|Photo: Al Gharakhanian|
The topics of machine learning and, particularly, deep learning are clearly among the hottest topics covered by many tech publications. While the amount of hype is not insignificant, there are many good reasons why the space deserves substantial attention and coverage. To name a few:
- The reach and impact of machine learning/deep learning (ML/DL) has been proven over and over in hundreds of applications in a variety of disciplines. Applications such as advertising, autonomous vehicles, chatbots, cyber security, drones, e-commerce, fintech, industrial machinery, healthcare, marketing, robotics, and search engines are just a few key areas that have been impacted by ML/DL in a big way.
- Benefits of ML/DL are no longer limited by only an elite few that can afford fancy gear. The popularity of product recommenders and affordable chatbots among the general population is undeniable. One does not have to be a supervisionary to see a plethora of new and unexplored frontiers that are yet to be harnessed
- The costs of developing and deploying ML/DL pipelines are on a rapid decline. Even the most ardent skeptics of this technology can easily examine its uses and most likely will find value in them
- According to several ML/DL luminaries, deep neural networks "work unreasonably well," even though they aren't sure why. Just imagine the realm of possibilities when we are able to get to the bottom of such a remarkable performance.
Emergence of unsupervised learning.
The first and the most important macro trend in ML/DL is a gradual shift from supervised to an unsupervised learning paradigm.
The great majority of legacy ML/DL implementations are supervised learners. In other words, they can only be useful if they are trained by large amounts of labeled training data. While supervised learners have served us well, gathering and labeling large datasets are time consuming, expensive, and prone to errors. These challenges become far more pronounced when the size of the datasets increases. Unsupervised learners, on the other hand come with a huge advantage because they don’t require large training datasets and they learn as they go. This should explain why much of the advanced research in ML has to do with unsupervised learning...
Mr. Gharakhanian ends his article with the following: "This leap puts machine intelligence a step closer to human intelligence, enabling machines sætte soft skills som feeling and intuition to Learning."