4 ways machine learning is improving web accessibility for everyone

4 ways machine learning is improving web accessibility for everyone

Digital accessibility refers to the ease with which websites, media content and digital services can be accessed by the widest possible number of people. The team at Savv-e are passionate advocates for inclusion in digital design and we love to hear about advances in technology that can elevate experiences for our learners and content consumers.

With an ageing and diverse workforce, it is becoming increasingly important to consider the needs of everyone, especially as learning experience designers. We want every learner to be able to access and engage with our solutions in their pursuit of knowledge and elevated capability.

Improving the accessibility of digital content is a great start, the only problem is, it has often been difficult to find cost effective ways to achieve this, especially as technology products can be both resource intensive and expensive to implement. This is why we are excited to share some of the new ways automation and machine learning can improve web accessibility.

1. Automated subtitles

Video streaming platforms such as YouTube are now using speech recognition technology and algorithms to improve the viewing experience of people who are hearing impaired. Whilst YouTube has used speech recognition technology to add captions to video for years now, more recently they have enhanced this capability within their algorithms to now identify laughter, music or applause; enriching the experience of video content for people with hearing difficulties.

2. Automated image descriptions

Image recognition is the technology that powers capabilities such as facial recognition for security purposes or self driving capabilities for autonomous cars. But it can also be used to identify images for people with vision impairment.

In 2016, Facebook launched an image recognition feature based on its research that enables text descriptions to be automatically generated for images. This elevates the experience for those who are visually impaired – their friend’s posted photos are described to them via text descriptions that can then be read out through the use of text to speech software.

Further in facial recognition capability, there are now adaptive learning delivery capabilities that use facial recognition capability to identify how learners are responding to e-learning experiences – if you look like you are struggling, then, perhaps you need a less challenging topic. On the other hand, if you are smiling and working through the content quickly, then your content can be changed instantly to deliver more challenging content. It is almost like a personal coach that reads a learners face, and uses the ‘tell-tale’ signs to design better, more personalised learning experiences.


3. Simplification software

With the ever increasing amount of information made available every day, those without the ability to quickly filter and sort information run the risk of being left behind. (As we become more and more reliant on communication via computers or mobile devices, there is a risk that people with intellectual or learning difficulties could be left behind.) One way of addressing this is via text simplification software.

This can be built directly into apps or integrated with commonly used platforms such as Gmail or Facebook. The software can improve readability by breaking up lengthy sentences and replacing complex, ambiguous or confusing terms or figures of speech with easier to understand words and phrases.


4. Automated learning intervention

Educational possibilities are greatly expanded by including computer based or e-learning educational content, but it’s important to ensure no student is being left behind.

Machine learning holds strong promise for helping to identify learning difficulties, by detecting language or learning disorders and prompting early intervention.

Recent research into this area has found that children who are affected by neural disorders that can affect speech or comprehension display certain patterns when performing set tasks. Data and Analytics are able to offer insights into learning experiences that are not set to the right level of challenge, allowing for early intervention.

Therefore automated screening tools can be used that ask children to narrate a series of pictures, where pauses or difficulty with certain tenses or language patterns can indicate known learning difficulties.

Using screening processes such as these can help provide wide scale and regular testing of students to identify which students need additional or targeted assistance.

Looking forward

The pervasiveness of the internet and digital services in our daily lives will only continue to increase. As this occurs it is important to ensure that everyone can benefit from these technological developments.

To help improve the accessibility your pages we've developed a handy checklist of important accessibility requirements - click the below image to download it.

digital_accessibility_checklist

Related Articles

How To Fake AI Without Really Trying

This is an example of how Savv-e uses gamification to match the user experience with specific learning outcomes.

In this case, we designed an emotional 'chat-bot' to train staff in the finance industry in dealing with angry customers. The results were surprising and a lot of fun.

Read More

Savv-e shorts...little bites of e-learning wisdom

 

Savv-e shorts from the Savv-e Team...little bites of e-learning wisdom to take you on your e-learning journey.

Read More

The importance of user experience and why you ignore it at your peril

Most of us have had the misfortune of taking a poorly designed online course. One of the worst I ever experienced had screens and screens of small text in a barely readable font coupled with audio delivered by a monotonous computer-generated voice. I was a hostage to the content unable to even skip ahead on the bits I already was really familiar with. Needless to say, I didn’t remember much of that course and I am pretty sure that despite the certificate I got at the end, I didn’t take much away from it.

Read More

Join 5000+ learning professionals and receive our email updates.