Stan Math Library  2.10.0
reverse mode automatic differentiation
grad_tr_mat_times_hessian.hpp
Go to the documentation of this file.
1 #ifndef STAN_MATH_MIX_MAT_FUNCTOR_GRAD_TR_MAT_TIMES_HESSIAN_HPP
2 #define STAN_MATH_MIX_MAT_FUNCTOR_GRAD_TR_MAT_TIMES_HESSIAN_HPP
3 
4 #include <stan/math/fwd/core.hpp>
6 #include <stan/math/rev/core.hpp>
8 #include <stdexcept>
9 #include <vector>
10 
11 namespace stan {
12 
13  namespace math {
14 
15 
16  // FIXME: add other results that are easy to extract
17  // // N * (fwd(2) + bk)
18  template <typename F>
19  void
21  const F& f,
22  const Eigen::Matrix<double, Eigen::Dynamic, 1>& x,
23  const Eigen::Matrix<double, Eigen::Dynamic, Eigen::Dynamic>& M,
24  Eigen::Matrix<double, Eigen::Dynamic, 1>& grad_tr_MH
25  ) {
26  using Eigen::Matrix;
27  using Eigen::Dynamic;
28  start_nested();
29  try {
30  grad_tr_MH.resize(x.size());
31 
32  Matrix<var, Dynamic, 1> x_var(x.size());
33  for (int i = 0; i < x.size(); ++i)
34  x_var(i) = x(i);
35 
36  Matrix<fvar<var>, Dynamic, 1> x_fvar(x.size());
37 
38  var sum(0.0);
39  Matrix<double, Dynamic, 1> M_n(x.size());
40  for (int n = 0; n < x.size(); ++n) {
41  for (int k = 0; k < x.size(); ++k)
42  M_n(k) = M(n, k);
43  for (int k = 0; k < x.size(); ++k)
44  x_fvar(k) = fvar<var>(x_var(k), k == n);
45  fvar<var> fx;
46  fvar<var> grad_fx_dot_v;
47  gradient_dot_vector<fvar<var>, double>(f, x_fvar, M_n, fx,
48  grad_fx_dot_v);
49  sum += grad_fx_dot_v.d_;
50  }
51 
52  stan::math::grad(sum.vi_);
53  for (int i = 0; i < x.size(); ++i)
54  grad_tr_MH(i) = x_var(i).adj();
55  } catch (const std::exception& e) {
57  throw;
58  }
60  }
61 
62 
63  } // namespace math
64 } // namespace stan
65 #endif
fvar< T > sum(const std::vector< fvar< T > > &m)
Return the sum of the entries of the specified standard vector.
Definition: sum.hpp:20
void grad_tr_mat_times_hessian(const F &f, const Eigen::Matrix< double, Eigen::Dynamic, 1 > &x, const Eigen::Matrix< double, Eigen::Dynamic, Eigen::Dynamic > &M, Eigen::Matrix< double, Eigen::Dynamic, 1 > &grad_tr_MH)
Independent (input) and dependent (output) variables for gradients.
Definition: var.hpp:31
static void grad(vari *vi)
Compute the gradient for all variables starting from the specified root variable implementation.
Definition: grad.hpp:30
double e()
Return the base of the natural logarithm.
Definition: constants.hpp:95
static void recover_memory_nested()
Recover only the memory used for the top nested call.
static void start_nested()
Record the current position so that recover_memory_nested() can find it.

     [ Stan Home Page ] © 2011–2016, Stan Development Team.