[#6] The Missing Piece of the Most Popular Online Data Science Courses
Explainability and ethics in machine learning and AI should be introduced to aspiring data scientists from the beginning
This week, I published my first piece on Medium.com.
As I write in my post:
[W]e have reached a point in society where the topic of ethics in AI is being discussed widely. If we are going to actually tackle these issues of biases in AI models and work towards making our models better and more easily interpretable, both aspiring and current data scientists need to learn how to actually develop such models.
I researched the most popular online courses in data science and machine learning, to see if/how they introduce issues about biases in machine learning models and making models more easily interpretable. By and large, these issues are lacking from these courses.
Hence, online courses need to be teaching these topics from the beginning, and they need to be constantly updating to keep up with the state-of-the-art explainable AI techniques in the field.
Please check out the full article, and let me know what you think!
The Explainability Corner
I like to end all of my posts with a helpful link pertaining to searching for jobs or programming in Python, etc. In the spirit of today’s main post, I will share a resource pertaining to explainable AI.
Here is an interesting post on Medium about a technique for interpreting XGBoost models. It goes through some issues with existing methods for measuring feature importances and describes their approach. Here is the corresponding paper on arXiv, and the GitHub repository.
Enjoy the rest of your week!