I was reading the beginning of the first chapter of Neural Networks and Deep Learning last night. In the introduction of perceptrons it reminded me that NAND gates can be used to construct all other logic gates. Surely I learned that in my first computer architecture course and explained it to others when I was a TA for “Computer Architecture I” during my first graduate school semester!
For a fun refresher, I decided to design other logic gates using NAND gates. While the design results aren’t novel, the figures I ended up with are attractive enough to share.
Here’s a better XNOR gate that I didn’t get until verifying my other gates on another website. It has three layers instead of the four layers above.