3 ms·
>I think it's more like not hiring a big name coding competition winner because they never bothered to learn how to use version control, or any coding best prac
by reader5000 9y ago
>I think it's more like not hiring a big name coding competition winner because they never bothered to learn how to use version control, or any coding best practice, or any language other than C.
Depends on what you're hiring for, but I'll take "competition winner with no version control" over "average programmer with expert VC capabilities".
>Bayes rule isn't some kind of deep magic
Yes, it's largely conceptually obsolete.
The people jamming out weekly SOTA machine learning models on arxiv aren't sitting around meditating on conditional probabilities. They're making little tweaks to giant models that are basically impossible for a human to comprehend.
- XiaomiFan 9y ago> Yes, it's largely conceptually obsolete. Wow
- mmierz 9y agoI'm not sure I agree with that. I don't know any ML researchers that don't know about probability, but maybe they exist somewhere. Machine learning research isn't a good model for "data science" writ large. Maybe there are some jobs and some problem spaces where you can just tweak big black box models and you don't ever need to think about what their output means. But if you're the kind of data scientist who helps make decisions with data -- you better believe statistics and probability is conceptually relevant. As soon as models meet the real world, you've got to understand probability in order to know what to expect.
- achompas 9y ago> Yes, it's largely conceptually obsolete. I'm sorry, what? How did you arrive at a point where you believe this is true? This is like calling compilers "obsolete." Is it because you believe deep learning has "taken over" or something?
- reader5000 9y agoTry to derive e.g. a face detector from bayes theorem. You immediately arrive at computationally intractable sums/integrals. Yet, we have super-human image classifiers. Therefore, bayes theorem is obsolete. Sure, you can try to retrofit bayes theorem on top of a neural net, but who cares?
- achompas 9y ago> You immediately arrive at computationally intractable sums/integrals. So we instead sample from that posterior. Unless you think MCMC is also obsolete, in which case I’ll see myself out.
- reader5000 9y agoYou're right, but a) you have comp efficiency issues with MCMC, and b) just empirically MCMC models don't work as well as gradient descent + NN for many tasks.
- achompas 9y agoAnd you don't have computational efficiency issues with NNs? We're also ignoring the benefits of a posterior distribution, which is useful for understanding the data-generating process.
- reader5000 9y agoYeah of course. I can't explain to you why NNs outperform bayesian approaches, probably just NNs are capturing the correct type of prior for vision/language tasks. And yeah bayesian models are more interpretable but when you have millions of latent variables I'm not sure interpretability is a thing.
- achompas 9y agoYep, we arrived at my larger point: if you care about interpretability, NNs are horrible and Bayesian techniques are pretty damn great.
- reader5000 9y agoWell certainly, but interpretability is obsolete.