SVMs and Deep Learning for Classification Tasks

While deep learning models are widely used, they come with the downside of being resource-intensive in terms of training and upkeep. In certain scenarios, opting for simpler models might be more practical. Support Vector Machines (SVMs) present a viable alternative for classification tasks. When should SVMs be chosen over deep learning models? Discuss the advantages and disadvantages of SVMs in comparison to both deep learning and other non-deep learning classification models like logistic regression.