1,721,062 research outputs found
Social educational robotics and learning analytics: a scoping review of an Emerging field
No full textSocial Educational Robotics and Learning Analytics (LA) are prominent fields in technology-enhanced learning, but their combined potential remains underexplored, despite methodological similarities. Increasingly, signs of joint interests have emerged, with a surge in publications mentioning both social robots and learning analytics in the last five years. We therefore conducted a scoping review to explore if a new research field is emerging. We identified 29 empirical studies that combine social robots and LA, but also found that few studies explicitly state that social educational robots and LA are used in combination. Several studies used social educational robots that adapted to the learners or the learning environment based on interaction data. This signifies that they are in fact employing the feedback cycle that is at the core of LA methodology, but as most of these studies update the learner model using post-session data (e.g., learner improvement or feedback), they are long-term studies with repeated interventions that are applying LA methodology inadvertently. There are also benefits for LA research to use social educational robots, since LA increasingly uses an array of equipment to collect multimodal data, and all studies in this review employ at least two input modalities (μ = 4.4). Social robots provide the possibility to collect this data non-intrusively with the robot itself, in addition to creating a pedagogically boosted interaction compared to traditional LA interventions (e.g., learning management systems). By raising researchers’ awareness of how close the fields of social educational robotics and LA are, substantial synergy effects could therefore be gained
The University of Birmingham 2017 SLaTE CALL Shared Task Systems
Get PDFThis paper describes the system developed by the University of Birmingham for the SLaTE CALL Shared Task on grammatical and linguistic assessment of English spoken by German-speaking Swiss teenagers. Our work focused on automaticspeech recognition (ASR) but we also improved the text-processing component of the system. Several approaches to training a DNN-HMM ASR system using the AMI and the German PF-STAR corpus, plus a limited amount of Shared Task data, are described. In cross-validation evaluations on the initial Shared Task data, our final ASR system achieved a word error-rate (WER) of 9.27%, compared with 14% for the official baseline Shared Task DNN-HMM system. For text processing we expanded the baseline template-based grammar to include additional correct response patterns from the original Shared Task transcriptions. Finally, we fused the outputs of several systems at the text processing stage using linear logistic regression. Our best single and fused systems submitted to the challengeachieved ‘D’ scores of 4.71 and 4.766, respectively, on the final test se
The University of Birmingham 2017 SLaTE CALL Shared Task Systems
No full textThis paper describes the system developed by the University of Birmingham for the SLaTE CALL Shared Task on grammatical and linguistic assessment of English spoken by German-speaking Swiss teenagers. Our work focused on automaticspeech recognition (ASR) but we also improved the text-processing component of the system. Several approaches to training a DNN-HMM ASR system using the AMI and the German PF-STAR corpus, plus a limited amount of Shared Task data, are described. In cross-validation evaluations on the initial Shared Task data, our final ASR system achieved a word error-rate (WER) of 9.27%, compared with 14% for the official baseline Shared Task DNN-HMM system. For text processing we expanded the baseline template-based grammar to include additional correct response patterns from the original Shared Task transcriptions. Finally, we fused the outputs of several systems at the text processing stage using linear logistic regression. Our best single and fused systems submitted to the challengeachieved ‘D’ scores of 4.71 and 4.766, respectively, on the final test se
Analysis of and feedback on phonetic features in pronunciation training with a virtual teacher
No full textPronunciation errors may be caused by several different deviations from the target, such as voicing, intonation, insertions or deletions of segments, or that the articulators are placed incorrectly. Computer-animated pronunciation teachers could potentially provide important assistance on correcting all these types of deviations, but they have an additional benefit for articulatory errors. By making parts of the face transparent, they can show the correct position and shape of the tongue and provide audiovisual feedback on how to change erroneous articulations. Such a scenario however requires firstly that the learner's current articulation can be estimated with precision and secondly that the learner is able to imitate the articulatory changes suggested in the audiovisual feedback. This article discusses both these aspects, with one experiment on estimating the important articulatory features from a speaker through acoustic-to-articulatory inversion and one user test with a virtual pronunciation teacher, in which the articulatory changes made by seven learners who receive audiovisual feedback are monitored using ultrasound imaging.QC 20120416</p
Going Beyond Counting First Authors in Author Co-citation Analysis
Get PDFThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Articulatory synthesis using corpus-based estimation of line spectrum pairs
No full textAn attempt to define a new articulatory synthesis method, in which the speech signal is generated through a statistical estimation of its relation with articulatory parameters, is presented. A corpus containing acoustic material and simultaneous recordings of the tongue and facial movements was used to train and test the articulatory synthesis of VCV words and short sentences. Tongue and facial motion data, captured with electromagnetic articulography and three-dimensional optical motion tracking, respectively, define articulatory parameters of a talking head. These articulatory parameters are then used as estimators of the speech signal, represented by line spectrum pairs. The statistical link between the articulatory parameters and the speech signal was established using either linear estimation or artificial neural networks. The results show that the linear estimation was only enough to synthesize identifiable vowels, but not consonants, whereas the neural networks gave a perceptually better synthesis.</p
Speaker adaptation of a three-dimensional tongue model
No full textMagnetic Resonance Images of nine subjects have been collected to determine scaling factors that can adapt a 3D tongue model to new subjects. The aim is to define few and simple measures that will allow for an automatic, but accurate, scaling of the model. The scaling should be automatic in order to be useful in an application for articulation training, in which the model must replicate the user's articulators without involving the user in a complicated speaker adaptation. It should further be accurate enough to allow for correct acoustic-to-articulatory inversion. The evaluation shows that the defined scaling technique is able to estimate a tongue shape that was not included in the training with an accuracy of 1.5 mm in the midsagittal plane and 1.7 mm for the whole 3D tongue, based on four articulatory measures.</p
- …
