To my understanding, transfer learning helps to incorporate data from other related datasets and achieve the task with less labelled data (maybe in 100s of images per category).

Few-shot learning seems to do the same, with maybe 5-20 images per category. Is that the only difference?

In both cases, we initially train the neural network with a large dataset, then fine-tune it with our custom datasets.

So, how is few-shot learning different from transfer learning?

  • 1
    $\begingroup$ Can you please put your main specific question in the title? "Few shot and transfer learning" is not a question and it's not specific. Take a look at this for the motivation. $\endgroup$ – nbro Jan 7 at 10:38
  • $\begingroup$ I edited your post to remove the question "Can somebody help me understand how approaches like meta-learning are different from transfer learning?", because it's a completely different question than "How is few-shot learning different from transfer learning?". By the way, we already have that question here. I also try to clarify certain parts of your post. Please, make sure that I didn't change the meaning of the post. $\endgroup$ – nbro Jan 7 at 17:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.