Recently I came across a very interesting TedxSF
talk by Jeremy Howard who is an Australian data scientist. This talk is from
2013 and may be dated but nonetheless I found it very interesting. In his talk,
he discusses machine learning devices and when he does so he uses an example of
playing checkers on a computer. Howard describes how remarkable these devices
are through checkers: by programming broad parameters to what you are trying to
accomplish (which in this case teaching a computer to play checkers) and then
defining all possible approaches to playing checkers. Once this is done, Howard
discusses how you can then get the computer to play itself thousands of times
and each time it does it would be taking the better approach until it actually taught
itself to play better than the programmer! If that was kind of confusing Howard
puts it quite eloquently in the talk by saying “Machine learning is giving a
computer a broad approach or parameters in which it can create a solution to a
problem better than your own solution to that problem.” (Howard, 2013) Algorithms are getting better and
better by comparing them to each other, getting a score on how well they do what
they are intended to do. This process is continual and new algorithms are
created to beat the one before it. Howard even says that there are now international
rankings for data scientists.
Howard then goes on to talk about
algorithm competitions and how good algorithms have gotten. This is what the
results of these competitions look like “There was recently a machine learning
competition where a computer actually graded hand written student essays.
30,000 of them, more consistently with two teachers than the teachers did with
each other. So, we are now at a point where computers can read better than
teachers.” (Howard, 2013) If a computer can grade more consistently
than teachers can it makes me ask the question, what is the need for teachers?
Are the services that we provide society becoming obsolete by new technologies?
Howard relates what is happening now, to the labour jobs lost during the
industrial revolution and how people then shifted to the service industry. Well
now this technology is encroaching on the service industry. So, what is next?
Do you think that society will allow new technological advances to disrupt in
the service industry and if so, what will be the next industry that these lost jobs
will shift to? Let me know what you think in the comments!
The sources
used in this article and Jeremy Howard’s TedxSF Talk can be found at the
following link:
Works
Cited
Howard, J. (2013,
November 17). Machine-Learning: Connecting Devices with Algorithms.
Retrieved from Datafloq:
https://datafloq.com/read/machine-learning-competitions-connecting-devices-a/327
This is a really interesting topic! I was not aware that computers had the ability to read and grade hand written essays, which is quite shocking. While this ability might make it seem like teachers, and the services of humans in general, are becoming obsolete, I would like to remain optimistic in believing that humans still have the unique ability to be thoughtful and emotional. A computer might not pick up and register the same components of an essay that a human reader would. A teacher can most likely understand each individual student and their writing style and thought process more effectively than a computer would. I would hope that society will continue to value humans and their unique abilities as computers become more advanced.
ReplyDeleteHowever, as for what's next, it unfortunately seems as though technology may in fact disrupt the service industry. For example, restaurants such as McDonalds have incorporated kiosks where the customer can order their own food, ultimately making the customer service employee's job obsolete. I think it's only a matter of time before we see technological advancements popping up in all aspects of society. The next industry could very well be all restaurant employees/waitresses/bartenders.
Thanks for your interesting topic!