On: Introduction to Support Vector Machines in Machine...

[bitforger]

ITT: Whether SVMs are still relevant in the deep learning era. Some junior researchers will say neural networks are all you need. Industry folks will talk about how they still use decision trees.

Personally, I'm quite bullish on the resurgence of SVMs as SOTA. What did it for me was Mikhail Belkin's talk at IAS.[1]

[1] https://m.youtube.com/watch?index=15&list=PLdDZb3TwJPZ5dqqg_...

[stu2b50 → bitforger]

I mean NNs are still quite bad at low n tabular data (and they may always be), which is honestly how a lot of real life data is, so there is clearly a need for not a neural network.

I feel like I've seem more tree ensembles in the wild than SVMs, though.

[rangerranvir → stu2b50]

Anyway the idea of NNs was introduced to work on data which a simple human brain couldn't make sense of.

For more general tabular data, like trees, regression and even rule based models are more realistic.

[ma2rten → bitforger]

I don't have time watch the video could you summarize why you think SVMs will become SOTA and on which problem?

[rangerranvir → bitforger]

Thanks for sharing. You actually shared a full playlist. Won't be able to get up before finishing a few of them.