Data Science Exam

The Exam… #

Last week was the In Class Assessment Week (ICAW), which was a week dedicate to in class assessments for school. It would have one of the three semesterly assessments for each class that I’m apart of, those being Data Science, Web Dev, Math, English and History.

The exam for Data Science was on Machine Learning, as that is the content that was being covered last semester and built on this semester. The question asked was along the lines of Machine learning implementations provide great benefits to society, which I interpreted to be about the ethics behind machine learning and the impacts that that has on society.

The stimulus given to us looked into the use of Machine Learning (ML) models to detect when a person is using their phone in parliament and for how much time. We were presented with the Cambridge Analytica scheme, and the way that it affected society. We had a bell curve showing the distribution of how people react to making sure ML algorithms are free of bias and don’t become racist, sexist (pretty much just evil). The final piece of stimulus we were given was a world map showing the countries that use facial recognition and those who have banned it. From there, I had to take what I knew and apply it to the question.

I agreed with the statement to some degree, but also had to take into account the issue with training a machine learning model with inherent bias and how that works, and how that can be combatted. I provided a bit of historical context on the topic with reference to the Cambridge Analytica scheme as I have recently been reading a book on it called Mindf*ck: Cambridge Analytica and the Plot to Break America. I brought up the issue with data handling and management, looking into a few examples of bad management and manipulation of data such as the scheme stated previously and the way that data was created to be used in other circumstances such as Microsoft’s TAY twitter bot which fell on its face due to internet trolls, and Tesla who have used machine learning to improve Tesla’s autopilot feature on cars and much more.

One of the highlights that I put in there, though, was that there is, and always will be, danger with how data is handled and used. Data isn’t inherently malicious, but it’s down to the people who use it that define whether it will be used in a good way or maliciously. There are people out there that will use it for good and try to figure out ways that it can applied to new problems to give new insights into the world, but there are also those that use the data for their own personal and financial gain, which doesn’t usually end well.

If I get permission, I’ll try to post the response here, but if not, then I won’t.

Conclusion #

Overall, the exam did go quite well, as I was able to smash out approximately 1100 words, just like the Web Dev exam, in the 70 minutes provided with useful information that related to, and answered the question. I am happy with how I went in the exam, and with my application of knowledge on the different concepts and machine learning implementations in the world to be able to to talk about those implementations as well as how they affected the world and the ethics behind them.