MultipleImportanceSampler Class Reference

"Optimally" combines samples taken from multiple sampling distributions with respect to a function, 'f', whose value we are trying to integate over a given domain,'D'. D is assumed to be the union of the domains of the individual underlying distributions. More...

#include <MultipleImportanceSampler.h>

Inheritance diagram for MultipleImportanceSampler:

PropertyMap

List of all members.

Public Member Functions

Constructors
Note:
ni, n, and N use the same notation from Veach and Guibas' paper


 MultipleImportanceSampler (const SamplerList &samplers, unsigned ni=1)
 MultipleImportanceSampler (const SamplerList &samplers, const IntList &ni)
virtual ~MultipleImportanceSampler ()
Initialization
virtual void init ()
 Initializes the weight computation function based on this class' PropertyMap.
Main usage interface
virtual void sample (WeightedEventList &results)
 Generates ni samples from each sampler pi in { p1,p1,...,pn }, along with sample weights, and stores the N weighted events in the out parameter results.

Protected Types

typedef void(MultipleImportanceSampler::* ComputeWeightFunc )(WeightedEvent &event, unsigned i)

Protected Member Functions

virtual void _computeWeightBalance (WeightedEvent &event, unsigned i)
virtual void _computeWeightCutoff (WeightedEvent &event, unsigned i)
virtual void _computeWeightPower (WeightedEvent &event, unsigned i)
virtual void _computeWeightMaximum (WeightedEvent &event, unsigned i)

Protected Attributes

SamplerList m_samplers
 list of distributions from which to draw samples from
unsigned m_N
 number of sampling distributions in m_samplers
unsigned m_n
 total number of samples
IntList m_ni
 sampler-by-sampler breakdown of number of samples
ComputeWeightFunc m_computeWeightFunc
 weight computation for individual samples (see init())
union {
   real_t   m_beta
 input to power heuristic
   real_t   m_alpha
 input to alpha heuristic
}; 

Detailed Description

"Optimally" combines samples taken from multiple sampling distributions with respect to a function, 'f', whose value we are trying to integate over a given domain,'D'. D is assumed to be the union of the domains of the individual underlying distributions.

Author:
Travis Fischer (fisch0920@gmail.com)

Matthew Jacobs (jacobs.mh@gmail.com)

Date:
Fall 2008
The goal of importance sampling is to reduce variance while approximating the integral of f over D by drawing random samples across D according to some probability density function p which is proportional to f. By drawing more samples from the regions where f is large (concentrating our sampling effort on the "important" regions of the domain), and similarly drawing less samples from the regions where f is small (focusing less effort in the "unimportant" regions of the domain), the variance of our estimate overall is reduced, as long as we compensate for our uneven sampling rate by weighting each sample by the inverse probability with which it was sampled. Think of importance sampling this way: if we have only one shot at sampling f (only one sample because of very limited 'resources'), we would like to concentrate that one sample in the region of the domain that will contribute the most to the value of the integral we are ultimately trying to approximate. By biasing our sampling technique to weigh "important", "large" regions of the domain with respect to f, we get more bang for our buck and correspondingly end up with a much lower variance estimator. If we were to instead naively sample uniformly across the domain, D, there's a good chance that our one, precious sample would be wasted by sampling a location, x, that is unimportant, where f(x) is relatively small. A perfect estimator would take samples from a distribution p, across D, whose probabilities were distributed directly proportional to f. This is, however, unreasonable, since if we knew the exact values of f across D, we wouldn't have to approximate the integral of f over D in the first place. We can, however, still greatly reduce variance by using knowledge about f to guide our sampling process and instead, sample according to either an approximation of f, or possibly according to some factorization of f into separate functions f1, f2, etc, some of which may be easier to draw samples from. As long as our sampling process attempts to draw samples from areas in D where f is "known" to be large and avoids regions in D where f is "known" to be small, our sampler will be more efficient and our overall variance will be reduced.

Multiple importance sampling generalizes this one step further. Say we have n different sampling distributions p1,p2,...,pn, and our function f, may be separated or factored into several functions, f1,f2,...fm. By designing different sampling distributions to focus on one or more of the simpler functions, fi, we can combine samples from the different sampling distributions, each of which may be better at estimating some important component (subfunction) of f, we can gain a much better final sampling distribution, p, which captures all of the relevant aspects of f in an intuitive, modularized manner (where each pi can be tuned towards capturing a different fi). How to go about combining samples from n distributions, all or some of which may be good or bad estimators of f, is a very difficult problem in this generic of a setting. Veach and Guibas (1995) showed several provably good ways of optimally combining sampling techniques for Monte Carlo integration. This class implements their multiple importance sampling model, including several heuristics they gave which attempt to combine samples in slightly different ways that may work better in one situation versus another. These include the balance, cutoff, power, and maximum heuristics.

Note:
for more information and theoretical details, see Veach and Guibas. Optimally Combining Sampling Techniques for Monte Carlo Rendering. In Proceedings of the 22nd Annual Conference on Computer Graphics and interactive Techniques S.G. Mair and R. Cook, Eds. SIGGRAPH '95. ACM, New York, NY, 419-428.
See also:
http://www-graphics.stanford.edu/papers/combine/

Definition at line 85 of file MultipleImportanceSampler.h.

Member Typedef Documentation

typedef void(MultipleImportanceSampler::* MultipleImportanceSampler::ComputeWeightFunc)(WeightedEvent &event, unsigned i) [protected]

Constructor & Destructor Documentation

MultipleImportanceSampler::MultipleImportanceSampler ( const SamplerList &  samplers,
unsigned  ni = 1 
) [inline]

Parameters:
ni is the constant number of samples to take per sampler

Definition at line 95 of file MultipleImportanceSampler.h.

MultipleImportanceSampler::MultipleImportanceSampler ( const SamplerList &  samplers,
const IntList &  ni 
) [inline]

Parameters:
ni holds the variable number of samples to take per sampler

Definition at line 104 of file MultipleImportanceSampler.h.

virtual MultipleImportanceSampler::~MultipleImportanceSampler (  )  [inline, virtual]

Definition at line 116 of file MultipleImportanceSampler.h.

Member Function Documentation

void MultipleImportanceSampler::init (  )  [virtual]

Initializes the weight computation function based on this class' PropertyMap.

Note:
defaults to balance heuristic

Definition at line 76 of file MultipleImportanceSampler.cpp.

void MultipleImportanceSampler::sample ( WeightedEventList &  results  )  [virtual]

Generates ni samples from each sampler pi in { p1,p1,...,pn }, along with sample weights, and stores the N weighted events in the out parameter results.

Note:
'results' is first cleared before adding any samples

Definition at line 96 of file MultipleImportanceSampler.cpp.

void MultipleImportanceSampler::_computeWeightBalance ( WeightedEvent &  event,
unsigned  i 
) [protected, virtual]

Definition at line 111 of file MultipleImportanceSampler.cpp.

void MultipleImportanceSampler::_computeWeightCutoff ( WeightedEvent &  event,
unsigned  i 
) [protected, virtual]

Definition at line 131 of file MultipleImportanceSampler.cpp.

void MultipleImportanceSampler::_computeWeightPower ( WeightedEvent &  event,
unsigned  i 
) [protected, virtual]

Definition at line 158 of file MultipleImportanceSampler.cpp.

void MultipleImportanceSampler::_computeWeightMaximum ( WeightedEvent &  event,
unsigned  i 
) [protected, virtual]

Definition at line 174 of file MultipleImportanceSampler.cpp.

Member Data Documentation

SamplerList MultipleImportanceSampler::m_samplers [protected]

list of distributions from which to draw samples from

Definition at line 163 of file MultipleImportanceSampler.h.

unsigned MultipleImportanceSampler::m_N [protected]

number of sampling distributions in m_samplers

Definition at line 166 of file MultipleImportanceSampler.h.

unsigned MultipleImportanceSampler::m_n [protected]

total number of samples

Definition at line 169 of file MultipleImportanceSampler.h.

IntList MultipleImportanceSampler::m_ni [protected]

sampler-by-sampler breakdown of number of samples

Definition at line 172 of file MultipleImportanceSampler.h.

ComputeWeightFunc MultipleImportanceSampler::m_computeWeightFunc [protected]

weight computation for individual samples (see init())

Definition at line 175 of file MultipleImportanceSampler.h.

real_t MultipleImportanceSampler::m_beta

input to power heuristic

Definition at line 179 of file MultipleImportanceSampler.h.

real_t MultipleImportanceSampler::m_alpha

input to alpha heuristic

Definition at line 182 of file MultipleImportanceSampler.h.

union { ... } [protected]

The documentation for this class was generated from the following files:
Generated on 28 Feb 2009 for Milton by doxygen 1.5.6