By Seth Berger
According to Tech Edvocate, machine learning has developed and evolved over the years. Starting off as a rule-based methodology, the concept of machine learning was not taken seriously. However, in recent years, machine learning has become a powerful tool, having become a more data-driven method, something only possible through the enormous amount of data that colossal tech companies such as Google, Facebook and Microsoft have.
Machine learning is a way to extract information and learn from data. It is a form of artificial intelligence, a method used to teach machines to recognize things based on previous data given. It can be considered a function approximation, a tool used to associate ground truth and data.
Machine learning is being developed and improved every single day. Its potential can be seen through the many successful programs that have been written. One such example is AlphaGo, an algorithm written to beat players in the board game “Go.” AlphaGo beat the highest ranked Go player in the world but was then defeated by its newer counterpart, AlphaGo Zero. As a field that is growing at an exponential rate, it has the capability of solving real-world issues in a more comprehensive manner than thought possible.
One issue that can be solved is a lack of proper education. Google machine learning researcher and Carnegie Mellon PhD Da-Cheng Juan believes that machine learning could be a powerful asset in providing a better education for children.
“There are usually 30 students in a high school classroom, sometimes up to 200 in college classes,” Juan said. “They do not get individual attention, and the teachers are sometimes inconsistent. Machines are a lot more scalable and can provide help to students individually. So that is something I think would be a lot more helpful for students to have.”
Juan states that there are three different types in artificial intelligence — the subhuman, which is less effective than a human being, the part-human, which is just as effective as a human, and the superhuman, which is more effective when a computer does the task.
At this point in time, machine learning is still at a stage where machines can only do part-human and superhuman tasks within a certain domain. For the most part, machine learning is within the subhuman domain. This can be seen by the fact that machines are simply not as capable as people in doing things such as art, poems and other works that are considered “creative.”
Juan believes that this could possibly be one drawback for machine learning in education, as it is too systematic, something that could be somewhat uncomfortable for students. It is also lacking emotionally, as it would not understand that a nervous student could use words of encouragement or that an overconfident student should make sure to check his work.
Java student and sophomore Jainam Doshi has strong opinions on the matter.
“In the future, computers may be able to mimic human emotions through machine learning,” Doshi said. “Although we humans are more efficient in our learning, machines can take in more data. It is definitely possible, and I think it will happen sooner than we think.”
With that being said, the question that remains is if machine learning will ever take over education completely. APCS student and junior Jennifer Jiang is firm in her beliefs that AI will never take over education completely.
“In every age, people find a way to take advantage of the things they create, which I think leads back to the Industrial Revolution,” Jiang said. “It’s kind of like a wave. We are the surfers and the tide happens to be AI this time.”
As an ever growing field, machine learning has nearly unlimited potential. Its ability to recognize and make decisions based on data is astounding, no matter how primitive it is right now. Although somewhat systematic, it will undoubtedly support students effectively. It can give students the individual attention they need, assisting the students in their learning without delay. Machine learning may be the key to opening the doors to a new education system, reshaping the way we learn.