Deep learning has recently produced a number of amazing success stories. As a result, you might wonder whether you should abandon standard machine learning methodologies in favor of deep learning. The response is contingent on the circumstances and the work at hand. In this part, we'll go through why you shouldn't utilize deep learning and when it's a good idea to use it.
When not to use deep learning?
Deep learning includes millions of parameters and, as a result, requires a large amount of data to train.
Deep learning isn't the way to go if you just have access to a restricted amount of characteristics that define each instance. The following apps are included in this category:
- Predict the scores of a student in their first university year based on only their
- scores in high school
- Predict the risk for a heart attack within the next year based on the sex, age, BMI (body mass index), blood pressure, and blood cholesterol concentration of a person
- Classify the sex of a turtle based on its weight, its height, and the length of its feet Also, in situations where you have only few training data and you know exactly which features determine the outcome of interest (and it’s easy for you to extract these features from your raw data), then you should go for these features and use those as abasis for a traditional ML model. Imagine, for example, you get images from a soccer player collection of different individual French and Dutch soccer players. You know that the jerseys of the French team are always blue, and those of the Dutch team are always orange. If your task is to develop a classifier that discriminates between players of these two teams, it’s probably best to decide if the number of blue pixels (the French team) in the image is larger than the number of orange pixels (the Dutch team). All other features (such as hair color, for example) that seem to discriminate between the two teams would add noise rather than help with the classification of new images. It’s therefore probably not a good idea to extract and use additional features for your classifier.
When to use deep learning?
When each instance is defined by complicated raw data (such as photos, text, or sound) and it is difficult to construct the crucial characteristics that distinguish the distinct classes, deep learning is the approach of choice. Deep learning models can then extract features from raw data that outperform models that rely on created features.
Picture displays various tasks in which deep learning recently changed the game.
The various tasks recently solved by deep learning that were out-of-reach for traditional ML for a long time