Thursday, July 2, 2009

A Majorization-Minimization Algorithm for (Multiple) Hyperparameters Learning


First we must understand what the majorization-minimization (MM) algorithm is. It is in fact the auxilliary function method, which might be regarded as a generalization of EM algorithm. To minimize (maximize) a function L(x), we find an upper (lower) bound for the objective which is more easier to minimize (maximize), Q(x; x') where the x' is the parameter for the upper (lower) bounding function. This can be seen as a generalization of the idea of CCCP, since in CCCP, we use linear functions to bound the convex (concave) functions and the parameter is simply the variable introduced by Legendre-Fenchel transform. And we require L(x) - Q(x; x') reaches its minimum (maximum) at x = x'. As we can see here, the following inequality holds
L(x) \leq Q(x; x') \qquad \text{or} \qquad L(x) \geq Q(x; x')
The idea is instead of optimizing L(x) directly due to its intractability, we may find suitable algorithm to optimize Q(x; x') where x' = x^{(n)}. Therefore for the minimization case we have the following inequality
L(x^{(n)}) = Q(x^{(n)}; x^{(n-1)}) + L(x^{(n)} - Q(x^{(n)}; x^{(n-1)})) \leq Q(x^{(n-1)}; x^{(n-1)}) + L(x^{(n-1)}) - Q(x^{(n-1)}; x^{(n-1)}) = L(x^{(n-1)})
where the inequality uses two optima, Q(x^{(n)}; x^{(n-1)}) is minimum of Q(x; x^{(n-1)}) and L(x) - Q(x; x^{(n-1)}) reaches its maximum at x = x^{(n-1)}.

Now let's come back to the Bayes learning problem. We know to select a proper hyperparameter, either we use cross validation, or we rely on maximization of Type II likelihood (EM solves it :-p). Here the author simply introduce another prior for the hyperparameter and integrate it out. E.g. for a Gaussian prior for the parameter, usually the hyperparameter is the precision, whose prior is then set to a Gamma distribution. But they kind of stealthly switch the concept. In Bayes framework, the joint distribution is then \Pr(\mathcal{D} \mid w) \Pr(w \mid \alpha) \Pr(\alpha). They first integrate out the hyperparameter \alpha, which leaves us \Pr(\mathcal{D} \mid w) \Pr(w). The we find the MAP estimation for simplicity, which results in
\min_w -\log \Pr(\mathcal{D} \mid w) - \log \Pr(w).
Now we may find the second term is of the form of - C\cdot\log (\| w \|^2 + \beta ), which is a different from the norm regularizer.

We solve this problem with MM optimization technique. We construct an upper bound for the second term
\log x \leq \log y + \frac{x - y}{y}
which is a linear function of x and therefore the optimization problem turns back to the original ``loss + norm regularizer'' form.

Don't see any thing special from this point of view.

24 comments:

Anonymous said...

buy xanax online order xanax bars - xanax bars mg

Anonymous said...

zolpidem high zolpidem cr side effects - zolpidem tartrate qualitest

Anonymous said...

where can i buy xanax online legally generic xanax just good - how to order xanax without an rx

Anonymous said...

buy diazepam cheap diazepam 10 mg injection - diazepam yellow tablets

Anonymous said...

ativan online ativan and alcohol addiction - ativan withdrawal taper schedule

Anonymous said...

cheap ativan online ativan dosage withdrawal - ativan uses anxiety

Anonymous said...

buy xanax 1mg xanax 2mg gg249 - xanax online

Anonymous said...

ativan price ativan overdose many - 6 mg ativan overdose

Anonymous said...

ativan online ativan overdose limit - ativan generic name lorazepam

Anonymous said...

buy ativan ativan withdrawal length - cost of generic ativan

Anonymous said...

buy valium online order valium online forum - legal buy valium online usa

Anonymous said...

soma online carisoprodol 350mg street value - carisoprodol ld50

Anonymous said...

ambien online no prescription ambience mall gurgaon features - ambien vomiting

Anonymous said...

soma without prescription many carisoprodol 350 mg get high - soma online offer code

Anonymous said...

buy valium online buy valium online no prescription cheap - diazepam valium vs xanax

Anonymous said...

Blogger: Paper Scanner - Post a Comment prednisone online - prednisone online pharmacy http://www.ourdailybreadmarket.net/#prednisone-online-pharmacy

Anonymous said...

Hello, lipitor without rx - price of generic lipitor http://www.costoflipitoronline.net/#price-of-generic-lipitor

Anonymous said...

Teen Drug Violence cost of topamax - topiramate online no prescription http://www.topamaxbetterprice.com/

Anonymous said...

2, buy stendra online no prescription - purchase stendra online http://www.stendraonlinecost.com/#purchase-stendra-online , [url=http://www.stendraonlinecost.com/#stendra-online-pharmacy ]stendra online pharmacy [/url]

Anonymous said...

tfv [url=http://www.lipitordiscountonline.net/#price-of-generic-lipitor ]lipitor without rx [/url] - price of generic lipitor - buy lipitor online no prescription http://www.lipitordiscountonline.net/#price-of-lipitor ,

Anonymous said...

ooo!!! Order Topamax - topamax without rx http://www.topamaxbestonline.net/#topamax-cost, [url=http://www.topamaxbestonline.net/#topamax-cost]Topamax Online[/url]

Anonymous said...

2013 zyban online - order zyban http://www.wellbutrinforsaleonline.net/, zyban drug

Anonymous said...

3, [url=http://www.onlinecelebrexdeal.net/]Cheap Celebrex[/url] - Celebrex Online - buy celecoxib http://www.onlinecelebrexdeal.net/.

Anonymous said...

6, Cheap Nexium - Esomeprazole Online - order nexium online http://www.nexiumpricewatch.net/ .