Cognitive Machine Learning Framework to Evaluate Competency during the Course Endorsement for Computer Science Postgraduates
SHIVANI A. TRIVEDf and REBAKAH JOBDAS
Faculty of Computer Science, Kadi Sarva Vishwavidyalaya, Gandhinagar, Gujarat 382015, India
The approach of content designing along with content delivery by the facilitator changes in accordance to the environment undergoing a dramatic shift. Career-oriented learners look forward for the endorsement in the course which persuades them in career advancement. In India looking at the current scenario having more than 789 universities and 20,000 colleges offering postgraduate courses, the challenging task is to find better course endorsement options. The research framework is based on neuro-education, the consequence of using brain imaging methods is to serve as analytical tools for measuring the educational interferences of the learner for shaping a better arrangement for increase learning ability. The nemo signals collected through the EEG device are analyzed with the machine learning algorithms resulting to a cognitive machine learning model to foresee the competency of student in picking a course labeling a output to Attention Span endeavor to choose the course. The research work will be organized in four phases. Phase 1 is discretize cognitive skill of a learner. Phase 2 is procuring nemo signals of a learner. Phase 3 is applying active learning methods to facilitate the content at the time of the course endorsement. Phase 4 is developing cognitive machine learning model to generate nemo feedback. The nemo feedback is generated, beneficial to facilitator as well as the learner. The learner can endorse the course to uplift knowledge domain. In this study, a sample size of 60 undergraduate program and postgr aduate progr am learners are taken into consideration while choosing elective course for the higher semester.
This chapter amalgamates cognitive machine learning (CoML) and the brain-computer interface (BCI) framework to elate competency during the course endorsement for computer science postgraduates. The main aim of this chapter is to represent the development of the CoML framework to predict the competency of a learner as observer, performer, listener, and reader. This chapter also aims to identify learner’s skill in learning a course, which is benefiting from gaining confidence after successful course completion. The other reason is to impact the effective teaching-learning method’s applicability on the learners. The CoML model could assist the facilitator in predicting learner’s performance and the effectiveness of an adapted approach in content designing [ 1 ]. These CoML techniques could provide guidelines to the facilitator for developing better teaching strategies to improve learner’s performance . This study can be further carried out on a prototype for a web-based assessment tool at a global level. The research work is organized in four phases. Phase 1 determines the cognitive skill of a learner. Various cognitive skills are responsible for critical thinking, appraised to the learner counting on their learning ability. The learning ability categorizes the learner categories as performer, learner, reader, and observer. Phase 2 procures neurosignals of a learner. The electroencephalography (EEG) signals in the form of alpha, beta, gamma, and theta waves are acquired through an EEG device. The raw signals are filtered and converted from analog to digital values ranging from 0 to 100 Hz. Phase 3 applies active learning methods to facilitate the content at the time of the course endorsement. Active learning methods are used to gain the attention of the learners. Active learning methods using an educational game with the audiovisual content can be adopted. To take more involvement of learners in the teaching-learning process, Phase 4 develops the CoML model to generate neurofeedback. The CoML model is the reverse engineering process to evaluate the cognitive ability of the learner by predicting the outcomes after performing the experiments based on active learning methods . The neurofeedback, when generated, is beneficial to the facilitator as well as the learner. The learner can endorse the course to the uplift knowledge domain. The facilitator can explore new pedagogies to amuse the learner's ability. In this chapter, for experiment purpose, a sample size of 40 postgraduate learners in computer science was taken into consideration. While choosing an elective course such as artificial intelligence (AI) and digital marketing (DM). AI is human intelligence in a simulation environment for information acquisition. DM is a marketing trend using digital technologies. These two subjects offered as elective for computer science postgraduates; the following section helps in course endorsement.
COURSE ENDORSEMENT OPTIONS
The approach of content designing along with content delivery by the facilitator changes by following the environment undergoing a dramatic shift. The endorsement into a particular course equips the learner for high-quality ability along with the academic rigor expected . The exploring of learning theory, strategies for communication between facilitator-learner and leamer- leamer, engagement techniques between facilitator-learner, leamer-leamer, and learner-content contingent upon the cognitive skill applied, and the knowledge retention  can be achieved. Career-oriented learners take the endorsement in the course, which persuades them in career advancement. In India, looking at the current scenario having more than 789 universities and 20,000 colleges offering postgraduate courses, the challenging task is to find better course endorsement options. Many prominent institutions in the world such as Digital Vidya and Coursera provide demo sessions online. Based on those demo sessions, students decide to endorse either AI or DM.
The prospective of neuroscience training packages is to surge the inclusive brain utility affected by the neurocognitive functions, which abode in the brain circuits while performing core tasks such as observing, experimenting, listening, and reading. This paves a way to augment academic skill acquisition . The consequence of using brain imaging methods is to serve as analytical tools for measuring the educational interferences of the learner. To read through the brain and to detect signals, norrinvasive methods such as functional magnetic resonance imaging (flVIRI) and EEG are used. These neuroimaging methods have a latent ability to enhance the cognitive subprocess at a profound level than reliance on only behavioral methods .
Neuroeducation is a promising area spanning the gap to progr ess bidirectional activity in the brain of learners and facilitators impacting the learning process . Neural science and cognitive bases of academic skills for shaping a better arrangement for learning environments optimally achieve these capabilities according to the learning ability . The neurosignals collected through the EEG device are analyzed, and the CoML model is developed to foresee the competency of the student in picking a course . Machine learning (ML) aims toward the development of computer programs that can access and utilize data for personal upgradation. In this chapter, the algorithm builds a CoML model from a set of EEG data, which contain input data from the brain, and labels the output as attention span or concentration . Therefore, unsupervised learning algorithms are used to find structure in the data, such as grouping or clustering of data points based on maximum attention, attention span, and duration of time spent by the learner in the two-step course endorsement activities. These outcomes have tossed a light for comprehending the root of disarrangements looked by learners. To endeavor and choose the course and its substance to complete the speculation of tune in discovering that course is better for professional success or not can be obtained. At the facilitator side, the insight is to revise the content and assure the learner to boost in course endorsement.
FIGURE 2.1 Description of brain frequency signals.
The essence of the thought process, sensations, and performances is the passing of flow of charge between the neurons within the brain. Harmonized electrical rhythms form brainwaves from many types of neurons passing information to each other, as shown in Figure 2.1. The delta band frequency records below 4 Hz are observed in children below the age of one year and decline as they gr ow. In grown-up individuals, delta rhythms are customarily known in deep, dreamless sleep and trifling in adults when awake. If these bands recorded in colossal number signify deformity then it is a link to a neurological disorder. Sometimes, these bands are taken as an artifact, which is a noise signal generated by the muscular movement of the neck or jaw. Theta band frequency ranges from 4 to 7 Hz. An insignificant quantity of theta waves chronicled if an individual is receptive and energetic. When the individuals are in drowsy or meditative, theta waves are recorded at the upper level. Compared to delta waves, a huge amount of theta activity hr receptive individuals is associated with a neurological disorder. The theta band has been concurrent with meditation or concentration and a wide series of cognitive processes such as reasoning or sensible state . Alpha band frequency is between 8 and 12 Hz. Their magnitude upsurges when their own eyes shut and the body relaxes, and they incapacitate when the eyes are open and rational thinking is on. Alpha activity is associated with logical reasoning. Cumulative mental effort and reasoning cause a conquest of alpha activity, particularly from the frontal areas . Subsequently, these rhythms might be applicable signals to measure cognitive skills. Beta band frequency ranges between 13 and 30 Hz  related to wakening mindfulness or responsiveness. Attention or concentration is measured when the brain signals are detected with beta waves. Beta brainwaves are classified into three bands depending upon the frequency: low beta (12-15 Hz) is associated with a “fast idle” or absorbed thought process, beta (15-22 Hz) is associated with logical thinking or dynamically problem-solving ability, and high beta (22-38 Hz) is associated with intricate thinking ability, incorporating new capabilities, and high apprehension. Persistent frequency of brain signal processing is not a proficient way to invoke the brain, as it takes remarkable energy. Gamma band frequency signals range from 30 to 100 Hz. The existence of gamma waves in the brain activity of a healthy individual connects to certain muscular movements. Certain research trials have publicized an association in individuals between motor actions, and gamma waves have been traced during muscle contraction . The EEG signal recording system consisted of electrodes with conductive media, amplifiers with filters, analog-to-digital converter, and the recording device. BCI readings are detected when electrodes are in use. They are basically of the following types: single-use electrodes (gel-based), reusable electrodes (gold-plated, silver-plated, stainless steel, or tin), headset and EEG cap, saline-based electrodes, and needle electrodes . There are some companies hosted BCI games in the market. A company named Emotiv [13,16] has technologically developed EPOC Neuroheadset. It introduced several sets of BCI-based games, such as Cortex Arcade and Spirit Mountain Demo Game. Besides, the company owns a highly efficient BCI, called EPOC Neuroheadset, with 14 electrodes at a very reasonable price along with a friendly interface for programmers. They set a fast development strategy to BCI-based applications. NeuroSky, another company, also arcades the headset named Mind Wave. Neuroheadset with the inbuilt interface for developing software applications can accumulate brainwaves and derivate into various mental states. Many other software companies such as Microsoft have revealed curiosity to research BCI, challenging the expansion of preliminary original applications that used BCI . The Mind Wave Mobile Headset consists of a measuring device and a sensor that contacts with the forehead. The orientation points are positioned on the ear pad. The microcontroller chip processes the signals and delivers data to applications in digital form. This instrument is connected to the analyzer via Bluetooth. The EEG electrode is positioned on the own forehead (on the frontal cortex) during an experiment. The headset securely processes the signals and delivers outputs in the form of EEG power scales (alpha waves and beta waves), which rate to the level of attention, meditation, and measure of an eye blink. Ranges of attention and meditation are specified and conveyed to a meter with a virtual e-Sense range of 1-100. Values 20-40 are abridged ranges, which are not suitable for consideration, and values ranging from 1 to 20 reflect strongly sunk e-Sense values. The values falling in the range of 40-60 are reflected to be unbiased. The resulting values greater than 60 are considered to be elevated values. Values in the range of 80-100 are considered to be high levels of e-Sense.