Copyright: nexusplexus / 123RF Stock Photo
What's up with the artificial intelligence scene in Singapore? Singapore City AI's latest gathering, held at Unilever's Level3 coworking space last week, attempted to answer that question by offering a range of the more up-and-coming AI companies that call Singapore a hub. Fittingly called “AI Use Cases from the Trenches," it focused specifically on personal stories, context from experience, and generally staying away from the hype that has been surrounding topics like AI and blockchain all year. The event was sponsored by City AI, Padang & Co, Eventnook, and 100offer.
Basically, in an atmosphere where walking the walk is way more important than talking the talk, Innovator.SG founder and Singapore City AI lead Victor Alexiev encouraged those there to talk the walk, presenting use cases that gave the audience a brief but thorough look into some of the problems and solutions AI is solving - as well as some of the tech and learning methods used to reach those solutions.
“How many of you have heard of Evie?" Victor asked the room.
Hands go up. “I'm using it!" one voice called out.
Evie is an artificial intelligence that tackles scheduling - you can speak to her like a person, and she'll be able to figure out if your meeting can be on a Tuesday or a Thursday. She can take into account schedules, time zones, international holidays, and other blocked-off times and organize someone's schedule to where it's most convenient for them.
Evie was designed four years ago as a bit of an anti-Siri, co-founder Jin Hian Lee explained. Where Siri started out by trying to do several things as more of an AI personal assistant, Evie honed in on just scheduling.
Think about it this way: you spend your whole day working on various tasks, and every time you have to change between them, you waste a little bit of time. You waste even more of your resources when your business is international and you need to coordinate between multiple people. Evie tackles that one problem and can communicate with people through messaging, much like a secretary.
“The biggest problem we solve is productivity," Lee said. At his previous job with Yahoo, he was working as a product lead and found that they lost lots of time to just trying to get the same people in the same room at the same time.
Evie came out of decomposing a problem, the solution of which was providing a service solution. The challenges that come with running Evie are less due to crunching a lot of data and more having to deal with scalability.
IBM Watson Cognitive Solutions
An easy perception to make of artificial intelligence is one that is black and white - either humans will do a task or AI will do a task. Manprit Singh, a consultant for IBM Watson Cognitive Solutions, pointed out that there's another area in here for AI and humans to work together in a way that best utilizes both of their strengths. In order to reach this point, IBM has gone through a few different developments to be able to work with the same uncertainty that humans face every day.
For example, Manprit said that AI won't replace your doctor “anytime soon," but having an AI assistant who can take a lot of different data points and weigh possible options for a diagnosis can result in a doctor who spends less time doing research and more time with his or her patients.
We're still several years off from getting our own holographic doctors like the one on Star Trek Voyager, but IBM's got a hell of an assistant in the pipeline.
The ideal AI for IBM is practical, capable, adaptable, and tailored to its location. Data gathered in California may not be applicable to other locations in the world. Segregating the AI to a certain area can keep insights relevant.
Watson is also trained for accuracy, meaning that, when giving an answer to a question - like a medical one, for example - it can also give a confidence rating, a measure of how confident it is that its answer is right. That in turn can help humans with their decision making as well.
Cambridge-based Prowler.io is a Ph.D-heavy (50 percent of the staff hold Ph.D's in machine learning) company that holds that the future of AI is not in deep learning but rather in principled decision making. It got its start just last year but has already landed funding from organizations including Amadeus Capital Partners, Passion Capital, and SGInnovate.
What Prowler has tried to do is break away from “decision tree-based" decision making because that does not account for uncertainty very well.
Prowler has focused on building up a system that learns autonomously, incorporates uncertainty in order to more quickly adapt to its environment, evolves those strategies in a complex environment, and can elaborate on its insights. It gets these qualities from a combination of mathematic principles, machine learning, and machine intelligence, said Zach Tan, the company's VP of product and business strategy.
Prowler uses its insights to drive game design, transportation, and logistics.
What Bitcoin was invented to do with money and what Ethereum does with businesses, Ocean Protocol does with data.
For example, Bitcoin enables money to be transferred via the Bitcoin currency between people, no matter where they are, without having to go through a bank or another payment processing system. Ocean Protocol will allow people to access, search for, and purchase data from wherever they are. This is important, because it will encourage greater collaboration over data (companies will be able to access each others' data - no one holds all of the data in the world). This, in turn, will decentralize data for the masses, encouraging companies to give up their hold on data so that everyone can access it.
And, because knowledge is power, data is power.
Ocean also works with securing the data to protect the rights of those who own the data. This means that it is taking steps to make sure that another Equifax breach - where some 175 million people's financial information was suddenly out in the open for the taking - wouldn't happen under Ocean.
“What we really need is to give power back to the people," said Chirdeep Singh, Ocean Protocol's founder.