Thursday, December 7, 2017

New AI system Can Predict Earthquakes

Scientists have developed an artificial intelligence (AI) system to successfully predict earthquakes, an advance that may help prepare for natural disasters and potentially save lives. This AI system successfully identified a hidden signal leading up to earthquakes and used this ‘fingerprint’ to train a machine learning algorithm to predict future earthquakes. This hidden signal used to be emitted by traditional algorithm. Researchers were impressed by the findings in this AI system. "This is the first time that machine learning has been used to analyse acoustic data to predict when an earthquake will occur, long before it does so that plenty of warning time can be given - it is incredible what machine learning can do," said Colin Humphreys of Cambridge University.

Compared to traditional system, AI system have Machine learning enabling the analysis of data sets too large to handle manually and looking at data in an unbiased way that enables discoveries to be made. Therefore, when traditional system and AI system analyze a same data set, they probably produce different results.

This characterization of AI should be emphasized in the mind of every MIS student. While most of computer system tries its best to imitate human and ignore some data due to human’s preconceptions, AI system begins to eliminate the intervention of human beings. Most of algorithms in AI system are principled and not related to human experiences. Therefore, AI system can unbiasedly treat every detail in a great number of data sets and successfully find a key to solve some problems which people currently cannot solve.

Question:
Do you think Artificial intelligence can always outperform the human intelligence?
If do, how can we handle AI when we develop future system?
What kind of area you suggest AI system should involve in?
Source:
“New AI system Can Predict Earthquakes.” News18, 23 Oct. 2017,
www.news18.com/news/tech/new-ai-system-can-predict-earthquakes-1554255.html





3 comments:

  1. Hello Yi, I believe that AI does not always outperform human intelligence but the ability to learn of a machine far outpaces the ability to learn as a human. An AI can perfectly correct a mistake and will always use that correction going forward whereas a human will remember the mistake but not always correct it going forward. This can be seen very simply in the AI system "Mari/O" which shows an AI learning to play mario and becoming perfect at it very quickly. I believe we should be careful with machine learning and take it slow because we do not know where it could ultimately take us for better or worse.

    ReplyDelete
  2. Hi Yi.
    I think that the technology behind Artificial Intelligence is growing at an extremely rapid pace and it could very easily become smarter or better than humans (at most tasks) very soon. With this in mind, I think your second question is a very important one. Since the technology is growing so rapidly, I think it is important for our governments to develop laws and guidelines surrounding its use before it continues to expand out of our control. I know that Elon Musk suggests all countries developing AI tech need to work together to prevent its use in warfare. I am in agreement with him about this subject, as this technology could prove to be even more dangerous that the atomic bomb.

    I would be very interested to see how Artificial Intelligence could be leveraged in a classroom environment and how it could assist teachers in certain settings or with certain tasks.

    ReplyDelete
  3. Maybe we could watch as AI teaches students and learn from it, rather than the other way around. :)

    ReplyDelete