The newsDesk speaks to Matt O’Connor and Dhruv Sahi, the brains behind a machine-learning education group, Reboot, about the importance of democratising information and learning about AI – and the dangers of Hong Kong falling behind in machine learning knowledge and talent
As we scroll through Instagram, use GPS services to locate our next appointments and message friends on third-party platforms, we enjoy a connected world of convenience that opens up to us through our fingertips and devices. This increasing technological convenience makes our lives run smoother, faster and more efficiently – but the complex mechanisms that power it remains a mystery to most. As the gap between Silicon Valley giants such as Google and Facebook and the rest of the world grows, so do the dangers of growing inequalities in wealth and technological know-how, putting the jobs of hundreds and millions at risk.
We speak to the brains behind Reboot.ai, a machine-learning education group that hopes to bridge the knowledge gap between Hong Kong’s existing talent and the demand for machine learning expertise. In other words, Reboot seeks to democratise machine learning and AI so the people who own machines are not the only ones to reap the benefits of this technology. The two behind Reboot, 29-year-old Matt O’Connor and 25-year-old Dhruv Sahi, both data scientists, are holding a workshop at theDesk coworking and events space in Sai Wan on the basics of intelligent algorithms on August 17th, between 7pm-8.30pm.
Okay. Full disclosure. When phrases like ‘AI’ and ‘machine learning’ get thrown around, some of us here at theDesk can’t even pretend to know what it all means. So O’Connor and Sahi patiently unpack these terms for us. Machine learning, explains Sahi, is the process of ‘giving a machine data so they can run from it and make better decisions.’ O’Connor furthers explains that, “traditional programming is where you have to tell the computer every single thing. Whereas machine learning and AI is where you create a more abstract infrastructure through machines can start learning if you reward it every time it does something right.”
In other words, it’s like ‘training a dog’, says O’Connor. A dog tries everything to receive a biscuit, and when it does something right and receives his treat, he assigns greater ‘probabilistic weight’ to that behaviour (i.e. sees that that particular behaviour is more likely to yield a reward) and so is more likely to repeat it again in the future. Only in our case, the dog is an algorithm used to achieve a specific outcome.
““AI and automation is going to lead to two paths. Either people who own the machines, like Google or Facebook, will reap the benefits and capture all the wealth. The alternate path is where machines help better people’s lives and not replace them.”
Now that we understand a little bit better what O’Connor and Sahi are talking about, it’s easier to see how quasi-apocalyptic assertions like ‘the machines are going to take over!’ or ‘Google will take over the world!’ perhaps aren’t just hyperboles to scoff at. The implications become clearer – why would you hire someone to organise and analyse spreadsheets when you can devise an algorithm to do it for free, ad infinitum? “AI and automation is going to lead to two paths,” says O’Connor. “Either people who own the machines, like Google or Facebook, will reap the benefits and capture all the wealth. The alternate path is where benefits are equally distributed, where machines help better people’s lives and not replace them.”
Through Reboot, the two hope to ‘make sure AI is more democratised’ and ensure that the mastery over these growing technologies aren’t reserved for just a handful of technological elites. “People are gonna need hard skills in the future,” asserts Sahi, whose personal motivation behind helming Reboot is also so he can gain more knowledge himself. The duo began in Hong Kong because “there are companies here that have tons of data but can be making better use of it, and the level of talent is not the same here as in San Fran,’ says O’Connor, adding that, ‘we’re trying to level the playing field a little bit.’
Most of their students are professionals taking their course part-time. They’re ‘people who want to prepare themselves now’ or ‘others who are already working in relevant fields, such as finance or data science, and want to perform better in their existing roles,’ they tell us. Through equipping these professionals with three fundamental abilities – first, how to automate data, gathering and cleaning it, second, how to carry out analysis, and finally how to visualise your model and explain it to someone else – the two hope to teach others how to make the most of the data at your fingertips, and explain the practical business implications of your discovery to even the least tech-savvy of stakeholders.
The duo are motivated by their love of the process of analysis, modelling and problem-solving, as well as desire to provide vital tech skills in a more personalised way than present curriculums which are ‘too academic’. “We’re designing a curriculum with the local atmosphere in mind,” says O’Connor, “talking to data-driven companies and asking about their needs.” On top of that, “it’s a very good feeling,” Sahi tells us. Continuing, O’Connor says, “it’s satisfying to not only provide someone with the answer, but to show people how to arrive at the answer themselves. This has an exponential effect.”
O’Connor and Sahi, graduates of Emory University and Connecticut College respectively, have full-time jobs outside of Reboot. O’Connor works at global supply chain manager Li & Fung, while Sahi works with online clothing retailer Grana. With firsthand experience in how AI and automation are impacting their respective fields, the two dedicate their free time to Reboot, ensuring that they do their part in democratising this growing technology. “Information is a form of wealth.” asserts O’Connor. “We’re not just talking about data, but macroeconomics.” And by helping democratize how we’re able to see and use data, Reboot is one small step in ensuring that human beings are lifted up by technology, rather than trampled beneath it.
Matt O’Connor & Dhruv Sahi, in short:
Names: Matt O’Connor & Dhruv Sahi
Ages: 29 & 25, respectively
Location: Hong Kong
To learn more about Reboot, come to Coding AI: Building Intelligent Game Playing Algorithms with Python on August 17, 7-8.30pm, and meet Matt O’Connor and Dhruv Sahi. Reboot is now accepting applications for their next semester, which begins September 12th 2017. Details can be found on their website http://reboot.ai.