Hello World – by Paul Gallanagh
Ever wondered how the Netflix ‘watch next’ manages to hook you in for an unplanned weekend of binge viewing? Or how Spotify’s ‘recommended for you’ song becomes the perfect soundtrack for your day? This is thanks largely to machine learning – a term I was completely unaware of when Apps for Good first asked if I would deliver their new course on the subject.
Stanford University defines machine learning as, “the science of getting computers to act without being explicitly programmed”. In recent years, it has brought us self-driving cars, practical speech recognition, effective web search, and a vastly improved understanding of the human genome.
Indeed, chances are that already today you have encountered machine learning going about your daily routine, and the impact of machine learning is going to grow exponentially, with many organisations seeking to maximise its potential.
Although we had reservations about our capacity and capability to successfully deliver a machine learning course to our students, we were keen to give it a go. We knew that that Apps for Good would support us as we progressed through the course, thanks to our experience of piloting their Internet of Things course a few years back.
The fear of this unknown was tempered with our desire to keep our computer science provision real and relevant to our students – and Apps for Good has really helped us keep our provision fresh, exciting, and current.
We were really impressed by the high-quality resources supplied by Apps for Good. Well researched, clear and concise, and accessible, these helped staff to enrich their own knowledge and skills in machine learning. Alongside these, we also used the first class, practical tasks offered by Dale Lane via his excellent Machine Learning for Kids site (machinelearningforkids.co.uk).
Having familiarised themselves with the resources, staff were equipped to successfully deliver this rich, experiential course to our S3 pupils (11- to 12-year-olds) who had chosen to study computer science.
It was no surprise that our pupils immersed themselves fully in this experience from the get-go. The convergence of currency of context, quality of resources, creativity, teamworking, and problem-solving proved to be a potent cocktail that brought the very best out of our young people.
As my colleague Darren Boyd, a computing teacher here at Dunoon Grammar School, commented, “I ‘freaked’ a little when Paul first asked us to deliver this new machine learning course. However, due to the quality of resources provided by Apps for Good, I was able to upskill myself and deliver high-quality, engaging lessons which the pupils thoroughly enjoyed. Pupils enjoyed it so much that they took it upon themselves to learn key programming concepts beyond the level we would expect at this stage of their learning. This was in order to create their machine learning prototypes, which is extremely exciting to witness for their development within the subject”.
The culmination of the Apps for Good year are the regional and national ‘market place’ events, where pupils have the opportunity to pitch their concepts and prototypes to industry experts.
To help prepare for these Apps for Good events, we hosted our own, inviting local industry experts to visit our school so our pupils could share their ideas with them. Over 50 teams shared their machine learning and Internet of Things concepts during a very exciting day. It was extremely pleasing to see the diverse range of ideas on show, with pupils successfully drawing on the skills and knowledge acquired on the course to produce a range of commercially viable products, with supporting prototypes.
Our pupils’ projects spanned topics from natural language processing to recommendation systems and decision-making systems. These included ‘Paws and Relax’ – a system that recommends a stray pet currently in a rescue centre to suit an owners’ needs and lifestyle, and ‘Mood Master’, a system that evaluates your mood based on the language used and tone of your voice when speaking on your mobile. This would message the user with recommended self-help support and guidance.
We were then delighted to hear that one of our machine learning teams had scooped the People’s Choice award at the Apps for Good Scotland event, as voted for by the industry experts in attendance. ‘Carbon Kicker’ is designed to evaluate your current impact on the environment and recommend ways to reduce this negative impact.
All teams received invaluable feedback during the event and this advice was supplemented by further support given by staff from the software engineering firm EPAM during a recent Google Hangout expert session. All 50 of our teams are currently finalising their entries for the Apps for Good UK awards, which will be held later this year in London.
Making the impossible possible
It is no understatement when I say that Apps for Good have revolutionised our provision of computer science over the last few years. We have record numbers of pupils continuing their computer science journey with us in certificated courses in our senior school next year, and increasing numbers are continuing this journey into tertiary education and industry.
This impact is perhaps best summed up by one of our current pupils, Olivia Robertson:
“I had not heard of machine learning, but once it was explained to us, it was clear that machine learning was already part of my life and I had already been encountering it through social media and services such as Netflix.
“I would never have believed I would have been able to code a machine learning solution, and it was really satisfying to produce our own prototype and get our code working effectively”.
Paul is Principal Teacher of Business & Computing at Dunoon Grammar School, Argyll and Bute, Scotland (@dunoongrammar1). He was named Apps for Good UK Educator of the Year 2018.
The original article can be found on page 70 here.