The history of machine learning begins in 1950 with Alan Turing and the “Turing Test”, a test he created to determine whether or not a computer has real intelligence. In order for a computer to pass the test, it must be able to fool a human into believing it is also human. The test basically was a guessing game. There was a judge, a female contestant and a male contestant and the based off the answers to a series of questions the judge had to choose which of the respondents was human and which was the computer. The most recent modification to the test however doesn’t involve contestants. Simply, it is a judge and a computer and the judge has to determine if the respondent is a computer or a human. Since Turing introduced this test, the results have proven to be highly influential and also regularly criticized. However, overall the test has become incredibly influential in the artificial intelligence and machine learning world.
Another major advancement that occurred in the machine learning world was in the 1990s when scientists began creating programs for computers to analyze. From the data, computers where able to draw conclusions based on the massive amounts of data they were presented with. Simply, the computers were able to “learn” from the data. Computers, like IBM Watson are now able to store endless amounts of data, ranging from different topics, and they are able to make connections between the data in order to draw conclusions. For example, in 2011 IBM Watson was put to the test. Watson went up against the two greatest Jeopardy champions. After the first round, Watson was ties for first with $5,000. However, Watson made an incredible comeback and entered into final Jeopardy with $36,681, the next player having only $5,400. This test was groundbreaking not only for the IBM team, but also for AI and machine learning. Watson was able to enter into its data base to collect data it needed to answer the question, analyze the data, and great an answer all within seconds.
In 2012, Google developed an algorithm that was able to autonomously look through YouTube videos to identify all the videos that contained cats in them. Similarly, in 2014, Facebook developed the DeepFace software which was an algorithm that was able to recognize and verify individuals based on their photos.
These singular advancements in machine learning have made huge impacts on technology today and how we use it daily. Seeing how computer intelligence went from the Turing machine up to iPhones with facial recognition is incredible and it just shows that it will continue to change rapidly and that there is still much more to learn.
I work at a market research firm and ever since the semester started I’ve begun to think about how disruptive technology will affect the particular industry I work in currently. Within my firm I work in the media and entertainment department which is why when it came to choosing what industry I wanted to focus my machine learning project on, I jumped at the opportunity to focus on entertainment. Spencer made a good point in one of our first classes, that when machines take over and people are out of jobs, there will always need to be a solid entertainment industry to keep people busy.
Google gives us a goo idea of what entertainment is like for us now because of the machine learning that how technology uses. It creates a more personalized experience for us and allows us to watch our content when we want.
According to Forbes, there are six major digital transformations in the media and entertainment industry. Multi-channel experiences are the norm now; Accenture Digital did a study that showed most people obviously use different devices to watch certain things, but often, people are viewing content on their devices simultaneously. Creators are becoming scared that it’s not necessarily the content that people care for, it’s more about the convenience. This is where the AI comes into play. AI is getting more and more creative, according to Forbes. The technology is used in the entertainment world in many ways, one being to create plots of shows and movies based off of box office ratings. This new wave of computer-human collaboration is already working effectively within the industry. The result of this collaboration and integration of AI in the entertainment industry is that the computer will learn how to collect box office ratings and Nielson TV ratings data on its own to then create a plot and compile a final trailer for review all within 24 hours, which is significantly less than the average 30 days spent editing manually.
A prime example of machine learning specifically in the entertainment industry is located within Netflix. Based on what you watch, Netflix recommends shows or movies that are similar to the shows you have watched or are currently viewing. To that point, in 2013, Netflix released “Max” your personal recommender on the app. However, this feature crashed and burned due to the fact that “Max” didn’t sync properly with consumers Netflix accounts making is results less and less accurate and his recommendations poor. This sent Netflix back to the drawing board to see how to use machine learning to their advantage.
Similarly to Netflix, most companies are researching how to benefit from AI and machine learning as both technologies become more prominent. There is a lot of research and development regarding machine learning for entertainment companies and they seem to have their finger on the pulse, as of right now.