Educational AI Application

Educational AI Application
Student’s Name
Institution
Course
Professor’s Name
Date
Educational AI Application
Technology has grown rapidly across the globe. Internet connectivity and smartphone penetration have enabled more people to access technological developments faster than before. The use of technology has penetrated more areas, including the educational sector. It is not long ago when teachers would consider mobile phones a source of distraction for learners. Teachers and parents tried everything to ensure that students do not use phones while in school. However, the widespread acceptance of mobile phones has seen more tutors encouraging students to use their mobile phones constructively. The development of mobile apps targeting the educational sector has been instrumental in enhancing education (García-Martínez et al., 2019). Mobile apps have enhanced the learning process and allow both teachers and students to interact more closely. The inclusion of artificial intelligence (AI) in educational apps enables teachers and students to benefit more from the technology. I would be making an app to enable students to learn online. The application would enable tutors to conduct online classes without the need for students to attend physical classes. The app would assist in reducing the cost of attending physical classrooms. The move would be critical in meeting the educational needs of learners from low-income backgrounds. Further, the app would minimize physical contact among students and assist in curbing the spread of COVID19 in schools. The app demonstrates that it is possible to apply AI in learning successfully.
Automatic Speech Recognition
Automatic Speech Recognition (ASR) enables individuals to use their voices to speak with a commuter interface in a way that resembles normal conversations (Raut & Deoghare, 2016). ASR would enhance the learners’ skills in the language. Firstly, the feature would facilitate reading and writing, where students can obtain information on critical elements of phonemic awareness, like the link between sounds and symbols. For instance, as learners speak, they can see the words appearing on the screen. Such correspondence directly indicates the relationship between the appearance and the sound of a particular word. Secondly, ASR would allow students to alternate between typing and speaking. Speech recognition would ensure that learners can shift between the two areas and improve both writing and speaking skills (Michael, 2017). Thirdly, the feature would reduce writing fatigue. The ability of the app to recognize speech means that students do not need to engage in the physical act of composing to paper or keyboard. Instead, learners can use their energy to address other aspects of learning.
Additionally, the feature minimizes the anxiety involved in the organization and editing of text. ASR can perform such tasks effectively and enable students to focus on other aspects of learning. Lastly, ASR would be useful to students with disabilities (Michael, 2017). It can assist struggling writers in boosting their writing skills and offer alternative access to the app for individuals with physical impairments.
The app would need to recognize both referential utterances and transactional utterances. Referential utterances provide information and would be useful in supplying students with the information they need in a particular area (Tribus, 2017). For instance, assuming a student is studying geography and wishes to know the world’s capital cities, the app should provide that information accordingly. Referential utterances will enable students and teachers to get the information they require in their respective fields. Transactional utterances emphasize getting something done. In this case, a speaker issues a command and expects it to be done. Transactional utterances would be needed in the app to ensure that users can issue commands to the app and have their directives followed. For instance, students may wish to submit assignments. Issuing transactional utterances would allow such students to submit their work. Both referential utterances and transactional utterances are essential to the app and would be incorporated. They would make it easier for users to interact with the application.
Models
The app would need a Knowledge Tracing (KT) model. The idea of KT is modeling student’s performance and acquisition of skills. KT is widely utilized to update a tutor’s estimate of the probability of a learner having a particular skill (Xu & Mostow, 2012). The tutor uses the learner’s observable performance on the steps that utilize the particular skill. In the app, the KT model will use the following five parameters. Firstly, the probability of slip failing on an already known skill (Xu & Mostow, 2012). Here, the app will check the probability that students will make an error after acquiring a skill. The second parameter is the probability guess of excelling in an unknown skill. Where the skill is not learned, this parameter will check if the learner can guess correctly. Thirdly, the model will determine the probability of knew. Here, the parameter will test how well learners can know a skill before they practice it. Fourthly, the model will determine the transition probability, where it will determine how students move from not knowing a skill to acquiring it (Xu & Mostow, 2012). The last parameter would be the probability forget. The parameter tests the probability of moving from knowing a skill to not knowing it at all.
The second model would be the Rasch model. The model is instrumental in measuring latent traits like attitude and ability. The model demonstrates the probability of a learner getting a correct response to a particular test. The model has the less difficult items on the left and the more difficult ones on the right. Students can also be placed on the same measurement scale, with those less able on the left and the more able on the right. For instance, assuming that a test has 10 questions, three students A, B, and C, can be categorized depending on their ability. A is expected to get 3/10 questions correctly, while B is expected to get 6/10 right and C score 10/10 correctly. Such categorization assists in determining the ability of learners in particular areas. In the app, using the Rasch model would be critical in showing the learners’ attitude towards a particular subject. The model would provide tutors with feedback regarding a student’s progress. The model would assist tutors in establishing learners’ abilities and hence customize the learning process to meet the needs of the students.
Educational Data Mining
The app would collect usage data. Here, the app would collect any data related to how the users interact with the software. All user interactions with the app would be captured. Usage data is critical in evaluating the relevance and value of the app (Schmitz, Bartsch & Meyer, 2016). For instance, if users open the app and close it immediately, chances that it is not relevant are high. However, where a user repeatedly spends a long time on the app, there is a high likelihood that it is very important. Monitoring usage data would demonstrate how the user...