Can someone please clarify if the below notation is correct or not. lecturer writes: Z_x = x^T*w is a GP on S = R^d I think this might be a mistake and should instead be: Z_x = x^T*w is a GP on S = R^n Since Z_x is a collection of n observed random variables and not over d dimensions. Therefore, Z_x is a GP on S = R^n. Am I correct in saying this?
Somebody already asked this, but I can't find it, so here goes: Bayesian Lin. Regression is a parametric model, correct? (You put prior on parameters) GP Regression is a nonparametric model. (Since you put prior on functions, as far as I get it. Is this right?)
I don't know if I understand it correctly. Is it true that a Gaussian Process Regression Inference is based on the assumption of w ~ N(0, a) ? In many videos and PRML, they use w ~ N(0, a) to derive the GP Regression from Baysian Linear Regression, however they did not mention the assumption.
hi it's been a 8yrs that this video is out, hope you ok dude, THANK YOU A LOT
Same for me. =)
@@iMuhannedHD Indeed he is: jwmi.github.io/ :D Very cool
Videos are very informative. Amazing stuff! Thank you!
perfect explanation! thank you sir!
Fantastic! I like a lot your videos....congratulations!
Very clear explanation! Thank you!
Can someone please clarify if the below notation is correct or not.
lecturer writes: Z_x = x^T*w is a GP on S = R^d
I think this might be a mistake and should instead be: Z_x = x^T*w is a GP on S = R^n
Since Z_x is a collection of n observed random variables and not over d dimensions. Therefore, Z_x is a GP on S = R^n.
Am I correct in saying this?
Somebody already asked this, but I can't find it, so here goes:
Bayesian Lin. Regression is a parametric model, correct? (You put prior on parameters)
GP Regression is a nonparametric model. (Since you put prior on functions, as far as I get it. Is this right?)
Yes, that is correct.
many thanks! is radial basis function network (RBFN) the same as GP then?
I like it....I am your fans
All of them?
Linear combination of multivariate gaussian is not univariate gaussian it’s also multivariate
hi dear Monk, would you do a cast on RNN? thanks
I don't know if I understand it correctly. Is it true that a Gaussian Process Regression Inference is based on the assumption of w ~ N(0, a) ? In many videos and PRML, they use w ~ N(0, a) to derive the GP Regression from Baysian Linear Regression, however they did not mention the assumption.