August 13, 2013 at 1:20 AM
Should machine learning be used to make decisions about movie scripts?
You heard it right. Machine learning has been used in many applications in the past few years from suggesting products on amazon.com to recommending movies on Netflix.com. These algorithms are complex to the core, however, at a high level they all work basically the same. You provide historical data to be "learned" by the algorithms, and then you test the model to make a prediction based on the model. These methods have proved to work exceptionally well in finding patterns in user's choices and behavior.
But the big question is, what happens when you apply these methods to things that aren't focused on a particular person? I feel confident that if an algorithm has enough data consisting of movies that I've seen, it should be able to provide a recommendation of other movies that I may like. But what about analyzing data on previous block buster films? Just because these movies did well at the box office, does that mean that a computer program and construct an award winning, enjoyable movie? I have my doubts about this but at the same time, I believe this stems from a fundamental issue that man kind will have to face in this age of Big Data. The more that we take away the human element from decision making, the more people will begin to worry until the we see the proof that the data is enough to predict a great story from a mediocre, redundant one.
What do you think? Is this a good use of technology, or will this create lazy producers and predictable film scripts?
Here is a link to the original story
Solving Equation of a Hit Film Script, With Data