The typhoon brought abundant rainfall. This was a blessing for Taiwan as there won’t be a shortage of water in the summer. However, stabilizing the water supply has always been a challenge for Taiwan in recent years.
🚰 30 years ago, plastic pipes were used to popularize tap water. Due to the cost, the plastic pipes cannot be completely replaced, resulting in a leakage rate of up to 15%. Nearly 1 billion tons of water are lost each year due to the pipeline.
👂 This is a big problem. Is there any solution? Experts are dealing with the leakage issue. The Taiwan Water Corporation now has 80 workers who carry a leakage detector that can amplify the sound of leakage 900 times. They “listen” back and forth in the sewer, some of them having 30 years of experience. However, it would take 80 of them three years to circle around Taiwan once. Even without rest, it would be very difficult to spot new leaks on time.
Resolving this hot issue has been given hope with recent AI technology.
📈 When I first saw the proposal of the “Water Saviours” team in the Presidential Hackathon, I found they had adopted the “supervised learning” technique, that is, having the workers tell the machine what sound indicates water leakage. Then the “deep learning” algorithm is used to predict and verify it. By the time the machine apprentice builds the model, it will save a lot of time and effort.
🕚 Taiwan is not the first to have invented such a system. There are also solutions developed and used by Israeli companies, but at a considerable cost. The Taiwan system, written with open source code, has reached an accuracy rate of 70%, and the time taken to spot the leakage has also been reduced to 1/10 of that in the past.
🇳🇿 One of the benefits of the self-developed technology is that it could later be used in other fields, such as smart meters or natural gas pipelines, as they all have pressure and velocity of flow. The “Water Saviours” team, representing Taiwan, will participate in an accelerated program next month organized by the New Zealand government to help solve the problem of water leakage there.
What’s interesting about this case is that there has always been a voice saying, “AI is here and human jobs will be replaced.”
👴 For example, will the 80 water leakage workers lose their jobs when AI comes? Exactly the opposite! The process of machine learning will continue to require human wisdom to “pour” in, and constantly require people and machines to learn from each other in order to solve the problem one by one.
💡 The “Water Saviours” team brings the wisdom of the skilled masters, solves the problem with collaboration, and finds a new position for the value of human wisdom. This is what AI should stand for —“Assistive Intelligences”.