kiara_logo_icon_red/w=32,quality=90,fit=scale-down){kind=icon}
Goal of building a 1 million A.I. community
日本語ブログはこちら→ 🇯🇵日本語
MISSION IMPOSSIBLE
Team AI is challenging Mission Impossible.
It is about achieving a true AI society through the power of Team AI.
As a numerical target, we are trying to create a machine learning community of 1 million people mainly in Tokyo.
At a time when there are said to be several 100,000 application engineers, I think it’s really hard to bring the nascent AI engineering community to 1 million.
But Team AI is serious. The reason is that we believe in the truth that it is valuable to mankind, and that what is valuable will always be mainstream.Team AI now has 2000 community members in 1 year. This 500 times is 1 million people. I don’t think this is impossible. we have already achieved 0.2% of our target. You can leverage your business model and scale it exponentially.
It’s hard to have 1 million people in direct proportion, but a quadratic function is possible.
It is often said in Silicon Valley that if we calculate backwards from the future, we can see what we need to do now, but when we think of “true AI society” the following things are realized.
Implement will become as easy as HTML
GUI machine learning tools such as SONY/Microsoft Azure are already coming out, but the popularization of “tools that can be used by non-engineers” is indispensable for the democratization of AI. Intuitively, a future where people with no prior knowledge can quickly use machine learning. I think this is the key to the spread of AI. Currently, the know-how is still concentrated in some engineers, and there are few cases of AI implementation.
In HTML, tools like Perich and Strikingly are popular for getting cool homepages in 30 minutes. I think these wonderful services have lowered the threshold largely for small and medium enterprises and individuals.
PlugIns, which extend the functionality of WordPress and are available at a low price in HTML, are also more convenient. Similarly, in AI, if complex mathematical models are modularized and fitted into the calculation flow in a simplified state, it will become more convenient if tools that can build and learn AI become widespread. This is simply not possible, but I think it will be solved in three to five years.
In addition, Silicon Valley’s enterprise machine learning open source H2O.ai, which recently had a guest event with Team AI, has already released a “Machine Learning for Machine Learning” called Driverless AI. It’s a cool way to simulate a variety of mathematical models on a dashboard without writing any code, and then optimize the actual data analysis “automatically pull in” into a more accurate model. In addition, there are actually many unsophisticated tasks involved in data analysis, but as the ideas of “Automating Data Preprocessing Open Source” which our company intern Mori is working on, become more familiar in the future, AI will become even more familiar.
AI education will become as familiar as convenience stores
There are 50,000 convenience stores in all over Japan. I think it would be nice if there were 50,000 machine learning education centers like Team AI Base all over Japan. I think there should be more opportunities for people to learn AI easily, such as community centers, cram schools and libraries in the city, offline.
AI is one of the most lucrative jobs today. Although personnel costs have not become substantial, there is an overwhelming shortage of human resources, and it is not difficult for young people with skills to obtain job offers from 5 to 10 companies. You can become a freelancer and live comfortably. Senior workers, who had no job opportunities at Hello Work, were encouraged to learn mathematics and AI, and they could become very attractive. Team AI wants to create more places to learn these amazing skills.
I think it is necessary to lower the hurdle significantly in order to make it familiar. I hope that Team AI will realize the concept of “AI School for Kids” and encourage elementary, junior high and high school students to learn more about AI and drive the future. At the same time, building educational content that can be understood by children will eliminate “Inaccessible AI” and can be used for AI learning by humanities professionals. Senior citizens and women can understand it. In this sense, I think AI education for children is very important.
Collaborate with engineers from around the world
Last year, Google Translate introduced a neural network to improve the accuracy of its translations. Thanks to this, by installing a plug-in called Google Translation for Chrome on the Chrome browser, I can read any website in Japanese with one click and without stress.
I think that this practically eliminates the English barrier that Japanese people are not good at. With a similar translation API, messages are translated in real time on Slack and LINE, making it easy for anyone to chat.
Our company’s use of Kaggle at hackathons is a great open innovation platform. The Python/R code, raw data, and analysis results posted by AI engineers around the world are available for free, and you can use a function called Kernel Fork to run it in your browser instantly without building an environment. Cloud computing resources are also free.
I think Kaggle is the ultimate platform for data analytics people. For example, if the diagnostic data of people suffering from incurable diseases in our company is published on Kaggle, we could work together as a hackathon problem in Shibuya, build diagnostic algorithms from Japan with the help of the community, and actually save African lives. Our efforts have only just begun, but I think we have a great contribution to society.
GitHub, StackOverFlow, Qiita, and open source all use the collective knowledge of engineers from around the world for innovation is a common standard, but so is data analytics.
Academia has been publishing machine learning research papers on the Web, simultaneously releasing GitHub source code of research results published by MIT and Stanford, and encouraging research institutes and hackers in various countries to study products and peripheral approaches.I think “An ecosystem of wonderful know-how sharing on a global scale (recycling mechanism)” is the future of AI.
Finally
- Make implementation as easy as HTML
- AI education will become as familiar as convenience stores
- Collaborate with engineers from all over the world
This is the AI society that Team AI should realize as its mission.
As a first step, they plan to introduce a free education platform called “Team AI Zero”
First of all, we will create a curated media of high-quality educational content, and gradually expand quizzes and problem collections.
We will create a knowledge base using Qiita, Youtube, Udemy, free e-book, O’Reilly and Google Translate to accelerate your learning and career support.
Our goal is to build a 1 million person machine learning community, with a success rate of 2000 = 0.2%.
Increase the white space by 99.8% and grow 500 times.
Thank you for your continued support.
＞＞ Read more from Kiara Blog ＜＜