SlideShare uma empresa Scribd logo
1 de 20
Baixar para ler offline
Brief Paper Review

B.Balaji (2009)
“Nonlinear filtering and quantum physics
 A Feynman path integral perspective”


Kohta Ishikawa
Oct 30. 2011   3


                                           1
2
xk = f (xk   1 , tk 1 )xk 1   + e(xk   1 , tk 1 )   System Model
yk = h(xk , tk ) + wk Observation Model



dx(t) = f (x(t), t)dt + e(x(t), t)dw(t) System Model(SDE)
y(tk ) = h(x(tk ), tk ) + wk Observation Model



dx(t) = f (x(t), t)dt + e(x(t), t)dv(t) System Model(SDE)
dy(t) = h(x(t), t)dt + dw(t) Observation Model(SDE)

                                                                   3
m
                                      X
     d (t, x) = LY (t, x)dt +                hi (x) (t, x)dyi (t)
                                       i=1


                 n
                X @                          n
                                            X @ 2 (t, x)     m
                                                             X
                                          1                1
LY (t, x) =             (fi (x) (t, x)) +                            h2 (x) (t, x)
                                                                      i
                i=1
                    @xi                   2 i=1 @x2 i      2   i=1

 (0, x) =     0 (x)




              dy(t)

                                                                                     4
dy(t)
                                m
                                                             !
                                X
     u(t, x) = exp                        hi (x)yi (t)             (t, x)
                                i=1

Then, DMZ equation translates to
                        0                                                       1
               n
               X @ 2 u(t, x)        n
                                    X                    m
                                                         X
 @u(t, x)    1                        @ fi (x) +            @hj (x) A @u(t, x)
          =                     +                    yj (t)
   @t       2   i=1
                        @x2
                          i       i=1            j=1
                                                             @xi        @xi
             ✓X
              n                   m
                                  X                    m
                                                       X
                      @fi (x)   1                    1
                              +           h2 (x)
                                           i                  yi (t) hi (x)
              i=1
                       @xi      2   i=1
                                                     2i=1
             m n
             XX                                      m
                                                     X X  n                                  ◆
                                @hi (x)            1                            @hi (x) @hj (x)
         +         yi (t)fj (x)                                    yi (t)yj (t)                 u(t, x)
           i=1 j=1
                                 @xj               2 i,j=1                       @xk     @xk
                                                             k=1

 u(0, x) =   0 (x)


                                                                                                          5
⇢
             y(⌧l )   post-measurement form
y(t) !
             y(⌧l 1 ) pre-measurement form




                                              6
⌧l    1     t  ⌧l
                                                                                          {⌧0 , ⌧1 , · · · }
                                          0                                      1
                n                n            m
@ul (t, x)   1 X @ 2 ul (t, x) X @           X            @hj (x) A @ul (t, x)
           =             2    +     fi (x) +     yj (⌧l )
   @t        2 i=1   @xi        i=1          j=1
                                                           @xi        @xi
            ✓X
             n                      m
                                    X                m
                                                     X
                    @fi (x)       1                1
                            +            h2 (x)
                                          i                 yi (⌧l ) hi (x)
             i=1
                     @xi      2    i=1
                                                   2i=1
            m n
            XX                                      m
                                                    X X n                                        ◆
                                   @hi (x)        1                                @hi (x) @hj (x)
          +         yi (⌧l )fj (x)                                yi (⌧l )yj (⌧l )                 ul (t, x)
            i=1 j=1
                                    @xj           2 i,j=1                           @xk     @xk
                                                            k=1


ul (⌧l , x) = ul   1 (⌧l 1 , x)




                                                                                                               7
m
                                           !
                       X                                                                        (t, x)
ul (t, x) = exp
˜                             yi (⌧l )hi (x) ul (t, x)
                        i=1

                     n
                     X @ 2 ul (t, x)       n
                                           X
@ ul (t, x)
  ˜                1       ˜                          @ ul (t, x)
                                                        ˜
            =                                  fi (x)
    @t        2      i=1
                              @x2
                                i          i=1
                                                         @xi
                                             n                       m
                                                                                  !
                                             X @fi (x)             1 X
                                                               +             h2 (x) ul (t, x)
                                                                              i     ˜
                                             i=1
                                                       @xi         2   i=1

                               m
                                                   !
                               X
 ul (⌧l
 ˜        1 , x)   = exp             yi (⌧l )hi (x) ul       1 (⌧l 1 , x)
                               i=1




                                          yi (⌧l   1)

                                                                                                         8
9
Stochastic Process


probability density of state                              path measure,
                                                          functional differentiation




      Fokker-Planck equation                        Path Integral representation

                                 green’s function




                                                                                       10
SDE for continuous filtering
  xi (t) = fi (x(t)) + vi (t)
   ˙
  yi (t) = hi (x(t)) + wi (t)
   ˙

Probability density
                 Z
 P (t, x, y|t0 , x0 .y0 ) =   [d⇢(v(t))][d⇢(w(t))]

                                              (x(t)
                                               ˙      f (x(t))       v(t))   n
                              ⇥ [dx(t)]                                          (x(t)
                                                                                  ˙       f (x(t))   v(t))
                                                        x(t)
                                              (y(t)
                                               ˙      h(x(t))       w(t))     m
                              ⇥ [dy(t)]                                           (y(t)
                                                                                   ˙      h(y(t))    w(t))
                                                        y(t)
                                  n                            m
                              ⇥       (x(t)     x)|x(t0 )=x0       (y(t)     y)|y(t0 ) = y0
                                                                                                         11
!
                                                      1     XZ t
                                                            n
     [d⇢(v(t))] = [Dv(t)] exp
                                                     2~v
                                                                       vi (t)2 dt           ~v , ~w
                                                            i=1 t0
                                                                                        !
                                                       1     XZ t
                                                              m
     [d⇢(w(t))] = [Dw(t)] exp                                              vi (t)2 dt
                                                      2~w      i=1    t0

                               Z    y(t)=y     Z    x(t)=x
P (t, x, y|t0 , x0 .y0 ) =                                     [Dv(t)][Dw(t)][Dx(t)][Dy(t)]
                                   y(t0 )=y0       x(t0 )=x0
                                                                                                                                      !
                                       1     XZ t
                                             n                             XZ t
                                                                           n
                                                                                      @fi (x(t))         1         XZ t
                                                                                                                   m
                                                            2                                                                  2
                    ⇥ exp                                  vi (t)dt                              dt                           wi dt
                                      2~v    i=1      t0                   i=1   t0     @xi             2~w        i=1   t0
                           n
                       ⇥    (x(t) f (x(t)) v(t)) m (y(t) h(y(t)) w(t))
                             ˙                        ˙
                             Z y(t)=y Z x(t)=x
                           =                   [Dx(t)][Dy(t)] exp ( S)
                                   y(t0 )=y0       x(t0 )=x0
         Z         "       n                                          n                               m
                                                                                                                                      #
   1          t
                     1     X                                          X       @fi (x(t))    1         X
                                                                2
S=                dt               (xi (t)
                                    ˙          fi (x(t))) +                              +                  ( yi
                                                                                                              ˙      hi (x(t)))2
   2         t0      ~v    i=1                                         i=1
                                                                                @xi        ~w         i=1

                                                                                                                                      12
measurement term in the action S
          Z    ti            m
                             X                                    Z    ti            m
                                                                                     X⇥
  1                                                        1                                                                           ⇤
                        dt         ( yi
                                     ˙    hi (x(t))) =                          dt         ˙2
                                                                                           yi (t)   +   h2 (x(t))
                                                                                                         i          2hi (x(t))yi (t)
                                                                                                                              ˙
 2~w          ti    1        i=1
                                                          2~w         ti    1        i=1




relevant term
     Z    ti            m
                        X                          ⇢      1
                                                              Pm
 1                                                       ~w    j=1 hj ([x(ti ) + x(ti
                                                              Pm                                    1 )]/2)[yj (tj )   yj (tj 1 )]
                   dt         hj (x(t))yj (t) ⇠
                                       ˙                  1
~w       ti    1        j=1                              ~w    j=1 hj ([x(ti ) + x(ti               1 )]/2)[yj (tj 1 )    yj (tj 2 )]




                                                                                                                                       13
leads to approximated probability density
P (ti , xi , yi |ti    1 , xi 1 , yi 1 )
                                                ˜
                                              ⇠ P (ti , xi |ti 1 , xi 1 )
                                   ⇣         Pm                                                     ⌘
                       ⇢              1
                            exp               j=1 hj ([x(ti ) + x(ti 1 )]/2)[yj (tj )   yj (tj 1 )]
                   ⇥               ⇣ ~w      Pm                                                        ⌘
                                        1
                            exp        ~w     j=1 hj ([x(ti ) + x(ti 1 )]/2)[yj (tj 1 )    yj (tj 2 )]

                                        Z    x(ti )=xi
  ˜
  P (ti , xi |ti    1 , xi 1 )     =                              [Dx(t)] exp( S(ti       1 , ti ))
                                            x(ti   1 )=xi     1

                       Z            "        n                                   n                       m
                                                                                                                           #
                  1         ti
                                      1      X                                   X     @fi (x(t))    1   X
                                                                          2
  S(ti 1 , ti ) =                  dt               (xi (t)
                                                     ˙            fi (x(t))) +                    +            h2 (x(t))
                                                                                                                i
                  2        ti    1
                                      ~v      i=1                                i=1
                                                                                         @xi        ~w   i=1

˜
P (t, x|ti         1 , xi 1 )




                                                                                                                           14
Z    x(t)=x                   ✓                   ◆
˜                                                              1
P (t, x|t0 , x0 ) =                        [Dx(t)] exp            S
                               x(t0 )=x0                       ~v
        Z          "   n                                                                 n                                   m
                                                                                                                                               #
   1         t         X                                        @fi (x(t))    ~v         X                                   X
S=        dt             + [x2 (t)
                            ˙i         ˙    fi2 (x)
                                     2xi (t)fi (x(t))] + ~v                +                                                       h2 (x(t))
                                                                                                                                    i
   2   t0    i=1                                            i=1
                                                                   @xi        ~w                                             i=1
      Z t "X  n                           n
                                         X @fi (x(t))           m
                                                                             #
    1                                                       ~v X 2
  =       dt     [x2 (t) + fi2 (x)] + ~v
                  ˙i                                     +         hi (x(t))
    2 t0     i=1                         i=1
                                                 @xi        ~w i=1
      X Z x(t)
       n
                dxi (t)fi (x(t))
        i=1       x(t0 )
        Z t                XZ
                           n         x(t)                                                    L=T                  V
    1
  ⌘              dtL                        dxi (t)fi (x(t))                                             Z    t        n
                                                                                                                       X
    2                                                                                                1
            t0             i=1     x(t0 )
                                                                                                 T =              dt         x2 (t)
                                                                                                                             ˙i
                                                                                                     2       t0        i=1
                                                                             "                                                                        #
                                                          1
                                                               Z    t            X
                                                                                 n
                                                                                                      @fi (x(t))   ~v
                                                                                                                                   m
                                                                                                                                   X
                                                      V =               dt             fi2 (x)   + ~v            +                        h2 (x(t))
                                                                                                                                           i
                                                          2        t0            i=1
                                                                                                        @xi        ~w               i=1

                                                                                                                                                   15
16
17
dx = (b(x, t) + Bu)dt + dv

                       ⌧                Z    tf         ✓                                 ◆
                                                            1
C(xint , tint , u) =       (x(tf )) +              dt         u(t)T Ru(t) + V (x(t), t)
                                            tint            2                                 xint




J(x, t) = min C(x, t, u(t ! tf ))
              u(t!tf )



                                                                                                     18
                T                                                           
 @J        1 @J                             @J          @J  1    @2J
    =                              BR 1 B T    + V + bT    + Tr ⌫ 2
 @t        2 @x                             @x          @x  2    @x



                                   Z                               ✓            ◆
                                                                            S
J(x, t) =                log                      [Dx(t)] exp
                                       x(t)=x
                 Z       tf        ✓                                                                            ◆
                                       1
S = (x(tf )) +                d⌧         (x(⌧ )
                                          ˙        b(x(⌧ ), ⌧ ))T R(x(⌧ )
                                                                    ˙           b(x(⌧ ), ⌧ )) + V (x(⌧ ), ⌧ )
                     t                 2


V (x(t), t)


                                                                                                                    19
20

Mais conteúdo relacionado

Mais procurados

11.solution of linear and nonlinear partial differential equations using mixt...
11.solution of linear and nonlinear partial differential equations using mixt...11.solution of linear and nonlinear partial differential equations using mixt...
11.solution of linear and nonlinear partial differential equations using mixt...
Alexander Decker
 
Dynamic Tumor Growth Modelling
Dynamic Tumor Growth ModellingDynamic Tumor Growth Modelling
Dynamic Tumor Growth Modelling
Alex (Oleksiy) Varfolomiyev
 
Gaussseidelsor
GaussseidelsorGaussseidelsor
Gaussseidelsor
uis
 
Quantum modes - Ion Cotaescu
Quantum modes - Ion CotaescuQuantum modes - Ion Cotaescu
Quantum modes - Ion Cotaescu
SEENET-MTP
 
Numerical solution of spatiotemporal models from ecology
Numerical solution of spatiotemporal models from ecologyNumerical solution of spatiotemporal models from ecology
Numerical solution of spatiotemporal models from ecology
Kyrre Wahl Kongsgård
 
Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)
IJERD Editor
 

Mais procurados (20)

On the solvability of a system of forward-backward linear equations with unbo...
On the solvability of a system of forward-backward linear equations with unbo...On the solvability of a system of forward-backward linear equations with unbo...
On the solvability of a system of forward-backward linear equations with unbo...
 
optimal control principle slided
optimal control principle slidedoptimal control principle slided
optimal control principle slided
 
11.solution of linear and nonlinear partial differential equations using mixt...
11.solution of linear and nonlinear partial differential equations using mixt...11.solution of linear and nonlinear partial differential equations using mixt...
11.solution of linear and nonlinear partial differential equations using mixt...
 
Solution of linear and nonlinear partial differential equations using mixture...
Solution of linear and nonlinear partial differential equations using mixture...Solution of linear and nonlinear partial differential equations using mixture...
Solution of linear and nonlinear partial differential equations using mixture...
 
NONLINEAR DIFFERENCE EQUATIONS WITH SMALL PARAMETERS OF MULTIPLE SCALES
NONLINEAR DIFFERENCE EQUATIONS WITH SMALL PARAMETERS OF MULTIPLE SCALESNONLINEAR DIFFERENCE EQUATIONS WITH SMALL PARAMETERS OF MULTIPLE SCALES
NONLINEAR DIFFERENCE EQUATIONS WITH SMALL PARAMETERS OF MULTIPLE SCALES
 
2003 Ames.Models
2003 Ames.Models2003 Ames.Models
2003 Ames.Models
 
Scientific Computing with Python Webinar 9/18/2009:Curve Fitting
Scientific Computing with Python Webinar 9/18/2009:Curve FittingScientific Computing with Python Webinar 9/18/2009:Curve Fitting
Scientific Computing with Python Webinar 9/18/2009:Curve Fitting
 
On estimating the integrated co volatility using
On estimating the integrated co volatility usingOn estimating the integrated co volatility using
On estimating the integrated co volatility using
 
Dynamic Tumor Growth Modelling
Dynamic Tumor Growth ModellingDynamic Tumor Growth Modelling
Dynamic Tumor Growth Modelling
 
Numerical solution of boundary value problems by piecewise analysis method
Numerical solution of boundary value problems by piecewise analysis methodNumerical solution of boundary value problems by piecewise analysis method
Numerical solution of boundary value problems by piecewise analysis method
 
Gaussseidelsor
GaussseidelsorGaussseidelsor
Gaussseidelsor
 
Midsem sol 2013
Midsem sol 2013Midsem sol 2013
Midsem sol 2013
 
Levitan Centenary Conference Talk, June 27 2014
Levitan Centenary Conference Talk, June 27 2014Levitan Centenary Conference Talk, June 27 2014
Levitan Centenary Conference Talk, June 27 2014
 
Fdtd
FdtdFdtd
Fdtd
 
Chapter 5 (maths 3)
Chapter 5 (maths 3)Chapter 5 (maths 3)
Chapter 5 (maths 3)
 
Quantum modes - Ion Cotaescu
Quantum modes - Ion CotaescuQuantum modes - Ion Cotaescu
Quantum modes - Ion Cotaescu
 
Numerical solution of spatiotemporal models from ecology
Numerical solution of spatiotemporal models from ecologyNumerical solution of spatiotemporal models from ecology
Numerical solution of spatiotemporal models from ecology
 
Lesson 27: Integration by Substitution (slides)
Lesson 27: Integration by Substitution (slides)Lesson 27: Integration by Substitution (slides)
Lesson 27: Integration by Substitution (slides)
 
MUMS: Bayesian, Fiducial, and Frequentist Conference - Statistical Sparsity, ...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Statistical Sparsity, ...MUMS: Bayesian, Fiducial, and Frequentist Conference - Statistical Sparsity, ...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Statistical Sparsity, ...
 
Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)Welcome to International Journal of Engineering Research and Development (IJERD)
Welcome to International Journal of Engineering Research and Development (IJERD)
 

Semelhante a Nonlinear Filtering and Path Integral Method (Paper Review)

Logics of the laplace transform
Logics of the laplace transformLogics of the laplace transform
Logics of the laplace transform
Tarun Gehlot
 
Introduction to inverse problems
Introduction to inverse problemsIntroduction to inverse problems
Introduction to inverse problems
Delta Pi Systems
 
Calculus cheat sheet_integrals
Calculus cheat sheet_integralsCalculus cheat sheet_integrals
Calculus cheat sheet_integrals
UrbanX4
 
EM algorithm and its application in probabilistic latent semantic analysis
EM algorithm and its application in probabilistic latent semantic analysisEM algorithm and its application in probabilistic latent semantic analysis
EM algorithm and its application in probabilistic latent semantic analysis
zukun
 
Formulario cuantica 2
Formulario cuantica 2Formulario cuantica 2
Formulario cuantica 2
Abraham Prado
 
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD Editor
 
Reflect tsukuba524
Reflect tsukuba524Reflect tsukuba524
Reflect tsukuba524
kazuhase2011
 
Emat 213 study guide
Emat 213 study guideEmat 213 study guide
Emat 213 study guide
akabaka12
 
Unconditionally stable fdtd methods
Unconditionally stable fdtd methodsUnconditionally stable fdtd methods
Unconditionally stable fdtd methods
physics Imposible
 

Semelhante a Nonlinear Filtering and Path Integral Method (Paper Review) (20)

Logics of the laplace transform
Logics of the laplace transformLogics of the laplace transform
Logics of the laplace transform
 
Introduction to inverse problems
Introduction to inverse problemsIntroduction to inverse problems
Introduction to inverse problems
 
Calculus cheat sheet_integrals
Calculus cheat sheet_integralsCalculus cheat sheet_integrals
Calculus cheat sheet_integrals
 
Hw2 s
Hw2 sHw2 s
Hw2 s
 
EM algorithm and its application in probabilistic latent semantic analysis
EM algorithm and its application in probabilistic latent semantic analysisEM algorithm and its application in probabilistic latent semantic analysis
EM algorithm and its application in probabilistic latent semantic analysis
 
Formulario cuantica 2
Formulario cuantica 2Formulario cuantica 2
Formulario cuantica 2
 
Funcion gamma
Funcion gammaFuncion gamma
Funcion gamma
 
New Mathematical Tools for the Financial Sector
New Mathematical Tools for the Financial SectorNew Mathematical Tools for the Financial Sector
New Mathematical Tools for the Financial Sector
 
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
IJERD (www.ijerd.com) International Journal of Engineering Research and Devel...
 
Limits and Continuity of Functions
Limits and Continuity of Functions Limits and Continuity of Functions
Limits and Continuity of Functions
 
Trig identities
Trig identitiesTrig identities
Trig identities
 
Reflect tsukuba524
Reflect tsukuba524Reflect tsukuba524
Reflect tsukuba524
 
Emat 213 study guide
Emat 213 study guideEmat 213 study guide
Emat 213 study guide
 
Fdtd
FdtdFdtd
Fdtd
 
Unconditionally stable fdtd methods
Unconditionally stable fdtd methodsUnconditionally stable fdtd methods
Unconditionally stable fdtd methods
 
Unit Circle - Trigonometry
Unit Circle - TrigonometryUnit Circle - Trigonometry
Unit Circle - Trigonometry
 
Acturial maths
Acturial mathsActurial maths
Acturial maths
 
Course notes2summer2012
Course notes2summer2012Course notes2summer2012
Course notes2summer2012
 
02 2d systems matrix
02 2d systems matrix02 2d systems matrix
02 2d systems matrix
 
CME Deliverable Interest Rate Swap Future
CME Deliverable Interest Rate Swap FutureCME Deliverable Interest Rate Swap Future
CME Deliverable Interest Rate Swap Future
 

Mais de Kohta Ishikawa (11)

[Paper reading] Hamiltonian Neural Networks
 [Paper reading] Hamiltonian Neural Networks [Paper reading] Hamiltonian Neural Networks
[Paper reading] Hamiltonian Neural Networks
 
Spherical CNNs paper reading
Spherical CNNs paper readingSpherical CNNs paper reading
Spherical CNNs paper reading
 
Model Transport: Towards Scalable Transfer Learning on Manifolds 論文紹介
Model Transport: Towards Scalable Transfer Learning on Manifolds 論文紹介Model Transport: Towards Scalable Transfer Learning on Manifolds 論文紹介
Model Transport: Towards Scalable Transfer Learning on Manifolds 論文紹介
 
A brief explanation of Causal Entropic Forces
A brief explanation of Causal Entropic ForcesA brief explanation of Causal Entropic Forces
A brief explanation of Causal Entropic Forces
 
量子アニーリング解説 1
量子アニーリング解説 1量子アニーリング解説 1
量子アニーリング解説 1
 
R Language Definition 2.2 to 2.3
R Language Definition 2.2 to 2.3R Language Definition 2.2 to 2.3
R Language Definition 2.2 to 2.3
 
Introduction to pairtrading
Introduction to pairtradingIntroduction to pairtrading
Introduction to pairtrading
 
Particle Filter Tracking in Python
Particle Filter Tracking in PythonParticle Filter Tracking in Python
Particle Filter Tracking in Python
 
Rでisomap(多様体学習のはなし)
Rでisomap(多様体学習のはなし)Rでisomap(多様体学習のはなし)
Rでisomap(多様体学習のはなし)
 
Introduction to statistics
Introduction to statisticsIntroduction to statistics
Introduction to statistics
 
PRML_2.3.1~2.3.3
PRML_2.3.1~2.3.3PRML_2.3.1~2.3.3
PRML_2.3.1~2.3.3
 

Último

CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
giselly40
 
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slideHistor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slide
vu2urc
 

Último (20)

08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slides
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 
08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men08448380779 Call Girls In Civil Lines Women Seeking Men
08448380779 Call Girls In Civil Lines Women Seeking Men
 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
Advantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your BusinessAdvantages of Hiring UIUX Design Service Providers for Your Business
Advantages of Hiring UIUX Design Service Providers for Your Business
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
 
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
Apidays Singapore 2024 - Building Digital Trust in a Digital Economy by Veron...
 
Histor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slideHistor y of HAM Radio presentation slide
Histor y of HAM Radio presentation slide
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?
 

Nonlinear Filtering and Path Integral Method (Paper Review)

  • 1. Brief Paper Review B.Balaji (2009) “Nonlinear filtering and quantum physics A Feynman path integral perspective” Kohta Ishikawa Oct 30. 2011 3 1
  • 2. 2
  • 3. xk = f (xk 1 , tk 1 )xk 1 + e(xk 1 , tk 1 ) System Model yk = h(xk , tk ) + wk Observation Model dx(t) = f (x(t), t)dt + e(x(t), t)dw(t) System Model(SDE) y(tk ) = h(x(tk ), tk ) + wk Observation Model dx(t) = f (x(t), t)dt + e(x(t), t)dv(t) System Model(SDE) dy(t) = h(x(t), t)dt + dw(t) Observation Model(SDE) 3
  • 4. m X d (t, x) = LY (t, x)dt + hi (x) (t, x)dyi (t) i=1 n X @ n X @ 2 (t, x) m X 1 1 LY (t, x) = (fi (x) (t, x)) + h2 (x) (t, x) i i=1 @xi 2 i=1 @x2 i 2 i=1 (0, x) = 0 (x) dy(t) 4
  • 5. dy(t) m ! X u(t, x) = exp hi (x)yi (t) (t, x) i=1 Then, DMZ equation translates to 0 1 n X @ 2 u(t, x) n X m X @u(t, x) 1 @ fi (x) + @hj (x) A @u(t, x) = + yj (t) @t 2 i=1 @x2 i i=1 j=1 @xi @xi ✓X n m X m X @fi (x) 1 1 + h2 (x) i yi (t) hi (x) i=1 @xi 2 i=1 2i=1 m n XX m X X n ◆ @hi (x) 1 @hi (x) @hj (x) + yi (t)fj (x) yi (t)yj (t) u(t, x) i=1 j=1 @xj 2 i,j=1 @xk @xk k=1 u(0, x) = 0 (x) 5
  • 6. y(⌧l ) post-measurement form y(t) ! y(⌧l 1 ) pre-measurement form 6
  • 7. ⌧l 1  t  ⌧l {⌧0 , ⌧1 , · · · } 0 1 n n m @ul (t, x) 1 X @ 2 ul (t, x) X @ X @hj (x) A @ul (t, x) = 2 + fi (x) + yj (⌧l ) @t 2 i=1 @xi i=1 j=1 @xi @xi ✓X n m X m X @fi (x) 1 1 + h2 (x) i yi (⌧l ) hi (x) i=1 @xi 2 i=1 2i=1 m n XX m X X n ◆ @hi (x) 1 @hi (x) @hj (x) + yi (⌧l )fj (x) yi (⌧l )yj (⌧l ) ul (t, x) i=1 j=1 @xj 2 i,j=1 @xk @xk k=1 ul (⌧l , x) = ul 1 (⌧l 1 , x) 7
  • 8. m ! X (t, x) ul (t, x) = exp ˜ yi (⌧l )hi (x) ul (t, x) i=1 n X @ 2 ul (t, x) n X @ ul (t, x) ˜ 1 ˜ @ ul (t, x) ˜ = fi (x) @t 2 i=1 @x2 i i=1 @xi n m ! X @fi (x) 1 X + h2 (x) ul (t, x) i ˜ i=1 @xi 2 i=1 m ! X ul (⌧l ˜ 1 , x) = exp yi (⌧l )hi (x) ul 1 (⌧l 1 , x) i=1 yi (⌧l 1) 8
  • 9. 9
  • 10. Stochastic Process probability density of state path measure, functional differentiation Fokker-Planck equation Path Integral representation green’s function 10
  • 11. SDE for continuous filtering xi (t) = fi (x(t)) + vi (t) ˙ yi (t) = hi (x(t)) + wi (t) ˙ Probability density Z P (t, x, y|t0 , x0 .y0 ) = [d⇢(v(t))][d⇢(w(t))] (x(t) ˙ f (x(t)) v(t)) n ⇥ [dx(t)] (x(t) ˙ f (x(t)) v(t)) x(t) (y(t) ˙ h(x(t)) w(t)) m ⇥ [dy(t)] (y(t) ˙ h(y(t)) w(t)) y(t) n m ⇥ (x(t) x)|x(t0 )=x0 (y(t) y)|y(t0 ) = y0 11
  • 12. ! 1 XZ t n [d⇢(v(t))] = [Dv(t)] exp 2~v vi (t)2 dt ~v , ~w i=1 t0 ! 1 XZ t m [d⇢(w(t))] = [Dw(t)] exp vi (t)2 dt 2~w i=1 t0 Z y(t)=y Z x(t)=x P (t, x, y|t0 , x0 .y0 ) = [Dv(t)][Dw(t)][Dx(t)][Dy(t)] y(t0 )=y0 x(t0 )=x0 ! 1 XZ t n XZ t n @fi (x(t)) 1 XZ t m 2 2 ⇥ exp vi (t)dt dt wi dt 2~v i=1 t0 i=1 t0 @xi 2~w i=1 t0 n ⇥ (x(t) f (x(t)) v(t)) m (y(t) h(y(t)) w(t)) ˙ ˙ Z y(t)=y Z x(t)=x = [Dx(t)][Dy(t)] exp ( S) y(t0 )=y0 x(t0 )=x0 Z " n n m # 1 t 1 X X @fi (x(t)) 1 X 2 S= dt (xi (t) ˙ fi (x(t))) + + ( yi ˙ hi (x(t)))2 2 t0 ~v i=1 i=1 @xi ~w i=1 12
  • 13. measurement term in the action S Z ti m X Z ti m X⇥ 1 1 ⇤ dt ( yi ˙ hi (x(t))) = dt ˙2 yi (t) + h2 (x(t)) i 2hi (x(t))yi (t) ˙ 2~w ti 1 i=1 2~w ti 1 i=1 relevant term Z ti m X ⇢ 1 Pm 1 ~w j=1 hj ([x(ti ) + x(ti Pm 1 )]/2)[yj (tj ) yj (tj 1 )] dt hj (x(t))yj (t) ⇠ ˙ 1 ~w ti 1 j=1 ~w j=1 hj ([x(ti ) + x(ti 1 )]/2)[yj (tj 1 ) yj (tj 2 )] 13
  • 14. leads to approximated probability density P (ti , xi , yi |ti 1 , xi 1 , yi 1 ) ˜ ⇠ P (ti , xi |ti 1 , xi 1 ) ⇣ Pm ⌘ ⇢ 1 exp j=1 hj ([x(ti ) + x(ti 1 )]/2)[yj (tj ) yj (tj 1 )] ⇥ ⇣ ~w Pm ⌘ 1 exp ~w j=1 hj ([x(ti ) + x(ti 1 )]/2)[yj (tj 1 ) yj (tj 2 )] Z x(ti )=xi ˜ P (ti , xi |ti 1 , xi 1 ) = [Dx(t)] exp( S(ti 1 , ti )) x(ti 1 )=xi 1 Z " n n m # 1 ti 1 X X @fi (x(t)) 1 X 2 S(ti 1 , ti ) = dt (xi (t) ˙ fi (x(t))) + + h2 (x(t)) i 2 ti 1 ~v i=1 i=1 @xi ~w i=1 ˜ P (t, x|ti 1 , xi 1 ) 14
  • 15. Z x(t)=x ✓ ◆ ˜ 1 P (t, x|t0 , x0 ) = [Dx(t)] exp S x(t0 )=x0 ~v Z " n n m # 1 t X @fi (x(t)) ~v X X S= dt + [x2 (t) ˙i ˙ fi2 (x) 2xi (t)fi (x(t))] + ~v + h2 (x(t)) i 2 t0 i=1 i=1 @xi ~w i=1 Z t "X n n X @fi (x(t)) m # 1 ~v X 2 = dt [x2 (t) + fi2 (x)] + ~v ˙i + hi (x(t)) 2 t0 i=1 i=1 @xi ~w i=1 X Z x(t) n dxi (t)fi (x(t)) i=1 x(t0 ) Z t XZ n x(t) L=T V 1 ⌘ dtL dxi (t)fi (x(t)) Z t n X 2 1 t0 i=1 x(t0 ) T = dt x2 (t) ˙i 2 t0 i=1 " # 1 Z t X n @fi (x(t)) ~v m X V = dt fi2 (x) + ~v + h2 (x(t)) i 2 t0 i=1 @xi ~w i=1 15
  • 16. 16
  • 17. 17
  • 18. dx = (b(x, t) + Bu)dt + dv ⌧ Z tf ✓ ◆ 1 C(xint , tint , u) = (x(tf )) + dt u(t)T Ru(t) + V (x(t), t) tint 2 xint J(x, t) = min C(x, t, u(t ! tf )) u(t!tf ) 18
  • 19. T  @J 1 @J @J @J 1 @2J = BR 1 B T + V + bT + Tr ⌫ 2 @t 2 @x @x @x 2 @x Z ✓ ◆ S J(x, t) = log [Dx(t)] exp x(t)=x Z tf ✓ ◆ 1 S = (x(tf )) + d⌧ (x(⌧ ) ˙ b(x(⌧ ), ⌧ ))T R(x(⌧ ) ˙ b(x(⌧ ), ⌧ )) + V (x(⌧ ), ⌧ ) t 2 V (x(t), t) 19
  • 20. 20