Deep learning antipatterns
Antipatterns in deep learning are common solutions applied in the recurring challenges, which result in suboptimal models and often lead to highly counterproductive development flows.
While some antipatterns are easy to recognize because they are general bad practices, others only become antipatterns if they are applied in the wrong situation. To illustrate this, I will present real-life examples of two antipatterns that both can be design patterns when they are used for the right challenges but are highly counterproductive otherwise.