How Machine Learning Can Improve Learning Disability
The best and well known classification for learning disabilities or disorders is the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders known as DSM-5. Mental disorders covered by DSM-5 may include: dyslexia, dyspraxia, dysgraphia, and more
Learning Disorders are defined as "subtle and regionally distributed variations in mind-body systems" and "should not be confused with learning difficulties that may be caused by visual, hearing or motor impairments, as well as social challenges."
In this article we have a closer look at dyslexia as a learning disorder. Machine learning will be assessed for its effectiveness in classifying, understanding and assisting users with dyslexia.
Dyslexia is defined as a learning disability affecting the ability to read. Motor dysgraphia affects the ability to write and may be a sign of developmental coordination disorder (DCD) like dyspraxia.
Understanding disabilities of this nature is important since roughly 10% of the population has some form of learning impairment. Learning impairments, if gone undiagnosed, may lead to academic underachievement and in the worst of cases academic failure.
Diagnosing impairments early along with the use of educational solutions can prevent lifelong academic hardships and potentially elevate the person to their "non-impaired" peers.
Dyslexia is a specific learning disability that is neurobiological in origin. It is characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities.
Source: Multisensory Reading Center
Dyspraxia is a neurodevelopmental disorder of movement and coordination in which messages sent from the brain to the muscles are interrupted. It is often identified in early childhood, but can also come on later in life after an illness or acquired brain injury.
Simply stated, "machine learning is a technique for ingesting and recognizing new patters of information from large amounts of data. It enables researchers and information professions to recognize the strategies and plans that will be devised with ease" AND "machine learning is a growing trend in healthcare that aids medical professional in better investigating, anticipating, and treating patients."
Why is machine learning important?
Machine learning is capable of sifting through massive amounts of data and finding patterns that would otherwise be invisible to a human researcher. In fact, a research could devote their entire lifetime to a discovery where a machine could come to the same conclusions in moments.
The algorithms and tools used to quantify, assess and understand data becomes the key to the overall success of machine learning. On its own machine learning is ambiguous, but with neural networks, decisions trees, k-means clustering and the like we open up the raw processing power of machine learning with the ingenuity of human intelligence. Processing power and human insight can be merged to locate helpful strategies much quicker than ever thought possible.
Machine learning is a branch of artificial intelligence (AI) and computer science which focuses on the use of data and algorithms to imitate the way that humans learn, gradually improving its accuracy.
- Diagnosis and Classification system to detect the negative effects of dyslexia. This would be used as tool for determining dyslexia
- Analysis of external appearance using neural networks in e-learning systems and employing preprocessing, feature extraction, and classification. Like point 1, this would determine the likelihood of dyslexia being present
- Computer programs for learning the alphabet for dyslexic children that are specifically tailored for their unique learning difficulties. The computer program can adjust to each individual child. These programs would allow immediate support before extensive study and analysis is conducted by practitioners. Therefore the time needed for diagnosis and planning is diminished drastically
- Mobile reading applications. The application can read out words for them so they can understand the text. Words can also be modified in a variety of ways based on their skill levels. In this way machine learning can be employed to improve the dyslexic users reading comprehension by employing techniques that are geared towards the specific user rather than employing generic frameworks
- Sensing user behavior based on engagement. Use machine learning and image categorization to predict student engagement. Results were determined by using Support Vector Machine (SVM), Nave Bayes and K-Nearest Neighbor (k-NN). If predictive results are found to be effective this could be a valuable tool for assessing student engagement which can then be used to analyze future learning technique amongst dyslexic users and those with a wide range of learning disabilities
- Using machine learning to analyze how people engage with a game and detect dyslexia early. This would be done in the following ways: a) an empirical etymological analysis of errors commonly made by dyslexic individuals and b) specific dyslexia-related cognitive functions in mind; being: language skills, working memory, executive functions and perceptual processes
- An eye-tracker is used to analyze eye movement to distinguish between dyslexics and non-dyslexics. Characteristics of eye movement that may be identified include: obsessions, saccades, transients and mutilations. Principal Component Analysis yields characteristics from the tracker data and in turn a Hybrid Kernel SVM-PSO based on Particle Swarm Optimization (PSO) can be used to predict dyslexia
Support Vector Machine
Particle Swarm Optimization (PSO)
The information for this article was inspired by a research paper entitled, "Application of machine learning techniques for improving learning disabilities."
The research methodology was summarized and the ways in which machine learning can aid those with learning disabilities were highlighted.
We did not cover all sources listed in the literature review. Instead, we took those that appeared to be most promising, interesting and unique. We recommend finding the research paper and reviewing it on your own if you are interested in learning more.
International Journal of Electrical Engineering and Technology (IJEET) . Volume 11, Issue 10, December 2020, pp. 403-411, Article ID: IJEET_11_10_051. ISSN Print: 0976-6545 and ISSN Online: 0976-6553 DOI: 10.34218/IJEET.11.10.2020.051