Additional layer types - convolutional layer, typo in the video slide?

From my understanding, no matter it is dense layer or convolutional layer, each neuron looks at the previous layer’s output rather than input. The key difference is that neurons in dense layers use all previous output but neurons in convolutional layers use some of the previous output.

I believe in convolutional layers, no neuron will look at anything “before” previous layer’s output (e.g., neurons in the 3rd layer will not look at the initial input and output of the 1st layer; it will only look at the output of the 2nd layer).

May I ask whether I am correct? If yes then the video slide might be adjusted.

Hello @Huixiang_Kong,

I agree with you. I will raise this to the course team. Thank you for posting this.