This study explores the integration of innovative technologies, such as video games and artificial intelligence (AI), in education, focusing on the use of Deep Reinforcement Learning (DRL) to create engaging educational tools. The project “Deep Neural Tower” presents a video game designed to teach fundamental AI concepts to high school students through interactive gameplay. Developed using the Unity engine and ML-Agents toolkit, the game incorporates AI agents that learn and adapt through interaction with the environment. Students encounter challenges illustrating key AI concepts, such as agent-environment interaction, reward systems, and neural networks, which promote active learning and knowledge retention. Tested with 111 high school students, the game showed positive results in usability and engagement, enhancing understanding of complex AI concepts. This paper discusses the pedagogical benefits of using AI in education and concludes that combining advanced AI techniques with interactive gameplay has immense potential to transform STEM education.

Nowadays, the learning of new technological areas such as artificial intelligence is taught in early educational cycles. This study will address the issues of education and software with the aim of developing a software tool that facilitates basic learning about neural networks. For this purpose, an application has been developed to allow high school students to learn basic concepts. Also, to check the degree of usability of the application, twenty participants have been surveyed and responded to a standardized usability test accepted by the research community.

Smith–Magenis syndrome (SMS) is a rare, underdiagnosed condition due to limited public awareness of genetic testing and a lengthy diagnostic process. Voice analysis can be a noninvasive tool for monitoring and detecting SMS. In this paper, the cepstral peak prominence and mel-frequency cepstral coefficients are used as disease monitoring and detection metrics. In addition, an efficient neural network, incorporating synthetic data processes, was used to detect SMS in a cohort of individuals with the disease. Three study cases were conducted with a set of 19 SMS patients and 292 controls. The three study cases employed various oversampling and undersampling techniques, including SMOTE, random oversampling, NearMiss, random undersampling, and 16 additional methods, resulting in balanced accuracies ranging from 69% to 92%. This is the first study using a neural network model to focus on a rare genetic syndrome using phonation analysis data. By using synthetic data (oversampling and undersampling) and a CNN, it was possible to detect SMS with high levels of accuracy. Voice analysis and deep learning techniques have proven to be a useful and noninvasive method. This is a finding that may help in the complex identification of this syndrome as well as other rare diseases.

The production of phonation involves very complex processes, linked to the physical, clinical, and emotional state of the speaker. Thus, in populations with neurological diseases, it is possible to find the imprint in the voice signal left by the deterioration of certain cortical areas or part of the neurocognitive mechanisms that are involved in speech. In previous works, the authors determined the relationship between the pathological characteristics of the voice of the speakers with Smith-Magenis syndrome (SMS) and a lower value in the cepstral peak prominence (CPP) with respect to normative speakers. They also described the presence of subharmonics in their voices. The present study aims to verify whether both characteristics can be used simultaneously to differentiate SMS voices from neurotypical voices. It will also be analyzed if there is variation in the trajectory of the formants coinciding with the subharmonics. To do this, the effect of subharmonics in the voices of 12 SMS individuals was isolated to see if they were responsible for the lower CPP values. An evaluation of the CPP was also carried out in the areas of subharmonic presence, from the peak that reflected the value of f0, rather than using the most prominent peak. This offered us a baseline for the CPP value in the presence of subharmonics. It was checked if changes in the formants occurred synchronously to the appearance of those subharmonics. If so, the muscles that control the position of the jaw and tongue would be affected at the same time as the larynx. The latter was difficult to observe since the samples were very short. A comparison of phonatory performance of a sustained /a/ between a normotypical group and non-normotypical group of children was carried out. These groups were balanced and matched in age and gender. The Spanish Association of Smith-Magenis Syndrome (ASME) provides almost 20% of the population in Spain. The CPP allows differentiating between normative speakers and those with SMS, even when isolating the effect of subharmonics. The CPP is a robust index for determining the degree of dysphonia. It makes it possible to differentiate pathological voices from healthy voices even when subharmonics are present. The presence of subharmonics is a characteristic of voices of SMS individuals and is not present in healthy ones. Both indexes can be used simultaneously to differentiate SMS voices from neurotypical voices.