Microsoft's student developer contest imagines the future

With mixed reality, AI, and big data added to the competition this year, which projects made the grade?

The panel of judges on stage at the Imagine Cup finals​

The panel of judges at the Imagine Cup finals

Microsoft's annual student developer contest, the Imagine Cup, is over for another year, after a busy and inspiring set of finals in Seattle. In a massive science fair-like event over three days, 49 teams from all over the world came to the Microsoft campus to show off their projects.

It's always fascinating to watch the Imagine Cup. In the past I've been a judge in the UK heats, and we've attended the finals several times now. By the time the teams reach the finals they've been part of a national competition in the their home countries, then a regional one that brings international teams together for the first time. They've also worked with both academic and industry mentors to refine their ideas and their solutions, using off-the-shelf technologies to build and deliver nearly product-ready applications, devices, and services.

This year's event added special prizes to highlight mixed reality, AI, and big data; themes that Microsoft is addressing across its entire developer relations program. Teams could enter those competitions alongside the main event.

SEE: Sensor'd enterprise: IoT, ML, and big data (ZDNet special report) | Download the report as a PDF (TechRepublic)

Initial judging took place in two rooms, where project teams could demonstrate their solutions and walk judges through the key elements of their applications. Each team got ten minutes to present to the judges, letting them pitch their solution and demonstrate key features. Wandering around the rooms I got to see a wide selection of projects: tools to help reunite refugee families, precision agriculture tools using smartphones, training applications for firefighters that mixed VR and AR, ways of using drones to spot and track wildfire hotspots, smart beehives, ways of using AI to improve interview techniques.

The ideas were fascinating, the execution well handled; all the teams had prototypes and some were almost ready to go into wider trials and even into production (though sadly one team's hardware had been lost by their airline). At times it was hard to remember that the teams had been working on their projects for less than a year while still studying, considering just how far advanced many of them were.

20180724183443798ios.jpg

Announcing the Imagine Cup Awards

Judging for the new Imagine Cup awards was handled separately, with finalists being announced shortly after the initial judging, and then going straight into a presentation session. Three teams were chosen for each award, and each award was judged separately. Presentations were followed by a Q&A session with specialist judges, who asked in-depth questions of the various teams. With three different topics in three different rooms it was hard for observers to see everything, but those I did follow presented well and had spokespeople able to handle the questions.

Day 2 started bright and early, with the announcement of the 15 teams that were going through to the semi-finals. The remaining teams were then given the opportunity to pitch for three Wild Card places in the semi-finals, with 90 seconds to sell their project to everyone else who hadn't made the cut. Some teams were better prepared than others, but all did their best to sell their projects, including one team who, in best Hamilton-style, gave their pitch via hip hop.

The semi-final presentations took the same format as used for the Imagine Cup Awards. Each team presented to a panel of judges, before answering questions. Again, three different rooms were used, with the 18 teams split across three different panels of judges. With just over an hour of presentations, the judges only had a short time to confer and come up with a list of three finalists.

SEE: How to implement AI and machine learning (ZDNet special report) | Download the report as a PDF (TechRepublic)

Announcement of the finalists was paired with announcing the winners of the three additional awards. SochWare from Nepal won the AI Award for a solution that used machine learning to identify plant diseases, and then suggest how a farmer could use basic techniques to either cure or prevent them spreading. The Big Data Award went to India's DrugSafe for a tool that could identify fake drugs, building a database of counterfeit or altered labels that could be used to reduce risks to patients. A US team, Pengram, won the Mixed Reality Award for an application that let engineers remotely assist in a mixed reality workspace using HoloLens to help them work with remote colleagues.

Three teams would go forward to the final the next day, one Canadian, one Greek, and one from Japan, all offering different assistive technologies. The Greek team, iCry2Talk, were using machine learning to identify the sounds a baby makes, using a large data set to determine what a cry means. The app could be used from phone, PC, and from smart speakers. Canadian team smartARM used IoT technologies and a cloud service to develop a smart prosthetic. Using machine learning, a camera in the prosthetic palm identifies objects and uses the identification to pick a grip from a shared library, triggered with a muscle-sensing armband. Japan's Mediated Ear worked with audio machine learning to pick a specific voice out from a crowd and amplify it for a headphone wearer, turning a phone into a tool to focus on specific voices. It only works on pre-trained voices, so won't listen in on random conversations.

smartARM's winning machine learning prosthetic hand​

smartARM's winning machine-learning prosthetic hand

The final took place in a hotel in Seattle, where three industry judges would choose the winning team: Erica Brescia, co-founder and COO of Bitnami, Anil Dash, CEO of Fog Creek and Glitch, and Peggy Johnson, Microsoft EVP of Business Development. Taking the same form as the heats, the three teams each got to give a short presentation and demonstration, followed by questions from the judges. It wasn't surprising that the finalists got a lot of questions about the future of their projects, as they were all solutions that showed a lot of real-word promise.

After a short deliberation, the judges returned with results. The 2018 Imagine Cup winners were smartARM, their plans for ever-smarter prosthetics -- matched with replaceable 3D printed parts -- resonating with the panel. It was certainly an impressive project, and one that caught my attention while walking round the initial showcase. A video they'd made of a colleague trying out a prototype of their prosthetic showed just how much impact the project would have on its intended audience, with a look of astounded joy as she made it grip for the first time.

Seeing so many projects with social impact was inspiring. Technology can do much to build a better world, and it's good to see the next generation of developers taking full advantage of that power.

More for developers

Python's rise: Could it soon edge out C++ in programming language popularity?

Python climbs up TIOBE's search engine-based index of programming language popularity.

Google AI on Raspberry Pi: Now you get official TensorFlow support

Google's TensorFlow team makes it a whole lot easier to get AI up and running on a Raspberry Pi.

Gmail app developers' employees might be reading your email (CNET)

The app makers defend the practice, but also say they've since stopped it.

Windows apps made on Linux hit by security fail

That Windows app you made on a Linux system actually isn't as protected from attacks as you thought.

Hiring kit: Android developer (Tech Pro Research)

Companies are increasingly dependent on mobile platforms to power their business operations and to enable a productive workforce--and that means hiring topnotch developers to build the apps they need.