Instruction: Identify and discuss specific limitations or challenges when applying Transfer Learning to computer vision.
Context: The question challenges candidates to critically analyze the application of Transfer Learning in computer vision, including issues such as data and domain specificity.
Official answer available
Preview the opening of the answer, then unlock the full walkthrough.
The way I'd explain it in an interview is this: In computer vision, transfer learning can struggle when the target domain is very different from the source, such as moving from natural images to specialized medical or industrial imagery. It can also inherit biases...
easy
medium
hard
hard