What is the difference between eager learning and lazy learning?

How does eager learning or lazy learning help me build a neural network system? And how can I use it for any target function?


What is eager learning or lazy learning?

Eager learning is when a model does all its computation before needing to make a prediction for unseen data. For example, Neural Networks are eager models.

Lazy learning is when a model doesn't require any training, but all of its computation during inference. An example of such a model is k-NN. Eager learning is also known as instance-based learning [1, 2, 3].

How does eager and lazy learning help me build a neural network system? And how can I use it for any target function?

To answer your second question, you can't employ lazy learning to train a neural network, because they are inherently eager models.

| improve this answer | |
  • 1
    $\begingroup$ Maybe you should cite some book or paper that discusses this terminology because, although what you write makes sense, I am not completely sure this is standard terminology. Maybe it's just been a while since I had to deal with this terminology and that's why I am not completely convinced. $\endgroup$ – nbro Jul 31 at 12:27
  • $\begingroup$ @nbro I added some links, hope it's OK. $\endgroup$ – Djib2011 Jul 31 at 13:46
  • $\begingroup$ Well, actually, you added some links to other non "reputable" sources, so I edited your post only to leave two links to papers that seem somehow reliable. I think that's enough, unless there's a better source. $\endgroup$ – nbro Jul 31 at 14:02
  • $\begingroup$ @nbro I think the terms are sufficiently well-established. I do recall having seen them in my education now that I think of it :) Maybe not as much of a commonly-discussed topic anymore in recent years with many people behaving as if (deep) neural networks are the only types of machine learning models that exist... $\endgroup$ – Dennis Soemers Jul 31 at 14:06
  • 1
    $\begingroup$ @nbro I think that the default notion of learning has shifted towards eager learning. This way NNs aren't described as eager learners because that's the default. In contrast lazy learners are described as such to differentiate with the other more popular models. TBH, I don't agree with the terminology, because no actual learning happens in lazy models (no knowledge is extracted from the data), but since they are considered ML models, they need a way of distinguishing themselves... $\endgroup$ – Djib2011 Jul 31 at 15:03

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.