 Hello everyone. I am Vikas Ashok. I am an assistant professor in the Department of Computer Science at Old Dominion University, Norfolk, Virginia. In this presentation, I will discuss the usability problems that students with visual disabilities typically face while interacting with computer applications. It should come as no surprise that the use of computers in education has been steadily increasing and in many ways has revolutionized traditional education. With the internet, students have access to a wealth of information and knowledge at their fingertips that can help them develop their skills while also preparing them for a future career in a workforce that is increasingly reliant on computer technology. Students rely on a diverse set of computer software including education, technology services such as Blackboard, online collaborative tools such as Google Docs, productivity tools such as PowerPoint and Word, web browsing tools such as Google Chrome, communication software such as Zoom, etc. With the increased adoption of computing applications in education, computer accessibility has become an important factor to ensure equity of access since students with disabilities too should be able to access and benefit from the assortment of computing applications like their peers without disabilities. The disabilities can be broadly categorized into different groups such as cognitive impairments, visual impairments, hearing impairments, motor impairments and so on and each group has unique set of requirements for conveniently interacting with computer applications. In this presentation, I will exclusively focus on students with visual disabilities such as blindness and low vision. Students with visual impairment access computing applications via assistive technologies such as screen readers and screen magnifiers. Screen readers are mostly used by blind users to listen to the application content and also navigate the content using special keyboard shortcuts. Screen magnifiers on the other hand are used by students with low vision to enlarge the application content and also pan the content in different directions to view the occluded portions. While the assistive technologies such as screen readers and screen magnifiers indeed make content accessible, they however do not ensure usability, that is the ease with which students with visual disabilities can interact with the applications. Usability involves subjective aspects of interaction such as user satisfaction, learning curve, that is how long it takes for users to become familiar with an application and comfortably interact with it. Efficiency as in how quickly users can complete their tasks and memorability, that is how easy it is for users to remember the application structure and layout so that they can quickly access the different features in subsequent interactions with the applications. Many studies have been conducted to assess the usability of web applications for people with visual disabilities and they have all found that there exists a wide usability gap between the interaction experiences of people with visual disabilities and those without disabilities. For example, a few studies have shown that it takes a few minutes for blind screen reader users to locate certain specific information in a website whereas sighted users can accomplish the same in a matter of few seconds. A mental load on users with disabilities is also significantly higher than those without disabilities since they have to remember a lot of information during interaction. Sighted users on the other hand can almost instantly access this information whenever desired by a quick visual scanning of the website. To illustrate the difficulties faced by students with visual disabilities, let us look at a couple of examples. This example illustrates how blind students interact with Google Docs and Blackboard applications using screen reader shortcuts. To activate press space bar, menu to move through items press up or down arrow. Leaving menus last in university-google-docs-goop-mode and view tool bar editing mode button menu press space to active side panel complementary region tab. Hide side panel toggle button to toggle the state press space bar listed with seven items comment accessible user CS graphic. Mark as resolved and hide discussion button to activate press space bar. More options dot dot dot button menu press enter. Page down filter. Main region courses filter button menu. 25 items per page button menu. 202.1.1.0 underline CS7.222 underline to more add to 202.1.1.0 underline CS more add to 202.1.2.0 underline CS6.25 add more info for 202.1.2.0 underline CS6.25 add to notice how the interaction is tedious and sequential for blind students. In contrast, content access is almost instantaneous for sighted students with the help of a visual input device such as a computer mouse. Now let's look at how low vision students interact with Google Docs and Blackboard using a screen magnifier. Notice how most of the content is occluded due to content enlargement and therefore the student has to pan all over the screen to access the occluded portions of the content. Notice how the student cannot view both the document and the comments at the same time under the magnifier lens and therefore they're forced to remember the content of one of them to interpret the other. The sighted student on the other hand can easily associate comments with their respective text in the document. Now let's look at Blackboard application. Even here, most of the content is occluded due to content enlargement and the student has to pan from left to right to access the occluded portions. To understand the usability issues faced by visually impaired students in detail, we conducted an interview study with 40 university students who had some kind of visual disability. Most of these students had low vision due to conditions such as cataract, glaucoma, retinitis pigmentosa, optic atrophy, etc. and therefore primarily relied on a screen magnifier to interact with computers. Most of them also required screen reader support in addition to screen magnifier for interacting with applications. Only 6 student participants were fully blind and therefore used only screen readers. The majority of the participants identified themselves as male and the remaining 12 participants identified themselves as female. The average age of the participants was 24. The interviews were semi-structured which included the questions shown in this slide. The interviews started with questions about the types of applications used by the students in general and for their coursework. Questions about training and external assistance required for interacting with these applications and questions about the specific usability issues the students typically face while interacting with these applications. Our study revealed that visually impaired students too use an assortment of computer applications for their everyday activities. Almost all students stated that they are active on social media such as Twitter and discussion forums such as Reddit. For coursework, they rely on learning management systems such as Blackboard, Canvas, or Piazza as well as productivity tools such as Word, Excel, PowerPoint and collaborative tools such as Google Docs. Students also stated that recently their dependence on communication tools such as Zoom and Skype has significantly increased due to the ongoing COVID-19 pandemic. Majority of students also mentioned that they use different devices to interact with applications based on their convenience. All students stated that they need to invest significant amount of time and effort to get acquainted with unfamiliar computing applications or even familiar applications after software updates. This process typically involves trial and error exploration of the application's graphical user interface using either screen reader shortcuts or screen magnifier panning to locate the various application features such as menu, controls, login, main work area, etc. and also understand how to navigate between these features during interaction. To enable smoother subsequent interactions with the same applications, the students mentioned that they devised interaction strategies. The screen magnifier users stated that they memorized landmarks on the application screen that helped them determine the general panning directions for accessing specific application features. They also memorized the appearance, example, color, relative font size, etc. in addition to the text label of the features so as to easily recognize them while panning. Similarly, screen reader users mentioned that they devised keyboard shortcut strategies that helped them quickly navigate to application features without having to listen to a lot of irrelevant content. However, learning and memorizing strategies both panning and keyboard shortcuts is both time-consuming and cognitively burdensome. Moreover, as software gets frequently updated, these students have to go through the arduous learning process over and over again to maintain familiarity with the applications. The study also revealed that unlike sighted students who can independently complete their coursework in most scenarios, students with visual disabilities constantly rely on their sighted peers and instructors to finish their tasks. These tasks include locating course materials on Blackboard and instructor's websites, filling out PDF forms online, inspecting written content for errors, locating the application feature that lets them submit assignments, accessing application features, and so on. This clearly highlights the shortcomings of modern assistive technologies with regard to computer applications used in education. All students mentioned that application heterogeneity was the fundamental cause underlying many of their interaction problems with computer applications. As applications are developed by different vendors, they tend to vary in their presentation of content and controls. While this difference in appearance between two applications, example Microsoft Word and Google Docs, designed for the same purpose, causes little to no inconvenience for sighted students. It causes a lot of problems for visually impaired students, especially for blind screen reader users, since they have to memorize different shortcut strategies for doing the same tasks in these applications. Screen magnifier users too need to remember different landmarks for different applications, so as to reduce spanning effort while doing tasks. Furthermore, assistive technologies are platform-specific. Example, the Jaws screen reader can only be used on the Windows platform, and therefore the students have to memorize the key bindings for different screen readers if they need to access applications on different platforms. As COVID-19 pandemic has significantly increased the reliance on remote instruction, students with visual disabilities are facing new usability problems due to the use of communication tools such as Zoom or Skype. For example, students relying on screen readers express difficulties in accessing the various Zoom controls such as screen sharing, since screen reader sounds overlap with human voice. Also, since the lecture presentations are related over the communication channel as live video streams, they are inaccessible to screen reader users, and therefore blind students can only rely on instructor speech to understand the content. Even students with low vision complain that they cannot see the entire presentation due to content enlargement and additionally, since the context is continuously changing during lecture, they cannot pan to view the occluded portions. Hence, they have no choice but to understand the lecture content based on the limited contextual information they can see in real time with the screen magnifier. Even reviewing recorded lecture videos is time-consuming and frustrating since they have to repeatedly pause and pan to view the full frames in these videos. Therefore, these students stated that they often spend extra time asking their instructors and sighted classmates to explain some of the content in the lecture videos. In sum, visually impaired students are not yet enjoying the benefits of computing applications to the same extent as their sighted peers, compared to sighted students, visually impaired students need to invest significantly more time and effort before they can comfortably interact with computing applications. While the universities typically provide support in terms of additional time for completing assignments and exams, this does little to address the stress and frustration experienced by the students due to the limitations of the available assistive technologies. Remote learning due to the pandemic has made it even more difficult for students to independently complete their tasks. To address these challenges, there is an imminent need for smart assistive technologies that can provide uniform access to different applications across different platforms so that the content and cognitive burden on students with visual disabilities is significantly reduced. Leveraging recent advancements in artificial intelligence techniques to substantially enhance the capabilities of assistive technologies can help mitigate usability issues faced by students with visual impairments when they interact with computing applications. Present assistive technologies are simple relay systems that either convert content to speech or enlarge it for comfortable viewing. With AI support, the burden of trial and error exploration and layout memorization can be transferred from students to the intelligent assistive technologies. Such AI-enabled assistive technologies will be capable of automatically analyzing the application content to determine all the supported features and then enable students to use voice commands to directly access the desired features. Alternatively, these assistive technologies can automatically and dynamically extend the present application layouts by adding proxy interfaces which are more suitable for either blind or low vision interaction. AI-based solutions also exist for convenient low vision interaction with videos and these solutions can be integrated into assistive technologies so as to provide a platform-wide support for all applications involving videos. In sum, AI-enabled assistive technologies has the potential to mask the adverse impacts of application heterogeneity on the computer interaction experiences of students with visual impairments. This concludes the presentation. Thank you very much for your patient hearing.