Videos by Brandon Rohrer:
https://www.youtube.com/watch?v=Q9Z20HCPnww - in this video I liked how he makes a connection between actual neurons and neural networks. Weights mean the strength of connections between neurons. Connection between dendrite of one neuron and axon of another is called "synapse". Although the shawarma guy example was not easy for me to understand...
The other one "How Neural Networks work https://www.youtube.com/watch?v=ILsA4nyG7I0 has a pretty simple step-by-step explanation of NN workflow based on a simple example of 2x2 "camera". Good explanation of how weights work, backpropagation, sigmoid, ReLU, etc.
A friendly introduction to Deep Learning and Neural Networks by Luis Serano (https://www.youtube.com/watch?v=BR9h47Jtqyw) is easy to understand and really friendly for the first 2/3, but then he jumps to calculating probabilities, errors, etc. and it becomes harder to follow. In general, it's a common problem of most of the tutorials I have seen on ML and DL: they start with 1+1 = 2, but pretty quickly they jump into really deep stuff without warning.
https://www.youtube.com/watch?v=Q9Z20HCPnww - in this video I liked how he makes a connection between actual neurons and neural networks. Weights mean the strength of connections between neurons. Connection between dendrite of one neuron and axon of another is called "synapse". Although the shawarma guy example was not easy for me to understand...
The other one "How Neural Networks work https://www.youtube.com/watch?v=ILsA4nyG7I0 has a pretty simple step-by-step explanation of NN workflow based on a simple example of 2x2 "camera". Good explanation of how weights work, backpropagation, sigmoid, ReLU, etc.
A friendly introduction to Deep Learning and Neural Networks by Luis Serano (https://www.youtube.com/watch?v=BR9h47Jtqyw) is easy to understand and really friendly for the first 2/3, but then he jumps to calculating probabilities, errors, etc. and it becomes harder to follow. In general, it's a common problem of most of the tutorials I have seen on ML and DL: they start with 1+1 = 2, but pretty quickly they jump into really deep stuff without warning.
Comments
Post a Comment