Archive for February, 2015

h1

MLSS 2015 Sydney tutorial

February 23, 2015

This Sydney 2015 MLSS summer school is organised by Edwin Bonilla and held in Sydney Feb 16-25th.  My tutorial is titled “Models for Probability/Discrete Vectors with Bayesian Non-parametric Methods.”  My final version of the slides is here in PDF.

Advertisements
h1

How many clusters?

February 20, 2015

PYP0a_2

Sometimes people think that a Dirichlet Process (DP) can be used to pick the “right number of clusters”.  The following plots done by Monash Matlab whiz Mark Carman show that this has to be done very carefully.

Given N samples, Mark’s first plot shows the expected number of clusters, M, that one would get with a DP using concentration parameter \alpha.  The thing to notice is that the number of clusters is moderately well determined by the concentration parameter.  In fact the mean (expected value) of M is given by:

\alpha \left( \psi(\alpha+N) - \psi(\alpha)\right)

where \psi(\cdot) is the digamma function; for details see the ArXiv report by Marcus Hutter and I, Section 5.3.  Morever, the standard deviation of M is approximately the square root of the mean (for larger concentrations).

So fixing a given concentration parameter means roughly fixing the number of clusters, which increases as the sample size grows.  So with a fixed concentration parameter you cannot really “estimate” the right number of clustersroughly you are fixing the value ahead of time when setting the concentration.

PYP0e_2

Mark’s second plot shows what we do to overcome this.  We have to estimate the concentration parameter as well.  So if we put a prior distribution, an Exponential with parameter \epsilon, on the concentration, we now smear out the previous plots.  So now we show plots for different values of \epsilon.  As you can see, these plots have a much higher variance, which is what you want.  With a given \epsilon, you are still determining the broad range of the number of clusters, but you have a lot more latitude.

In implementation, this means we estimate the concentration \alpha (usually) by sampling it.  If we use Chinese restaurant processes, there is a simple auxiliary variable sampling formula for the concentration (presented in the original HDP paper by Teh et al.).  If we use our blocked table indicator sampling, the posterior on the concentration is log concave so we can use either slice sampling or adaptive rejection sampling (ARS). The implementation is moderately simple, and it works well.  However, it does mean now your algorithm will expend more time as it has to try and find the right concentration as well.

h1

Video of a Lecture

February 20, 2015

The talk I gave at JSI (Jozef Stefan Institute in Ljublana) on 14th Jan 2015 was recorded.  The group here, with Dunja Mladenić and Marko Grobelnik, are expert in areas like Data Science and Text Mining, but they’re not into Bayesian non-parametrics, so in this version of the talk I mostly avoided the statistical details and talked more about what we did and why.  The talk is up on Video Lectures.  The original PDF of the EU talk sequence was on this post.

h1

Lancelot James’ general theory for IBPs

February 9, 2015

When I visited Hong Kong last December, Lancelot James gave me a tutorial on his generalised theory of Indian Buffet Processes, which includes all sorts of interesting extensions that are starting to appear in the machine learning community.

The theory is written up in his recent Arxiv report, “Poisson Latent Feature Calculus for Generalized Indian Buffet Processes.”  I’m a big fan of Poisson processes, so its good to see the general theory written up, including a thorough posterior analysis.  In my recent tutorial (given at Aalto) I’ve included a simplified version of this with the key marginal posterior for a few well known cases.  Piyush Rai from Duke also has great summary slides as part of Lawrence Carin’s reading group at Duke.

Within this framework, you can now handle variations like the Beta-negative-Binomial Process of Mingyuan Zhou and colleagues presented at AI Stats 2012 (Arxiv version here), for which many different variations can be developed.

I’ve presented the general case as an application of improper priors because the Poisson process theory in this case is mostly just formalising their use for developing “infinite parameter vectors”.  Lancelot uses his Poisson Process Partition Calculus to develop a clean, general theory.  For me it was interesting to see that his results also apply to multivariate and auxiliary parameter vectors:

  • an infinite vector of Gamma scales \vec\beta created using a Gamma process can have an auxiliary vector of Gamma shapes \vec\alpha sampled pointwise, so \alpha_k is associated with \beta_k
    • most of the scales are infinitesimal
    • most of the shapes are not infinitesimal
  • an infinite set of K+1-dimensional multinomial’s can be created where the first K dimensions are generated using a Dirichlet style Poisson rate:
    • most of the multinomials will have infinitesimal values in the first K dimensions
    • the Beta process corresponds to the case where K=1