Catálogo de publicaciones - libros

This paper presents LECTOR, a system that helps educators in understanding when students have stopped paying attention to the educational process and assists them in reengaging the students to the current learning activity. LECTOR aims to take advantage of the ambient facilities that “smart classrooms” have to offer by (i) enabling educators to employ their preferred attention monitoring strategies (including any well-established activity recognition techniques) in order to identify inattentive behaviors and (ii) recommending interventions for motivating distracted students when deemed necessary. Furthermore, LECTOR offers an educator friendly design studio that enables teachers to create or modify the rules that trigger “inattention alarms”, as well as tailor the intervention mechanism to the needs of their course by modifying the respective rules. This paper presents the rationale behind the design of LECTOR and outlines its key features and facilities.

Developing an early warning model based on mental workload (MWL) to predict the driver’s performance is critical and helpful, especially for new or less experienced drivers. This study aims to investigate the correlation between human’s MWL and work performance and develop the predictive model in the driving task using driving simulator. The performance measure (number of errors), subjective rating (NASA Task Load Index) as well as six physiological indices were assessed and measured. Additionally, the group method of data handling (GMDH) was used to establish the work performance model. The results indicate that different complexity levels of driving task have a significant effect on the driver’s performance, and the predictive performance model integrates different physiological measures shows the validity of the proposed model is well with R = 0.781. The proposed model is expected to provide a reference value of their work performance by giving physiological indices. Based on this model, the driving lesson plans will be proposed to sustain the appropriate MWL as well as improve work performance.

Conversational user interfaces (CUIs) are rapidly evolving towards being ubiquitous as human-machine interfaces. Often, CUI backends are powered by a combination of human and machine intelligence, to address queries efficiently. Depending on the type of conversation issue, human-to-human conversations in CUIs (i.e. a human end-user conversing with the human in the CUI backend) could involve varying amounts of emotional content. While some of these emotions could be expressed through the conversation, others are experienced internally within the individual. Understanding the relationship between these two emotion modalities in the end-user could help to analyze and address the conversation issue better. Towards this, we propose an emotion analytic metric that can estimate experienced emotions based on its knowledge about expressed emotions in a user. Our findings point to the possibility of augmenting CUIs with an algorithmically guided emotional sense, which would help in having more effective conversations with end-users.

Multiword expressions (MWEs) are units in language where multiple words unite without an obvious/known reason. Since MWEs occupy a prominent amount of space in both written and spoken language materials, identification of MWEs is accepted to be an important task in natural language processing.

In this paper, considering MWE detection as a binary classification task, we propose to use a semi-supervised learning algorithm, standard co-training [] Co-training is a semi-supervised method that employs two classifiers with two different views to label unlabeled data iteratively in order to enlarge the training sets of limited size. In our experiments, linguistic and statistical features that distinguish MWEs from random word combinations are utilized as two different views. Two different pairs of classifiers are employed with a group of experimental settings. The tests are performed on a Turkish MWE data set of 3946 positive and 4230 negative MWE candidates. The results showed that the classifier where statistical view is considered succeeds in MWE detection when the training set is enlarged by co-training.

The recognition performance of a classifier is affected by various aspects. A huge influence is given by the input data pre-processing. In the current paper we analysed the relation between different normalisation methods for emotionally coloured speech samples deriving general trends to be considered during data pre-processing. From the best of our knowledge, various normalisation approaches are used in the spoken affect recognition community but so far no multi-corpus comparison was conducted. Therefore, well-known methods from literature were compared in a larger study based on nine benchmark corpora, where within each data set a leave-one-speaker-out validation strategy was applied. As normalisation approaches, we investigated standardisation, range normalisation, and centering. These were tested in two possible options: (1) The normalisation parameters were estimated on the whole data set and (2) we obtained the parameters by using emotionally neutral samples only. For classification Support Vector Machines with linear and polynomial kernels as well as Random Forest were used as representatives of classifiers handling input material in different ways. Besides further recommendations we showed that standardisation leads to a significant improvement of the recognition performance. It is also discussed when and how to apply normalisation methods.

Vigilance or sustained attention is an extremely important aspect in monotonous and prolonged attention seeking tasks. Recently, Event Related Potentials (ERPs) of Electroencephalograph (EEG) have garnered great attention from the researchers for their application in the task of vigilance assessment. However, till date the studies related to ERPs and their association with vigilance are in their nascent stage, and requires more rigorous research efforts. In this paper, we use P200 and N200 ERPs of EEG for studying vigilance. For this purpose, we perform Mackworth’s clock test experiment with ten volunteers and measure their accuracy. From the measured accuracy and recorded EEG signals, we identify that amplitude of P200 and N200 ERPs is directly correlated with accuracy and thereby to vigilance task. Thus, both P200 and N200 ERPs can be applied to detect vigilance (in real-time) of people involved in continuous monitoring tasks.