For centuries seismologists convinced us that it is impossible to predict earthquakes. Like when they said in the 1600s that it was impossible to predict if it would be raining or sunny the next day.
Today technology is revealing to us that in the not too distant future artificial intelligence will let us know earlier when a violent earthquake will arrive.
This is incredible news, it translates into a lot of human lives saved from the disasters of earthquakes and tsunamis.
A bit like when we are able to predict the strength of the seas or the arrival of a storm of water or snow. The ships stay in ports and do not sail. Planes don’t fly, and people avoid walking into the mountains.
In short, knowing how mother nature will behave gives us a lot of advantages and lets us keep danger at bay.
To this day persistent rumors of science echo that it was impossible to predict.
How many times have you heard this? I have, a lot of times. Here’s how things are… or rather how they will be in a very short time…
THE CHALLENGE OF THE CENTURY
Among the most particular challenges that could affect the next few years, is the one I am describing, which is one of the most useful for the entire planet.
At the moment, it is not possible to predict with certainty an earthquake. It is only possible to estimate, with a degree of uncertainty, that a certain event or consequence will happen.
In short, very few events …
But two scientists are trying to win this challenge.
Geophysicist Paul Johnson and his team at Los Alamos National Laboratory have developed a tool that could make it possible to predict certain types of earthquakes.
“Like many scientific investigations conducted in recent years, Johnson and his team’s approach is based on the use of artificial intelligence in the form of machine learning.”
What all this means is they basically provide a lot of earthquake data to the AI and then she works on it.
A bit like a very difficult math problem you had to solve in school. It may be difficult for most people.
To get the help you submit it to a knowledgeable professor at the University of Mathematics. Those brainy talents who are very articulate in certain disciplines and remember a lot of numbers by heart and can carry out extremely difficult operations by heart.
They are able to process that problem easier than and give you the solution.
Neural networks and artificial intelligence need large amounts of training data to teach them what to look for.
They are proficient in processing information on topics, they do not know.
THE SLIDING OF TWO PLATES AT THE SAN ANDREAS FAULT
The cycle of earthquakes on most of the faults involves a process that lasts decades. It is called “stick-slip”, which is a fast movement that occurs between two sides of a fault when these sides “disengage” from each other.
Before this sudden shift occurs, not much happens, other than the accumulation of tension that involves little data at the geophysical level.
When the catastrophic earthquake arrives, it’s too late. It is that catastrophic earthquake that provides most of the data, even though it’s not very useful for the purpose of forecasting.
How do you predict it?
“The Los Alamos team believes it takes about ten cycles of earthquake data to train a system that can predict earthquakes.”
Since seismology is a “young science”, at the moment this is nowhere near possible.
If we had invented artificial intelligence earlier we would have predictions available and it could have processed the data collected in the past.
Like the law of relativity had to wait for Einstein himself before being revealed to man. Before the arrival of such a brilliant mind, we would not have known it.
California’s famous and dangerous San Andreas Fault, for example, generates a major earthquake every forty years or so.
But, at the moment, scientists have only twenty years at their disposal, which is half a cycle of data that is detailed enough to be useful for predictive purposes.
ALGORITHMS WORK WELL ON SOME EARTHQUAKES
The Los Alamos team has also tried machine learning on a different seismic activity. The “slow-slip” events.
These are silent earthquakes. Unlike a traditional earthquake which usually lasts a few seconds, this one can last hours or days or even months before it ends.
This is where researchers can collect a lot of data on these phenomena of the Earth. Which become a great training ground for deep neural networks.
“Johnson and his team did studies on the ‘Cascadia subduction zone’, which stretches about a thousand kilometers along the coast of North America, from Vancouver Island in Canada to Northern California.”
Every fourteen months or so there is an earth-moving problem here.
Geophysicists have been gathering information since the early 1990s. This means that there are several complete cycles of data.
They allow the Los Alamos team’s machine learning system to predict what will happen in the coming weeks, with some degree of certainty.
I find it amazing. It is the first case of earthquake prediction, which disproves what other earth scientists, seismologists, and geophysicists were saying with confidence to date.
In fact, the slow-slip project by Johnson and his team tells us that machine learning techniques really work in the presence of certain types of seismic events
It works like this:
They created a numerical simulation, a computer model, which makes the essential elements their own in a physical event.
He trained his machine learning system on a lab-created phenomenon to see if he can actually learn to predict.
It’s now augmenting with data from real experiments.
Combining simulated data and real data is significantly more effective in predicting when an earthquake will occur in the laboratory.
I would say this is astounding and heralds another great opportunity that technology is giving us.
Saving lives, eliminating risks, and the possibility of saving a lot of money spent on the reconstruction and damage that occurs every time an earthquake happens.
Once again, technology exceeds the beliefs of past knowledge.
If you’re interested in how to profit from AI or just looking for a new way to make more money, then our free report on AI money-making strategies is for you. Press the IMAGE above or the button below and find out more.
Explore LUMI: your one-stop for top AI and finance Vertical Video. Stay informed, gain insights, and stay ahead with the latest trends. With LUMI, the future is now.
More From Inspirational