The secret behind this success is the ability of CNNs to automatically learn to identify relevant low and high level characteristics of the input images. The network generates an increasingly explicit image representation, learning to combine low and high level features to finally care about the actual content, instead of individual pixel information. This final representation is used to perform the classification of the emotions in several categories, such as, sadness, joy, anger, fear, surprise and neutrality. A very detail…
Just like machines, artificial neural networks, convoluted or otherwise cannot learn or “deep” learn. Only of a (mostly) whole human person (and some non human animals) with a (mostly) fully functional nervous system including a brain can we say they are capable of learning. To say otherwise is to commit the mereological fallacy and/or the compulogical fallacy. Of course, machine learning is already a term composed of two words that when combined in that order result in a logical contradiction. To say it has logical problems is the understatement of the century. Also, “deep” is a bullshit hedge term that means nothing. Try rewording this statement in a way that is not absurd. Example below.
“The secret behind this success is the ability of CNNs to automatically compute relevant low and high level characteristics of the input images. The network generates an increasingly explicit image representation as more computations are completed which result in a higher probability of it selecting the actual content for additional computations than the data from the individual pixels. The final results of these computations is the selection of the variable (emotion in this case) that best fits the computed data (probability distribution curve). Clever statistical and mathematical techniques are applied to the data (which attempt to mimic the hypothetico deductive reasoning process but are at root a series of feedback loops using weighted averages) resulting in higher probabilities of the correct variable being selected over time.”