{"id":219,"date":"2020-08-26T11:46:27","date_gmt":"2020-08-26T15:46:27","guid":{"rendered":"https:\/\/www.macloo.com\/ai\/?p=219"},"modified":"2020-08-26T13:23:23","modified_gmt":"2020-08-26T17:23:23","slug":"interrogating-the-size-of-ai-algorithms","status":"publish","type":"post","link":"https:\/\/www.macloo.com\/ai\/2020\/08\/26\/interrogating-the-size-of-ai-algorithms\/","title":{"rendered":"Interrogating the size of AI algorithms"},"content":{"rendered":"\n

I have watched so many videos<\/em> in my journey to understand how artificial intelligence and machine learning work, and one of my favorite YouTube channels belongs to Jordan Harrod<\/a>. She’s a Ph.D. student working on neuroengineering, brain-machine interfaces, and machine learning.<\/p>\n\n\n\n

I began learning about convolutional neural networks in my reading about AI. Like most people (?), I had a vague idea of a neural network being modeled after a human brain, with parallel processors wired together like human synapses. When you read about neural nets in AI, though, you are not<\/em> reading about processors, computer chips, or hardware. Instead, you read about layers<\/em> and weights<\/em>. (Among other things.)<\/p>\n\n\n\n

A deep<\/em> neural network has multiple layers. That’s what makes it “deep.” You’ll see these layers in a simple diagram<\/strong> in the 4-minute video below. A convolutional neural network has hidden<\/em> layers. These are not hidden as in “secret”; they are called hidden because they are sandwiched in between the input layer and and output layer. <\/p>\n\n\n\n

\n