], presentation scoring [28], and emotion recognition [29,30]. The dialog-level acoustic vectors that integrates both the general representation and the dyad-specific dynamics is derived by concatenating the general Fisher-scoring vector with the dyad-specific Fisher-scoring vector.…”

Section: Dialog-level Emotion Recognitionmentioning

confidence: 99%

Encoding Individual Acoustic Features Using Dyad-Augmented Deep Variational Representations for Dialog-level Emotion Recognition

Lee

2018

Interspeech 2018

Self Cite

View full text Add to dashboard Cite

Face-to-face dyadic spoken dialog is a fundamental unit of human interaction. Despite numerous empirical evidences in demonstrating interlocutor's behavior dependency in dyadic interactions, few technical works exist in leveraging the unique pattern of dynamics in task of advancing emotion recognition during face-to-face settings. In this work, we propose a framework of encoding an individual's acoustic features with dyadaugmented deep networks. The dyad-augmented deep networks includes a general variational deep Gaussian Mixture embedding network and a dyad-specific fine-tuned network. Our framework utilizes the augmented dyad-specific feature space to incorporate the unique behavior pattern emerged when two people interact. We perform dialog-level emotion regression tasks in both the CreativeIT and the NNIME databases. We obtain affect regression accuracy of 0.544 and 0.387 for activation and valence in the CreativeIT database (a relative improvement of 4.41% and 4.03% compared to using features without augmenting the dyad-specific representation), and we obtain 0.700 and 0.604 (4.48% and 4.14% relative improvement) for regressing activation and valence in the NNIME database.

show abstract

Toward Automating Oral Presentation Scoring During Principal Certification Program Using Audio-Video Low-Level Behavior Profiles

Cited by 6 publications

References 63 publications

Artificial Intelligence and Learning Analytics in Teacher Education: A Systematic Review

Artificial Intelligence and Learning Analytics in Teacher Education: A Systematic Review

An Interlocutor-Modulated Attentional LSTM for Differentiating between Subgroups of Autism Spectrum Disorder

Encoding Individual Acoustic Features Using Dyad-Augmented Deep Variational Representations for Dialog-level Emotion Recognition

Contact Info

Product

Resources

About