DATA 310 - Antonio Marra Informal Response 2/26
Q: Convolve the two 3x3 matrices that were assigned to you with your 9x9 matrix and calculate the resulting two matrices
Q: What is the purpose of using a 3x3 filter to convolve across a 2D image matrix?
Using a 3x3 filter can create a more usable image for processing in ML models and otherwise, as it can retain certain desired features. Convolving across a 2D image matrix can compress a larger image into a smaller variant without altering the fundamental features of the image to a large degree.
Q: Why would we include more than one filter? How many filters did you assign as part of your architecture when training a model to learn images of numbers from the mnist dataset?
Using multiple filters allows the user to highlight distinct features within the photo, for example, emphasizing vertical lines over horizontal, or vice versa. In my mnist model I used three separate filters within the architecture, one for horizontal emphasis, one for vertical, and the final 2x2 filter.