The Machine Learning Specialization from Coursera.org is a great online course I just completed. It is a course created by DeepLearning AI and taught mainly by Andrew Nig who graduated from Stanford University and has much more expierience outside of it. The specialization itself is made up of three courses. The first taught only the very basics touching on linear regression, gradient descent, and some simple techniques. The second course goes into more complex topics like Neural Networks, Bias and Variance, and even Decision Trees. Lastly the third course went into the much more underdeveloped topics of unsupervised learning.
Personally, I am very glad that I chose this course. It is simple and does not require a lot of time or dedication to learn each subject while still covering a broad range of topics and machine learning skills. As a new freshman at the Hill School and for other even busier people, this is a huge bonus. It also includes simple exercises to further hone our skills and to help us better absorb the content.
While learning, I could not help but be amazing at how powerful machine learning is, and how perfect the terms artificial intelligence are. Having gained deeper insights on their inner workings, they are no less mysterious, the way that AI’s and become such experts on tasks even though they are really just massive mathematical models and lots of data. I was also one again reminded of how new and explosive these recent years have been. In just teh 15, 20 years since the repopularization of AI, there have been so many studies and discoveries like the Adam Optimizer, making gradient descent much more efficient or the in-depth studies on Varience and Bias, important ideas for tuning machine learning models.
In contrast, I noticed that unsupervised learning and more specifically reinforced learning seem a lot more underdevopled. The techniques used are very brute force like and very computationally expensive, well more than the other learning models. However I beleive that this only shows that despite Machine Learning’s massive growth, there are still areas for improvement.
After learning these skills, I can scarcly wait to implement them on my own. Not only this, I know that I have only scratched the tip of the iceberg and is so excited to learn more, and in the future I dare not say when, I may make some discoveries myself so I can give back to this fascinating field of Computer Science.
Lingyuan's blog 