[This page is currently UNDER CONSTRUCTION]
Imagine a machine learning algorithm with interpretable parameters. What would you do with those interpretable parameters? Think about that for a second. You will probably have a hard time to answer that question.
In this project, I look for answers for such questions. Answering those questions would also clarify the definition of "interpretable machine learning" concept for us. Recently, I introduced an interpretable machine learning algorithm with interpretable kernel parameters. Since the computed kernel parameters are now interpretable, those computer parameters now can be used in various new ways. Consider that the parameters you computed can be used as a measure of the homogenity around them. Would not such a property be useful for many tasks? One obvious task would be the filtering operation. Another task would be data abstraction. There are many other possibilities. More of those possiblilities will be discussed (in details) soon here.
You can read more about my interpretable machine learning algorithm here (published at IEEE Transactions on Image Processing, 2018).