scishow
Detecting Earthquakes: AI vs. Citizen Scientists
YouTube: | https://youtube.com/watch?v=tfsrOLnKTG4 |
Previous: | These Superpowered Animals Use Your 5 Senses, But Better |
Next: | Catching Alzheimer's 25 Years Earlier |
Categories
Statistics
View count: | 106,716 |
Likes: | 5,679 |
Comments: | 237 |
Duration: | 05:01 |
Uploaded: | 2021-09-27 |
Last sync: | 2024-12-06 10:15 |
Citation
Citation formatting is not guaranteed to be accurate. | |
MLA Full: | "Detecting Earthquakes: AI vs. Citizen Scientists." YouTube, uploaded by SciShow, 27 September 2021, www.youtube.com/watch?v=tfsrOLnKTG4. |
MLA Inline: | (SciShow, 2021) |
APA Full: | SciShow. (2021, September 27). Detecting Earthquakes: AI vs. Citizen Scientists [Video]. YouTube. https://youtube.com/watch?v=tfsrOLnKTG4 |
APA Inline: | (SciShow, 2021) |
Chicago Full: |
SciShow, "Detecting Earthquakes: AI vs. Citizen Scientists.", September 27, 2021, YouTube, 05:01, https://youtube.com/watch?v=tfsrOLnKTG4. |
There are over 13,000 active seismic stations out there, producing far more data than seismologists have time to go through. So, researchers set up a showdown of humans versus machines to sift through all this information and, in the process, crown the heavyweight earthquake detection champion.
The project we described in this video has ended, but many scientists recruit for citizen science projects over at https://www.zooniverse.org/.
There’s also the Did You Feel It project, where anyone who feels a rumble can fill out a quick survey and report it to the US Geological Survey: https://earthquake.usgs.gov/data/dyfi/.
SciShow is supported by Brilliant.org. Go to https://Brilliant.org/SciShow to get 20% off of an annual Premium subscription.
Hosted by: Hank Green
SciShow is on TikTok! Check us out at https://www.tiktok.com/@scishow
----------
Support SciShow by becoming a patron on Patreon: https://www.patreon.com/scishow
----------
Huge thanks go to the following Patreon supporters for helping us keep SciShow free for everyone forever:
Chris Peters, Matt Curls, Kevin Bealer, Jeffrey Mckishen, Jacob, Christopher R Boucher, Nazara, charles george, Christoph Schwanke, Ash, Silas Emrys, Eric Jensen, Adam, Brainard, Piya Shedden, Alex Hackman, James Knight, GrowingViolet, Sam Lutfi, Alisa Sherbow, Jason A Saslow, Dr. Melvin Sanicas, Melida Williams
----------
Looking for SciShow elsewhere on the internet?
SciShow Tangents Podcast: http://www.scishowtangents.org
Facebook: http://www.facebook.com/scishow
Twitter: http://www.twitter.com/scishow
Instagram: http://instagram.com/thescishow
----------
Sources:
https://www.frontiersin.org/articles/10.3389/feart.2020.00321/full
https://www.zooniverse.org/projects/vivitang/earthquake-detective
https://www.essoar.org/doi/10.1002/essoar.10507021.1
http://www.isc.ac.uk/cgi-bin/stations?sta_list=&stnsearch=GLOBAL&stn_ctr_lat=&stn_ctr_lon=&stn_radius=&max_stn_dist_units=deg&stn_bot_lat=&stn_top_lat=&stn_left_lon=&stn_right_lon=&stn_srn=&stn_grn=
https://www.annualreviews.org/doi/abs/10.1146/annurev-earth-060313-054648?journalCode=earth
https://pubs.er.usgs.gov/publication/70192475
https://earthquake.usgs.gov/data/dyfi/background.php
Images:
https://www.istockphoto.com/photo/earthquake-seismogram-recording-by-a-seismograph-image-gm184093354-1828513
https://www.storyblocks.com/video/stock/apartment-earthquake-shaking-bmzzumdrqjowveyvg
https://www.storyblocks.com/video/stock/mesh-earthquake-macslkl
https://www.istockphoto.com/vector/vector-illustration-of-an-abstract-scheme-which-contains-people-icons-gm1247980427-363437930
https://commons.wikimedia.org/wiki/File:JAGI_Waveform_-_19_August_2018.gif
https://www.istockphoto.com/vector/earthquake-wave-seismograph-vector-gm1047684770-280241725
https://www.istockphoto.com/vector/3d-neural-network-with-six-layers-gm1165885651-320957853
https://www.storyblocks.com/video/stock/silhouette-of-people-rejoicing-and-lifting-up-his-hands-a-group-of-successful-businessmen-happy-and-celebrate-the-victory-on-the-roof-of-the-business-center-slow-motion-hhrqhg5lzj2tckekb
https://www.storyblocks.com/video/stock/animation-of-seismic-energy-wave-pattern-with-earthquake-ripple-s77fuidixjnhg69pr
https://www.istockphoto.com/vector/abstract-network-background-particle-wave-blockchain-neural-network-gm1195230552-340619965
https://svs.gsfc.nasa.gov/2893
https://www.istockphoto.com/vector/businesswoman-and-robot-shaking-hands-gm1141452215-305810039
The project we described in this video has ended, but many scientists recruit for citizen science projects over at https://www.zooniverse.org/.
There’s also the Did You Feel It project, where anyone who feels a rumble can fill out a quick survey and report it to the US Geological Survey: https://earthquake.usgs.gov/data/dyfi/.
SciShow is supported by Brilliant.org. Go to https://Brilliant.org/SciShow to get 20% off of an annual Premium subscription.
Hosted by: Hank Green
SciShow is on TikTok! Check us out at https://www.tiktok.com/@scishow
----------
Support SciShow by becoming a patron on Patreon: https://www.patreon.com/scishow
----------
Huge thanks go to the following Patreon supporters for helping us keep SciShow free for everyone forever:
Chris Peters, Matt Curls, Kevin Bealer, Jeffrey Mckishen, Jacob, Christopher R Boucher, Nazara, charles george, Christoph Schwanke, Ash, Silas Emrys, Eric Jensen, Adam, Brainard, Piya Shedden, Alex Hackman, James Knight, GrowingViolet, Sam Lutfi, Alisa Sherbow, Jason A Saslow, Dr. Melvin Sanicas, Melida Williams
----------
Looking for SciShow elsewhere on the internet?
SciShow Tangents Podcast: http://www.scishowtangents.org
Facebook: http://www.facebook.com/scishow
Twitter: http://www.twitter.com/scishow
Instagram: http://instagram.com/thescishow
----------
Sources:
https://www.frontiersin.org/articles/10.3389/feart.2020.00321/full
https://www.zooniverse.org/projects/vivitang/earthquake-detective
https://www.essoar.org/doi/10.1002/essoar.10507021.1
http://www.isc.ac.uk/cgi-bin/stations?sta_list=&stnsearch=GLOBAL&stn_ctr_lat=&stn_ctr_lon=&stn_radius=&max_stn_dist_units=deg&stn_bot_lat=&stn_top_lat=&stn_left_lon=&stn_right_lon=&stn_srn=&stn_grn=
https://www.annualreviews.org/doi/abs/10.1146/annurev-earth-060313-054648?journalCode=earth
https://pubs.er.usgs.gov/publication/70192475
https://earthquake.usgs.gov/data/dyfi/background.php
Images:
https://www.istockphoto.com/photo/earthquake-seismogram-recording-by-a-seismograph-image-gm184093354-1828513
https://www.storyblocks.com/video/stock/apartment-earthquake-shaking-bmzzumdrqjowveyvg
https://www.storyblocks.com/video/stock/mesh-earthquake-macslkl
https://www.istockphoto.com/vector/vector-illustration-of-an-abstract-scheme-which-contains-people-icons-gm1247980427-363437930
https://commons.wikimedia.org/wiki/File:JAGI_Waveform_-_19_August_2018.gif
https://www.istockphoto.com/vector/earthquake-wave-seismograph-vector-gm1047684770-280241725
https://www.istockphoto.com/vector/3d-neural-network-with-six-layers-gm1165885651-320957853
https://www.storyblocks.com/video/stock/silhouette-of-people-rejoicing-and-lifting-up-his-hands-a-group-of-successful-businessmen-happy-and-celebrate-the-victory-on-the-roof-of-the-business-center-slow-motion-hhrqhg5lzj2tckekb
https://www.storyblocks.com/video/stock/animation-of-seismic-energy-wave-pattern-with-earthquake-ripple-s77fuidixjnhg69pr
https://www.istockphoto.com/vector/abstract-network-background-particle-wave-blockchain-neural-network-gm1195230552-340619965
https://svs.gsfc.nasa.gov/2893
https://www.istockphoto.com/vector/businesswoman-and-robot-shaking-hands-gm1141452215-305810039
Thanks to Brilliant for supporting this episode of SciShow.
Go to Brilliant.org/SciShow to check out Brilliant’s Introduction to Neural Networks course! [♪ INTRO]. It’s tough to miss a big building-rattling earthquake, but even if we do not feel a rumble beneath our feet, it may still contain useful information for scientists.
That’s why 24 hours a day, seismographs are listening closely to the rumbles of our planet. In fact, there are over 13,000 active seismic stations out there. And that is far more data than seismologists have time to go through.
Which is why, in a 2020 study, researchers set up a showdown of humans vs machines to sift through all this information and in the process, crown the heavyweight earthquake detection champion. One type of rumble that researchers are particularly interested in is called a teleseismic event. That’s when the waves from a big earthquake, often a long way away, trigger another, usually much smaller earthquake.
Geophysicists haven't come to a consensus on how one earthquake triggers another in this way, so figuring out when it happens and when it doesn’t may help us predict which areas are most at risk of these remotely triggered rumbles. For the study, published in the journal Frontiers in Earth Science, the researchers took data from seismic stations around Alaska to detect both small earthquakes and tremors caused by 30 larger quakes. Tremors are slower and weaker releases of energy than earthquakes are, which release their energy near instantaneously.
Then, the researchers provided the data to our competitors. First up, in the human corner: citizen scientists! Over 2000 “earthquake detectives” volunteered for the study.
They were not experts in seismology, just regular people who wanted to help out. They watched and listened to a small piece of the seismic data, converted into both a visible waveform and an audio file. The idea is that when we listen to sounds, our brains naturally do a great job of analyzing and classifying what we’re hearing.
In this case, you can easily identify earthquakes by a telltale sound similar to a slamming door. On the other hand, tremors sound like a train traveling over railroad tracks, and background noise can sound like anything from wind to static. So converting the initial data into a sound made it easier to spot the differences.
Each of these pieces of data was classified by 10 detectives as an earthquake, a tremor, or just background noise. But they had competition. In the opposite corner was team
AI: a machine learning algorithm. This was a deep learning model, meaning it had the ability to classify data based on multiple features, as well as learn from its past guesses and refine its predictions. It identified earthquakes by comparing a section of unknown seismic data to a reference database where each event was identified by experts. Essentially, the algorithm took a piece of unknown data and found the most similar examples in that database to classify it.
To figure out which team came out on top, the seismologist referees needed to identify some of the earthquakes themselves to figure out what the correct answer actually was. And after all that, the winner is… team citizen science! Their earthquake labels were 85% accurate, nearly 10% better than the AI.
They weren’t quite as good at classifying tremors as they were earthquakes, but at the time there was no AI that could detect tectonic tremors at all. So the humans still won out. Now, that does not mean the AI is out of the fight.
After this initial showdown, things took an unexpected turn. The researchers used the citizen science data to train the AI to do a better job. In another study published a year later, the AI was able to surpass the humans at earthquake detection.
This training even taught the AI to detect tectonic tremors for the first time ever. So maybe in the end, the moral of the story is humans and machines accomplish more by cooperating than by competing. It’s likely AI will become an increasingly valuable tool for finding earthquakes.
But the efforts of volunteer citizen scientists will still be needed to first point it in the right direction. This is the first time an AI has been trained to detect this kind of remotely triggered earthquake. So in the paper, the researchers encouraged other scientists to recruit volunteers to classify their data and grow the size of the AI training database.
So humans will continue to help the robots get better at detecting these rumbles, which in turn could help predict the risk of future quakes and keep those humans safe! And while this particular project is over, if you are interested in doing some citizen science yourself, we will leave some links in the description. If you want to learn more about the machine learning we talked about today, you can over on Brilliant, where they have a whole course devoted to it.
They have plenty more science, math and computer science courses too. And Brilliant has recently upped the interactivity factor to the next level! Now, courses like algebra and algorithm fundamentals are extra-engaging to guide you the whole way through.
If you’re interested, you can sign up at brilliant.org/scishow to get 20% off an annual Premium subscription. So thank you for your support! [♪ OUTRO].
Go to Brilliant.org/SciShow to check out Brilliant’s Introduction to Neural Networks course! [♪ INTRO]. It’s tough to miss a big building-rattling earthquake, but even if we do not feel a rumble beneath our feet, it may still contain useful information for scientists.
That’s why 24 hours a day, seismographs are listening closely to the rumbles of our planet. In fact, there are over 13,000 active seismic stations out there. And that is far more data than seismologists have time to go through.
Which is why, in a 2020 study, researchers set up a showdown of humans vs machines to sift through all this information and in the process, crown the heavyweight earthquake detection champion. One type of rumble that researchers are particularly interested in is called a teleseismic event. That’s when the waves from a big earthquake, often a long way away, trigger another, usually much smaller earthquake.
Geophysicists haven't come to a consensus on how one earthquake triggers another in this way, so figuring out when it happens and when it doesn’t may help us predict which areas are most at risk of these remotely triggered rumbles. For the study, published in the journal Frontiers in Earth Science, the researchers took data from seismic stations around Alaska to detect both small earthquakes and tremors caused by 30 larger quakes. Tremors are slower and weaker releases of energy than earthquakes are, which release their energy near instantaneously.
Then, the researchers provided the data to our competitors. First up, in the human corner: citizen scientists! Over 2000 “earthquake detectives” volunteered for the study.
They were not experts in seismology, just regular people who wanted to help out. They watched and listened to a small piece of the seismic data, converted into both a visible waveform and an audio file. The idea is that when we listen to sounds, our brains naturally do a great job of analyzing and classifying what we’re hearing.
In this case, you can easily identify earthquakes by a telltale sound similar to a slamming door. On the other hand, tremors sound like a train traveling over railroad tracks, and background noise can sound like anything from wind to static. So converting the initial data into a sound made it easier to spot the differences.
Each of these pieces of data was classified by 10 detectives as an earthquake, a tremor, or just background noise. But they had competition. In the opposite corner was team
AI: a machine learning algorithm. This was a deep learning model, meaning it had the ability to classify data based on multiple features, as well as learn from its past guesses and refine its predictions. It identified earthquakes by comparing a section of unknown seismic data to a reference database where each event was identified by experts. Essentially, the algorithm took a piece of unknown data and found the most similar examples in that database to classify it.
To figure out which team came out on top, the seismologist referees needed to identify some of the earthquakes themselves to figure out what the correct answer actually was. And after all that, the winner is… team citizen science! Their earthquake labels were 85% accurate, nearly 10% better than the AI.
They weren’t quite as good at classifying tremors as they were earthquakes, but at the time there was no AI that could detect tectonic tremors at all. So the humans still won out. Now, that does not mean the AI is out of the fight.
After this initial showdown, things took an unexpected turn. The researchers used the citizen science data to train the AI to do a better job. In another study published a year later, the AI was able to surpass the humans at earthquake detection.
This training even taught the AI to detect tectonic tremors for the first time ever. So maybe in the end, the moral of the story is humans and machines accomplish more by cooperating than by competing. It’s likely AI will become an increasingly valuable tool for finding earthquakes.
But the efforts of volunteer citizen scientists will still be needed to first point it in the right direction. This is the first time an AI has been trained to detect this kind of remotely triggered earthquake. So in the paper, the researchers encouraged other scientists to recruit volunteers to classify their data and grow the size of the AI training database.
So humans will continue to help the robots get better at detecting these rumbles, which in turn could help predict the risk of future quakes and keep those humans safe! And while this particular project is over, if you are interested in doing some citizen science yourself, we will leave some links in the description. If you want to learn more about the machine learning we talked about today, you can over on Brilliant, where they have a whole course devoted to it.
They have plenty more science, math and computer science courses too. And Brilliant has recently upped the interactivity factor to the next level! Now, courses like algebra and algorithm fundamentals are extra-engaging to guide you the whole way through.
If you’re interested, you can sign up at brilliant.org/scishow to get 20% off an annual Premium subscription. So thank you for your support! [♪ OUTRO].