At the heart of Microsoft’s European headquarters building in the south Dublin suburb of Leopardstown is a facility known as DreamSpace. Part of Microsoft’s wide-ranging schools programme, the specially-designed facility offers teachers and students from all over Ireland opportunities to learn about the latest technologies and to use them in fun ways.
“We created DreamSpace when we opened the new building in December 2017,” says Kevin Marshall, head of education in Microsoft. “It provides experiential learning for primary and secondary school children. We have two core curriculum models: one for 10 to 12-year-olds and the other for 16 to 17-year-old transition year kids.”
The aim is to give the children an appreciation for what Stem (science, technology, engineering and maths) is and what it can mean for them in the future. The children get to learn about coding, technologies like AI, wearables, the augmented reality and virtual reality features of Windows and much else besides.
“We get kids to engage in two to three hours of short thematic lessons,” says Marshall. “For the transition year students it’s a little more focused on the application of the technology in the world of work and the coding might be more advanced. We host two schools a day, five days a week and it’s been up and running since last April. Since then we’ve had approximately 12,000 kids through the door from schools around the country and we’re fully booked up until the end of June.”
During school breaks and holidays the facility offers learning opportunities for Coder Dojo clubs and trainee teachers and others.
“We run a series of coding camps for kids in the summer,” Marshall adds. “It’s been fantastic. I knew there was a need and a desire there, but I didn’t realise the scale. I’ve been blown away by the extent of the demand.”
This is just one example of Microsoft’s engagement with the Irish education system. “We do a lot of work on the development of curriculum resources to support the education system. We have been working with Springboard for the past seven years on helping provide the right third-level courses for the current economic environment. As part of a global programme we are helping a number of showcase schools across Ireland with digital transformation.”
The company is also using AI to improve the educational experience. “It’s all about using the technology to enhance learning,” Marshall explains. “It doesn’t have to be about Stem. We are using technology to enhance learning and education.”
Teaching aids
Propensity modelling, which helps predict educational outcomes for children, is an effective example of how AI can assist the education system. “We did a piece of work with Washington State University where we created an early warning system for kids in danger of dropping out. The system uses machine learning and predictive analytics to warn if a child needs an intervention.”
Another very practical use for the technology is automated assessment of oral examinations
Access to data is critical. “The big difference between schools here and in America is that they have a lot of data as a result of government programmes,” Marshall points out. “They had struggled to use that data previously but are now able to understand it better thanks to AI.”
Care must be taken to ensure this data is not used inappropriately, he warns. “There is always a danger it could be used for league tables and so on. We need to manage it appropriately.”
Another very practical use for the technology is automated assessment of oral examinations. “A student can speak into a recording device which converts it to text. A machine learning programme analyses it and scores it.”
Microsoft is currently comparing the machine performance with that of human examiners. “We have got to 86 per cent and we are looking to get to 96 per cent. It will be great to be able to use this technology for practice tests for oral examinations in the Leaving Cert.”
Soapbox Labs
Accurate speech recognition is critically important for such applications and Irish company Soapbox Labs is at the forefront of developments in that area. “We develop speech recognition technology for children,” says Patricia Scanlon, chief executive and founder. “We build the technology and licence it to other people who integrate it into educational tools, toys, games and so on.”
She has been researching speech recognition for the past 20 years, working at UCD, Colombia University, Bell Labs, and IBM. “I always knew that voice technology was going to be huge. The key is speech recognition and there is a big difference between an adult and child. A child under the age of 12 has very different speech patterns and behaviours to adults, yet all speech technology was designed for adults.”
That’s why many of the popular voice activated digital assistants get quite a lot wrong with children and the younger the child the more they get wrong, she said. Scanlon founded Soapbox Labs to solve this problem six years ago and has been working on it since.
“Education was the first use case we were interested in,” she says. “The technology has to be accurate. Off-the-shelf speech recognition technology, particularly when it has been built for adults is not sufficient to teach children to read. If we think about how children learn, an adult reads to them, then they read to the adult who listens, prompts and encourages them. You can’t use voice technology to teach someone how to read if it’s not highly accurate. If the system can’t listen accurately to how they pronounce words, sentences and paragraphs, you can’t use it.
“Reading is a foundational skill,” she adds. “All education starts with reading. If a child can’t read by the age of eight it affects their outcomes. At that age you stop learning to read and start reading to learn. Our technology helps to rapidly advance a child’s literacy skills in a scalable, cost effective way.”
You need to look at the ethics of what you are doing
She sees the technology being used as a teaching aid as well as a standalone learning solution. “There are going to be situations where a child doesn’t have access to a teacher, or due to limited resources, a teacher has limited time to spend with each child reading and assessing,” Scanlon notes. “Using our technology, a child can learn to read and advance their skills anywhere if they have access to a smartphone or tablet. We are making sure the technology works in the real world, in normal environments. We have a number of clients around the world and we will announce new product roll-outs in Asia and the US in the coming months.”
Marshall sees the use of AI in education continuing to grow. “There are a number of practical things we are looking at,” he says. “We are looking at how to help kids in the classroom. Assessment is a really interesting area. If you can get a test score instantly instead of in three weeks time that’s really useful. Kids can see where they are going wrong and get better or see how well they are doing. AI can be used in all sorts of ways as a teaching aid in the classroom. We are actively working with the Science Foundation Ireland Adapt and Insight centres to build new solutions.”
But again, he warns that the technology must be handled with care. “It’s all built on data and lots of it and AI sits on top of that,” he says. “You need good solid research underpinning it and you need to look at the ethics of what you are doing. If you do it properly there will be huge benefits for the system, but it’s going to take time to do it properly.”
BRIDGING THE GAP
Tacoma Public Schools, Washington, USA use machine learning and AI to:
- predict student success with 90 per cent accuracy, reducing student dropout rates through early assessment and intervention
- enable educators to help students improve their academic results, leading to the rise of graduation rates from 55 per cent to over 80 per cent
The Catholic Education of Western Australia uses the cloud to:
- deliver AI-powered personalised education to 78,000 students in 163 schools
It also uses Microsoft Power BI’s predictive analytical abilities to identify important patterns from collected data, which helps teachers identify at risk students and provide interventions as early as possible.
