Empowering the Next Generation of AI Technicians (Part1)
A University of Tokyo Graduate School Alumna, she founded “Naked Technology” as a student. She has developed and managed a UI middleware for iOS, Android, and “Garakei” phones. After selling this company to mixi in 2011, she founded Cinnamon in Singapore in 2012, and later expanded into Vietnam and Thailand. Aiming to be the number 1 visual private communications service in Asia, she has delved into more Artificial Intelligence (AI) businesses, including establishing and managing the video app Tuya.
In this first interview, she opens up about her background and visions for AI powered Businesses.
Miss Hirano, you are currently based in Japan and starting a new business related to Artificial Intelligence. Can you tell us about this business?
Miku Hirano (a.k.a Hirano) : At the moment, I am working on developing a service that covers everything from consulting to systems development and caters specifically to businesses that have yet to make use of AI Technology. Currently in Japan, AI is not fully understood in many existing businesses as it is often discussed in academic contexts such as “AI has surpassed the human brain in Chess and Shogi” or “how can AI be made use of in self-driving cars”. Of course, we have seen AI at work in some businesses for over a decade now, such as Amazon’s recommendation function. However, if AI were to be fully implemented in the medical industry, human resources industry, law and so on, we would see a drastic change towards efficient problem solving in each industry. Of course, AI is not yet capable of physical tasks such as drawing pictures or cooking, however it does help narrow down what to prioritize in various tasks and in turn makes processes more efficient.
AI is instrumental in refining targets and goals. Specifically, what kinds of industries can benefit from AI?
Hirano: As an example, it can improve efficiency and accuracy in recruitment. Let’s say, for example, that human labor results in only 10 applicants being matched with companies per day. However, with AI, once it has studied the matching patterns and there is plenty of data on the matching process, thousands if not tens of thousands of matches can take place per day.
The medical industry is another that can also accomplish dramatic change with the introduction of AI. Firstly, it is not possible for a single doctor to be an expert in all medical fields. However, if we can program AI to capture cases and medical science that span a wide range of fields, then not only will this alleviate the pressure on doctors, it will also lead to an increase in diagnosis accuracy. Moreover, being able to use the objective AI analysis results as evidence as to whether a patient has cancer or not can dramatically reduce the psychological burden on doctors who are at risk of lawsuits in the case of a false diagnosis.
There are many potential uses for AI across various industries, aren’t there. Did you originally have an interest in AI?
Hirano: Actually, it started when I incorporated AI services such as recommendation functions into the company I founded, Naked Technology. I was involved in some AI, and my CTO was specialized in it. At the time, I had to change the service due to complications, but compared to 10 years ago, the AI scene has dramatically changed in several key ways. In order to incorporate AI into businesses, you need “input”, “process”, and “output”. For “input”, a large data set is necessary, and with smartphones as prevalent as they are, retrieving data on specific topics, for example “what words are most frequently used by women in their 20s on Twitter”, is incredibly easy. For “process”, processing speed has become unbelievably fast with recent innovations. Lastly “output”, or the analysis results of AI, is much more accessible by the general public due to the prevalence of smartphones, tablets, and IoT. With essentially all 3 conditions currently being met, I believe it is the perfect time to return to business focused AI.
What was the deciding factor that got you involved in this business?
Hirano: The deciding factor was when I expanded into Vietnam and I realized how many talented people there were who could put AI to good use in businesses. When I recruited for AI scientists and engineers in Vietnam, 150 applications came flooding in. These applicants were the equivalent of top of the class students from University of Tokyo or Tokyo Institute of Technology. Partly due to its national policies, Vietnam is in fact a treasure trove of talented engineers.
Wow 150! That is very different from Japan.
Hirano: When looking at Japan, there is a shortage of people with that skill set. Currently there are only about 300 people in the country researching AI. With most of the population employed in electrical and automobile companies, the number of people who can implement AI technology into businesses is extremely limited, which is an imminent problem that needs to be solved. At Cinnamon, our objective is to train people in AI development so that AI is more effectively implemented into businesses. So far, we have employed 15 engineers from the applicants in Vietnam and are training them as business AI specialists to promote growth in this area.
When you are training business specialists, what qualities do you believe are most important?
Hirano: Not only do we want them to be well versed in AI, mathematical training, and programming, but we also want them to possess entrepreneurship. The reason for this is that to get to a level where we are implementing AI in a streamlined and practical way, there will be a lot of trial and error. There are numerous existing algorithms, but we can’t know whether they will work or not in say, medical diagnostic imaging, until it is tried. Failing is inevitable in the process of testing and improving patterns, so it is important that they possess the spirit to try over and over again without fear of failure. Moreover, the ability to communicate problems that do occur during system development to fellow team members is also extremely crucial. For that reason, individuals who are caught up in strict research paper rules and pursue only the academic side of AI technology, will have difficulty adapting to the entrepreneur spirit necessary for Start Ups. Consequently, in order to train AI business specialists, there is a divide between the academic and the necessary skills.
Continues in Part 2.