With our deep learning, we'll see an object in front of us. We need to tell whether she is a human, a pile of bricks, or a chair. And even if you tell it is a human, we must think, 'What isa human going to do next?' Are you going to turn left? Are you gonna turn right? Going to jump in front of us? [Robot] Hi, so sorry,but you're in my way. Could you please move? LionsBot, we makeprofessional cleaning robots so the cleaners don'thave to work so hard.
Singapore is the Lion City, so the lion is the emblem of Singapore. Hence, our robots are LionsBot, and at the heart of every robot, there is one grain of sand from Singapore, and it brings the love andthe technology of Singapore to the rest of the world. We have multiple sensors,each feeding in information multiple times a second. In robots, anyone canput in a lot of sonars, a lot of sensors, butit is how we use them, how we make intelligent decisions with that information that counts. [Laurence Liew] Where you are is ourAI Singapore office. Singapore has a long historyof willing to spend money to get its citizens tore-skill, deep-skill, or upgrade their skills.
Our mission, really, isto promote the use of AI, get more researchers toembark on a career in AI, to do AI research. We have one very popular program. We call the AI for everyone, and the intent is to demystifyAI for the man in the street, for everyone in that sense. When the audience walkedout of the auditorium, they say, 'Ah, OK, AI is not so scary.' AI is actually nothing morethan just another piece of code, obviously very sophisticated code, but it is just another ITsystem or infrastructure. [Annabelle Kwok] Hi, I'm Annabelle. I founded NeuralBay, whichis a software AI company that looks into imageand video processing. So I was very lucky to be in Singapore where the hackathon scenewas slowly starting, and it was still kindof ahead of its time. So when this whole field ofimage processing came up, I think that opened a lotof doors for opportunities to not just analyze still photos but also to look at real-life events. So for example, in traffic flowmanagement in crowded areas, you can help to better direct human traffic. So we're in our office, and we have a lot of people walking around. So what we can do with thissoftware is that we can count the number of people in this area, as well as to track their movements.
So in recognizing people,it's a very tough problem because when they look away, can you still recognizethat it's the same person? So Zeldon, if you can justturn around very gracefully. So you can seethat in this software, it still tracks thatZeldon is the same person. [Laurence] I think when we design AI systems or any smart city technology, ultimately the question to ask is how will it affect thecitizen in the country? We do have several healthcarerelated AI projects that are undergoing,and I think there's lots of interesting areas where AIcould be used in education. When we launched AI for Everyone, the original targetwas 10,000 Singaporeans to be trained in three years. 1-1/2 years down the road,we are already at 7,000. I told my team can we do 100,000? Let's go from 10 to 100, all right? Training the people, the apprentice, they again, at eight or nine months, they will go out to the industry. There is an economic implication in that.
Singapore has a widepool of talented engineers. The government has spenta lot of money developing and training these engineers, so with such a big latency pool of people that we can tap on, whynot build in Singapore? [Annabelle] In terms of the software, I think the next step wouldbe accessibility to more data and also the diversity of data available. Most of the open-sourcedata is from the West and not necessarily from Southeast Asia because Southeast Asiancountries may not necessarily have the infrastructureto capture that data. So recognizing a woman ofcolor within Southeast Asia, the confidence intervalmight not be as high as recognizing one from the West. By doing it and making it available for the small enterprises,hopefully that might help correct some of the cultural bias in technology. We don't always have to give back in terms of time and money, but we can also give back interms of knowledge and skills. As for myself, I'm goodat building things, so why not build things to help people?
Comments
Post a Comment