Natural language processing is the future assistive educational technology for the intellectually disabled: A deep dive
Recent advances in information technology, networking, and interface research have provided new tools that will completely redefine the interface concept. Since accessibility is essentially a human interface issue, the entire area of disability access, including the definition of accessibility and assistive technology, will need to be rethought as will the strategies used in the past to create access. However, when we focus on the application of these technologies to enable the intellectually disabled to receive a proper education and training, what assistive technology is superior?
Let’s take a deep dive into explaining why natural language processing (NLP) is the future assistive educational technology for the intellectually disabled.
What are the effects of this technology — why does it matter?
Many people suffer from intellectual disabilities, and it is widely known that there is a high barrier to entry into education for them. We must fulfill our obligation to provide them with an opportunity to receive a proper education, and advancing NLP applications is a way to do so. With all the current technological advancements happening, NLP apps can be sufficiently improved to be properly implemented as assistive tools in educating the intellectually disabled. And when this happens, we may even be able to discover more use-cases for the technology, helping us to solve more problems for the disadvantaged and disabled.
What is the current situation?
As stated in Article 23 of the United Nations Convention on the Rights of the Child, disabled children shall be allowed to have effective access to and receive:
“Education, training, health care services, rehabilitation services, preparation for employment and recreation opportunities in a manner conducive to the child’s achieving the fullest possible social integration and individual development, including his or her cultural and spiritual development.”
This reinstates our moral and legal obligation to pursue technologies that provide more equitable educational and training resources to the intellectually disabled. Let’s look through some previous developments in technology that attempted to fulfill this obligation.
Common GUI Technology
To make computers more user-friendly, efficient assistive tools, GUIs (Graphical User Interfaces) were created. This allowed for people to interact with computers with ease, and as Poole et al (2005) point out, “[w]hile the GUI was not designed for people with disabilities, it made the computer more
accessible for them just as it has made the computer more accessible to the general population.” Nevertheless, companies like IBM have had a long history in making GUIs for the disabled, like a talking web browser (Home Page Reader), a screen reader for the visually impaired, and a speech recognizer to help the disabled interact with the computer using voice commands (ViaVoice). However, none of these technologies were ever targeted for the intellectually impaired, not to mention educational assistive technologies.
OS Tools
Although quite similar to IBM’s technologies, Windows OS, iOS, Andriod, and more have had a variety of accessibility options to assist users with vision, hearing, and mobility disabilities.
W3C Guidelines
The World Wide Web Consortium (W3C) has established certain accessibility guidelines that require Web content providers to be aware of assistive technologies designed to facilitate access to the Web for users who have a disability.
Assistive Software Products
Although not specific to intellectual disabilities, there are some software products called Augmentative and Alternative Communications. These include, but aren’t limited to:
- EZKeys
- Clicker5
- Access for Windows
- Plockaw
- Switch Access for Windows
- Proloquo2Go
- SonoFlex
A Case Study of NLP Applications
NLP contains many different components and processes, but to use NLP for Augmentative and Alternative Communications (AAC), n-gram models are used.
However, let’s start with the most basic type of algorithm used in word processing, prediction, and completion. This is the use of a word/phrase to search a wider known database of words/phrases. For example, let’s say that a user is typing “the ca..,” the system can use its database and provide potential candidates for what the user is trying to type, such as car, cat, cane, can, etc. However, there has to be a more efficient method of doing this, as the number of possible predictions would exceed the number of prediction slots provided by the representative interface (3, 4, 5, etc.). Let’s look at a couple of ways to narrow down the predictions to the number of slots provided by the interface.
The first method is to rank the predictions based on the frequency of occurrence. This is only possible if there is a database with the statistics necessary to use this method. If we do have a corpus with all the necessary data points, we would be able to easily predict the top words that are prefixed with “ca” (e.g.: car, cat, cane, can, call) would be suggested to end the phrase “the ca…”.
The second method uses the basics of the first but uses more semantic information to make a proper prediction. We know that “ca” is preceded by “the,” which means we have two words in this phrase at the moment. This is why we use the n-gram model, where n is the number of words that are being examined. In this case, we have a 2-gram model, which will use the previous word, alongside the letter sequence, to predict the current word. A 4-gram model, for example, would use the previous three words and the letter sequence to predict the current word. We can pair this model with more grammar models to predict the POS of the current word. In our case, we know that we have “the” in our phrase, and the word “the” is usually followed by nouns or noun-modifiers. Thus, we will be presented with more accurate predictions from the system.
Overall, there are many ways to use NLP in various applications, but in the status quo, n-gram models are used by being trained on previous system use.
A Case Study of Intellectual Disabilities
People with disabilities face several challenges in today’s society. According to the European Commission, the overall employment rate of people with disabilities in Europe is 48%. Only 27.8% of people with a disability obtain a tertiary level degree or diploma, and approximately 70% of people with disabilities face poverty or issues that are related to social inclusion.
Let’s look more deeply at specific intellectual disabilities, firstly, reading and writing disabilities amongst students. Nearly 20% of the school population has some kind of reading difficulty, including about 5–8% of the global population with dyslexia.
Secondly, let’s look more deeply at autism, another common intellectual disability. Being the result of a neurological disorder that affects the functioning of the brain, autism and its associated behaviors occur in approximately 15 of every 10,000 individuals. Autism interferes with the normal development of the brain in communicative, cognitive, and social areas.
For these learners, simply ‘trying harder’ is not enough, and ‘more of the same’ training risks that the gap between their reading and writing skills and school requirements will increase over time. In most cases, targeted exercises don’t completely benefit these students, which is when we turn to use assistive technologies, such as listening to text instead of reading, as a way to learn properly.
Now, does NLP help the intellectually disabled?
Over multiple studies conducted in the past 10 years, NLP applications have become more sophisticated, thus more effective in helping people with intellectual disabilities.
In a recently published study, by Svensson et al (2019), 149 students representing various urban and rural schools in Sweden took part in a study “to investigate if an intervention using assistive technology for students with reading disabilities affected their reading ability, ability to assimilate text, and motivation for schoolwork.” Over the 4 weeks of testing, students used NLP applications, such as SayHi (speech-to-text = STT), VoiceDreamReader (text-to-speech = TTS), Prizmo (scanning from written text to digitalized text), Skolstil-2 (an easy word processor and text-to-speech app that even pronounces each sound-letter, words, sentences, and the whole text while writing a text), Legimus (an audiobook reader), and Ruzzle (a word game). The results show gains in reading ability despite using nothing but assistive technology during the intervention period, and these gains were comparable to the enhancement in a control group during 1 year. Approximately 50% of the students and their parents reported an increase in motivation for schoolwork after they had finished the interventions.
In an older study by Fälth et al (2013), experimenters wanted to examine the effects of three computerized interventions on children's reading skills with reading disabilities in 2nd grade. They used five test sessions over one year, and in each of these sessions, a benchmark assessment was taken. During each of the periods over the year, the test groups were given NLP applications to improve word decoding skills, phonological abilities, and sentence creation. The control group received a normal education. The results showed that the groups that received the education with NLP applications showed much greater improvement than the one with ordinary instruction.
To summarize: NLP applications have proven their worth by greatly helping students with intellectual disabilities in learning more efficiently, quickly, and effectively. As evident of this deep dive, the current development focus on the NLP technology is itself improving the processing and applying of NLP to new problems.
🔑 Takeaways
- NLP is quite an intricate technology with many applications and methods.
- Current uses of NLP are somewhat effective but still basic and not quite focused on helping the intellectually disabled.
- If implemented properly, NLP can be a critical assistive tool in bettering the education of intellectually disabled people.
Citations
Fälth, Linda, et al. “Computer-Assisted Interventions Targeting Reading Skills of Children with Reading Disabilities — A Longitudinal Study.” Dyslexia, vol. 19, no. 1, 2013, pp. 37–53., doi:10.1002/dys.1450.
Higginbotham, D. Jeffery, et al. “The Application of Natural Language Processing to Augmentative and Alternative Communication.” Assistive Technology, vol. 24, no. 1, Apr. 2011, pp. 14–24., doi:10.1080/10400435.2011.648714.
Kiliçaslan, Yilmaz, et al. “AN NLP-BASED ASSISTIVE TOOL FOR AUTISTIC AND MENTALLY RETARDED CHILDREN: AN INITIAL ATTEMPT.” (2006).
Manzoor, Mirfa, and Vivian Vimarlund. “Digital Technologies for Social Inclusion of Individuals with Disabilities.” Health and Technology, vol. 8, no. 5, 2018, pp. 377–390., doi:10.1007/s12553–018–0239–1.
Svensson, Idor, et al. “Effects of Assistive Technology for Students with Reading and Writing Disabilities.” Disability and Rehabilitation: Assistive Technology, vol. 16, no. 2, 2019, pp. 196–208., doi:10.1080/17483107.2019.1646821.
Thank you so much for reading!
Shivam Syal is a 16 y/o disruptive innovator, computer science enthusiast, and emerging entrepreneur. Currently, he is looking for ways to use management and technology to address social inequalities arising out of unequal opportunities that are caused by disabilities, socioeconomic status, and global disparities.
Connect with him here 👇
shivamsyal.com | linkedin.com/in/shivamsyal | github.com/shivamsyal
