NICE_GP_HIK_Core/FastMinKernel.h

/** 
* @file FastMinKernel.h
* @brief Efficient GPs with HIK for classification by regression (Interface)
* @author Alexander Freytag
* @date 06-12-2011 (dd-mm-yyyy)
*/
#ifndef FASTMINKERNELINCLUDE
#define FASTMINKERNELINCLUDE

#include <iostream>

#include <core/vector/MatrixT.h>
#include <core/vector/SparseVectorT.h>
#include <core/vector/VVector.h>
#include <core/basics/Exception.h>
#include "core/basics/Persistent.h"

#include "FeatureMatrixT.h"
#include "Quantization.h"
#include "gp-hik-core/parameterizedFunctions/ParameterizedFunction.h"

namespace NICE {


/** 
 * @class FastMinKernel
 * @brief Efficient GPs with HIK for classification by regression
 * @author Alexander Freytag
 */  
  
  /** interface to FastMinKernel implementation*/
  class FastMinKernel : NICE::Persistent
  {

    protected:
      /** number of examples */
      int n;

      /** dimension of feature vectors */
      int d; 

      /** noise added to the diagonal of the kernel matrix */
      double noise;
      
      /** sorted matrix of features (sorted along each dimension) */
      NICE::FeatureMatrixT<double> X_sorted;
      
      //! verbose flag for output after calling the restore-function
      bool verbose;
      //! debug flag for output during debugging
      bool debug;      

      /** 
      * @brief Set number of examples
      * @author Alexander Freytag
      * @date 07-12-2011 (dd-mm-yyyy)
      */
      void set_n(const int & _n){n = _n;};
      
      /** 
      * @brief Set number of dimensions
      * @author Alexander Freytag
      * @date 07-12-2011 (dd-mm-yyyy)
      */
      void set_d(const int & _d){d = _d;};     

      /** 
      * @brief Prepare the efficient HIK-computations part 1: order the features in each dimension and save the permutation. Pay attention: X is of dim n x d, where as X_sorted is of dimensionality d x n!
      * @author Alexander Freytag
      * @date 07-12-2011 (dd-mm-yyyy)
      */
      void hik_prepare_kernel_multiplications(const std::vector<std::vector<double> > & X, NICE::FeatureMatrixT<double> & X_sorted, const int & _dim = -1);
      
      void hik_prepare_kernel_multiplications ( const std::vector< NICE::SparseVector * > & X, NICE::FeatureMatrixT<double> & X_sorted, const bool & dimensionsOverExamples, const int & _dim = -1);
      
      void randomPermutation(NICE::Vector & permutation, const std::vector<int> & oldIndices, const int & newSize) const;
      
      enum ApproximationScheme{ MEDIAN = 0, EXPECTATION=1};
      ApproximationScheme approxScheme;

    public:

      //------------------------------------------------------
      // several constructors and destructors
      //------------------------------------------------------
      
      /** 
      * @brief dummy constructor
      * @author Alexander Freytag
      * @date 20-04-2012 (dd-mm-yyyy)
      */
      FastMinKernel();      
      
      /** 
      * @brief initialize with some data
      * @author Alexander Freytag
      * @date 06-12-2011 (dd-mm-yyyy)
      */
      FastMinKernel( const std::vector<std::vector<double> > & X, const double noise , const bool _debug = false, const int & _dim = -1);

      
      /**
      * @brief Just another sparse data structure
      *
      * @param X vector of sparse vector pointers
      * @param noise GP noise
      */
      FastMinKernel( const std::vector< SparseVector * > & X, const double noise, const bool _debug = false, const bool & dimensionsOverExamples=false, const int & _dim = -1);

#ifdef NICE_USELIB_MATIO
      /**
      * @brief intialize with some data given in a matlab-sparse struct and restricted with an example index
      *
      * @param X matlab-struct containing the feature vectors
      * @param noise additional noise variance of the labels
      * @param examples set of indices to include
      */
      FastMinKernel ( const sparse_t & X, const double noise, const std::map<int, int> & examples, const bool _debug = false , const int & _dim = -1);
#endif

      /** 
      * @brief Default destructor
      * @author Alexander Freytag
      * @date 06-12-2011 (dd-mm-yyyy)
      */
      ~FastMinKernel();

      //------------------------------------------------------
      // several get and set methods including access operators
      //------------------------------------------------------
      
      
      void setApproximationScheme(const ApproximationScheme & _approxScheme = MEDIAN) {approxScheme = _approxScheme;};
      
      virtual void setApproximationScheme(const int & _approxScheme = 0);
      
      /** 
      * @brief Get number of examples
      * @author Alexander Freytag
      * @date 07-12-2011 (dd-mm-yyyy)
      */
      int get_n() const {return n;};
      
      /** 
      * @brief Get number of dimensions
      * @author Alexander Freytag
      * @date 07-12-2011 (dd-mm-yyyy)
      */
      int get_d() const {return d;};

      /** 
      * @brief Computes the ratio of sparsity across the matrix
      * @author Alexander Freytag
      * @date 11-01-2012 (dd-mm-yyyy)
      */
      double getSparsityRatio(){return X_sorted.computeSparsityRatio();};
      
      /** set verbose flag used for restore-functionality*/
      void setVerbose( const bool & _verbose);
      bool getVerbose( ) const;  
      
      /** set debug flag used for debug output*/
      void setDebug( const bool & _debug);
      bool getDebug( ) const;        
      
      //------------------------------------------------------
      // high level methods
      //------------------------------------------------------
      
      /**
      * @brief apply a parameterized function to the feature matrix
      * @author Alexander Freytag
      * @date 04-05-2012 (dd-mm-yyyy)
      *
      * @param pf the parameterized function (optional), if not given, nothing will be done
      */         
      void applyFunctionToFeatureMatrix ( const NICE::ParameterizedFunction *pf = NULL );
          
      /** 
      * @brief  Prepare the efficient HIK-computations part 2: calculate the partial sum for each dimension. Explicitely exploiting sparsity!!! Pay attention: X_sorted is of dimensionality d x n!
      * @author Alexander Freytag
      * @date 17-01-2012 (dd-mm-yyyy)
      */
      void hik_prepare_alpha_multiplications(const NICE::Vector & alpha, NICE::VVector & A, NICE::VVector & B) const;
            
      /**
      * @brief Computing K*alpha with the minimum kernel trick, explicitely exploiting sparsity!!!
      * @author Alexander Freytag
      * @date 17-01-2012 (dd-mm-yyyy)
      */
      void hik_kernel_multiply(const NICE::VVector & A, const NICE::VVector & B, const NICE::Vector & alpha, NICE::Vector & beta) const;
      void hik_kernel_multiply_fast(const double *Tlookup, const Quantization & q, const NICE::Vector & alpha, NICE::Vector & beta) const;

      /**
      * @brief Computing k_{*}*alpha using the minimum kernel trick and exploiting sparsity of the feature vector given
      *
      * @author Alexander Freytag
      * @date 20-01-2012 (dd-mm-yyyy)
      * @param A pre-computation matrix (VVector) (use the prepare method) 
      * @param B pre-computation matrix (VVector)
      * @param xstar new feature vector (SparseVector)
      * @param beta result of the scalar product
      * @param pf optional feature transformation
      */
      void hik_kernel_sum(const NICE::VVector & A, const NICE::VVector & B, const NICE::SparseVector & xstar, double & beta, const ParameterizedFunction *pf = NULL ) const;
      
      /**
      * @brief compute beta = k_*^T * alpha by using a large lookup table created by hik_prepare_alpha_multiplications_fast
      * @author Erik Rodner
      *
      * @param Tlookup large lookup table calculated by hik_prepare_alpha_multiplications_fast
      * @param q Quantization object
      * @param xstar feature vector (indirect k_*)
      * @param beta result of the calculation
      */
      void hik_kernel_sum_fast(const double* Tlookup, const Quantization & q, const NICE::Vector & xstar, double & beta) const;
      void hik_kernel_sum_fast(const double *Tlookup, const Quantization & q, const NICE::SparseVector & xstar, double & beta) const;

      /**
      * @brief compute lookup table for HIK calculation using quantized signals and prepare for K*alpha or k_*^T * alpha computations
      * @author Erik Rodner
      *
      * @param alpha coefficient vector
      * @param A pre-calculation array computed by hik_prepare_alpha_multiplications
      * @param B pre-calculation array computed by hik_prepare_alpha_multiplications
      * @param q Quantization
      *
      * @return C standard vector representing a q.size()*n double matrix and the lookup table T. Elements can be accessed with
      * T[dim*q.size() + j], where j is a bin entry corresponding to quantization q.
      */
      double *hik_prepare_alpha_multiplications_fast(const NICE::VVector & A, const NICE::VVector & B, const Quantization & q, const ParameterizedFunction *pf = NULL ) const;
      
      /**
      * @brief compute lookup table for HIK calculation using quantized signals and prepare for K*alpha or k_*^T * alpha computations
      * @author Alexander Freytag
      *
      * @param alpha coefficient vector
      * @param q Quantization
      * @param pf ParameterizedFunction to change the original feature values
      *
      * @return C standard vector representing a q.size()*n double matrix and the lookup table T. Elements can be accessed with
      * T[dim*q.size() + j], where j is a bin entry corresponding to quantization q.
      */
      double* hikPrepareLookupTable(const NICE::Vector & alpha, const Quantization & q, const ParameterizedFunction *pf = NULL) const;

      /**
      * @brief update the lookup table for HIK calculation using quantized signals and prepare for K*alpha or k_*^T * alpha computations
      * @author Alexander Freytag
      *
      * @param T previously computed LUT, that will be changed
      * @param alphaNew new value of alpha at index idx
      * @param alphaOld old value of alpha at index idx
      * @param idx index in which alpha changed
      * @param q Quantization
      * @param pf ParameterizedFunction to change the original feature values
      */
      void hikUpdateLookupTable(double * T, const double & alphaNew, const double & alphaOld, const int & idx, const Quantization & q, const ParameterizedFunction *pf ) const;

      /**
      * @brief return a reference to the sorted feature matrix
      */
      FeatureMatrix & featureMatrix(void) { return X_sorted; };
      const FeatureMatrix & featureMatrix(void) const { return X_sorted; };
      
      /**
       * @brief solve the linear system K*alpha = y with the minimum kernel trick based on the algorithm of Wu (Wu10_AFD)
       * @note method converges slowly for large scale problems and even for normal scale :(
       * @author Paul Bodesheim
       * 
       * @param y right hand side of linear system
       * @param alpha final solution of the linear system
       * @param q Quantization
       * @param pf ParameterizedFunction to change the original feature values
       * @param useRandomSubsets true, if the order of examples in each iteration should be randomly sampled
       * @param maxIterations maximum number of iterations
       * @param sizeOfRandomSubset nr of Elements that should be randomly considered in each iteration (max: y.size())
       * @param minDelta minimum difference between two solutions alpha_t and alpha_{t+1} (convergence criterion)
       * 
       * @return C standard vector representing a q.size()*n double matrix and the lookup table T. Elements can be accessed with
       * T[dim*q.size() + j], where j is a bin entry corresponding to quantization q.
       **/
      double *solveLin(const NICE::Vector & y, NICE::Vector & alpha, const Quantization & q, const ParameterizedFunction *pf = NULL, const bool & useRandomSubsets = true, uint maxIterations = 10000, const int & _sizeOfRandomSubset = (-1), double minDelta = 1e-7, bool timeAnalysis = false) const;


      //! set the noise parameter
      void setNoise ( double noise ) { this->noise = noise; }

      //! get the current noise parameter
      double getNoise (void) const { return noise; }
      
      double getFrobNormApprox();
      
      
      /** 
      * @brief  Prepare the efficient HIK-computations for the squared kernel vector |k_*|^2 : calculate the partial squared sums for each dimension.
      * @author Alexander Freytag
      * @date 10-04-2012 (dd-mm-yyyy)
      */
      void hikPrepareKVNApproximation(NICE::VVector & A) const;
      
      /** 
      * @brief  Compute lookup table for HIK calculation of |k_*|^2 assuming quantized test samples. You have to run hikPrepareSquaredKernelVector before
      * @author Alexander Freytag
      * @date 10-04-2012 (dd-mm-yyyy)
      * 
      * @param A pre-calculation array computed by hikPrepareSquaredKernelVector
      * @param q Quantization
      * @param pf Parameterized Function to efficiently apply a function to the underlying data
      *
      * @return C standard vector representing a q.size()*d double matrix and the lookup table T. Elements can be accessed with
      * T[dim*q.size() + j], where j is a bin entry corresponding to quantization q.
      */
      double * hikPrepareKVNApproximationFast(NICE::VVector & A, const Quantization & q, const ParameterizedFunction *pf = NULL ) const;
      
      /**
      * @brief Compute lookup table for HIK calculation of |k_*|^2 assuming quantized test samples ( equals hikPrepareSquaredKernelVector + hikPrepareSquaredKernelVectorFast, but is faster). Approximation does not considere mixed terms between dimensions.
      * @author Alexander Freytag
      * @date 10-04-2012 (dd-mm-yyyy)
      *
      * @param q Quantization
      * @param pf ParameterizedFunction to change the original feature values
      *
      * @return C standard vector representing a q.size()*d double matrix and the lookup table T. Elements can be accessed with
      * T[dim*q.size() + j], where j is a bin entry corresponding to quantization q.
      */
      double* hikPrepareLookupTableForKVNApproximation(const Quantization & q, const ParameterizedFunction *pf = NULL) const;
      
      /**
      * @brief Approximate norm = |k_*|^2 using the minimum kernel trick and exploiting sparsity of the given feature vector. Approximation does not considere mixed terms between dimensions.
      * @author Alexander Freytag
      * @date 10-04-2012 (dd-mm-yyyy)
      * 
      * @param A pre-computation matrix (VVector) (use the prepare method) 
      * @param xstar new feature vector (SparseVector)
      * @param norm result of the squared norm approximation
      * @param pf optional feature transformation
      */
      void hikComputeKVNApproximation(const NICE::VVector & A, const NICE::SparseVector & xstar, double & norm, const ParameterizedFunction *pf = NULL ) ;
      
      /**
      * @brief Approximate norm = |k_*|^2 using a large lookup table created by hikPrepareSquaredKernelVector and hikPrepareSquaredKernelVectorFast or directly using hikPrepareLookupTableForSquaredKernelVector. Approximation does not considere mixed terms between dimensions.
      * @author Alexander Freytag
      * @date 10-04-2012 (dd-mm-yyyy)
      *
      * @param Tlookup large lookup table
      * @param q Quantization object
      * @param xstar feature vector (indirect k_*)
      * @param norm result of the calculation
      */
      void hikComputeKVNApproximationFast(const double *Tlookup, const Quantization & q, const NICE::SparseVector & xstar, double & norm ) const;

      /**
      * @brief Compute the kernel vector k_* between training examples and test example. Runtime. O(n \times D). Exploiting sparsity
      * @author Alexander Freytag
      * @date 13-04-2012 (dd-mm-yyyy)
      *
      * @param xstar feature vector
      * @param kstar kernel vector
      */      
      void hikComputeKernelVector( const NICE::SparseVector & xstar, NICE::Vector & kstar) const;
      
      /** Persistent interface */
      virtual void restore ( std::istream & is, int format = 0 );
      virtual void store ( std::ostream & os, int format = 0 ) const; 
      virtual void clear ();
      
      // ----------------- INCREMENTAL LEARNING METHODS -----------------------
      
      /**
      * @brief Add a new example to the feature-storage. You have to update the corresponding variables explicitely after that.
      * @author Alexander Freytag
      * @date 25-04-2012 (dd-mm-yyyy)
      *
      * @param _v new feature vector
      */       
      void addExample(const NICE::SparseVector & _v, const ParameterizedFunction *pf = NULL);
      /**
      * @brief Add a new example to the feature-storage. You have to update the corresponding variables explicitely after that.
      * @author Alexander Freytag
      * @date 25-04-2012 (dd-mm-yyyy)
      *
      * @param _v new feature vector
      */       
      void addExample(const std::vector<double> & _v, const ParameterizedFunction *pf = NULL);
      
      /**
      * @brief Updates A and B matrices for fast kernel multiplications and kernel sums. You need to compute the new alpha value and run addExample first!
      * @author Alexander Freytag
      * @date 25-04-2012 (dd-mm-yyyy)
      *
      * @param _v new feature vector
      * @param alpha new alpha value for the corresponding feature
      * @param A precomputed matrix A which will be updated accordingly
      * @param B precomputed matrix B which will be updated accordingly
      * @param pf optional feature transformation
      */       
      void updatePreparationForAlphaMultiplications(const NICE::SparseVector & _v, const double & alpha, NICE::VVector & A, NICE::VVector & B, const ParameterizedFunction *pf = NULL) const;
      /**
      * @brief Updates LUT T for very fast kernel multiplications and kernel sums. You need to compute the new alpha value and run addExample first!
      * @author Alexander Freytag
      * @date 26-04-2012 (dd-mm-yyyy)
      *
      * @param _v new feature vector
      * @param alpha new alpha value for the corresponding feature
      * @param T precomputed lookup table, which will be updated
      * @param q quantization object to quantize possible test samples
      * @param pf optional feature transformation
      */       
      void updateLookupTableForAlphaMultiplications(const NICE::SparseVector & _v, const double & alpha, double * T, const Quantization & q, const ParameterizedFunction *pf = NULL) const;
      
      /**
      * @brief Updates matrix A for approximations of the kernel vector norm. You need to run addExample first!
      * @author Alexander Freytag
      * @date 26-04-2012 (dd-mm-yyyy)
      *
      * @param _v new feature vector
      * @param A precomputed matrix A which will be updated accordingly
      * @param pf optional feature transformation
      */       
      void updatePreparationForKVNApproximation(const NICE::SparseVector & _v, NICE::VVector & A, const ParameterizedFunction *pf = NULL) const;
      /**
      * @brief Updates LUT T for fast approximations of the kernel vector norm. You need to run addExample first!
      * @author Alexander Freytag
      * @date 26-04-2012 (dd-mm-yyyy)
      *
      * @param _v new feature vector
      * @param T precomputed lookup table, which will be updated
      * @param q quantization object to quantize possible test samples
      * @param pf optional feature transformation
      */       
      void updateLookupTableForKVNApproximation(const NICE::SparseVector & _v, double * T, const Quantization & q, const ParameterizedFunction *pf = NULL) const;

  };

} // namespace

#endif