So we have a book on machine learning called "artificial mind system - kernel memory approach". This topic will be to explain what exactly the kernel memory approach is.
What is an artificial mind?
In 1997 a book was released titled "Artificial minds" which explores some of the deep philosophical and technical questions around this concept. The book also touches on the symbolic vs connectionist AI approaches. The concept of mind also can be explored in the extended mind theory by Chalmers and Clark.
The kernel memory approach
First we have a video on kernels which educates us on what kernels are in the context of AI:
The book a artificial mind system - kernel memory approach introduces an interesting concept called the self organizing kernel memory (SOKM). This approach is capable of data pruning and seems to offer some advantages over traditional neural net approaches. The kernel memory approach relies on the concept of "kernel units" and "link weights" representing the strength of the connections in between. The book is filled with simulation results, formulas/algorithms and examples. This provides for the necessary evidence demonstrating how the kernel memory approach works both theoretically by the math and by the results of the simulations.
Disclosure, I have not finished this book but was recommended to read this book. The concepts in this book are not something to be understood overnight but are important for a deep understanding of machine learning concepts which at the foundation are based on statistics.
References
Franklin, S., Wolpert, S., McKay, S. R., & Christian, W. (1997). Artificial minds. Computers in Physics, 11(3), 258-259.
Hoya, T. (2005). Artificial Mind System: Kernel Memory Approach (Vol. 1). Springer Science & Business Media.