diff --git a/GPy/core/gp.py b/GPy/core/gp.py
index 3cfa38ec..01190101 100644
--- a/GPy/core/gp.py
+++ b/GPy/core/gp.py
@@ -283,7 +283,7 @@ class GP(Model):
 
     def log_likelihood(self):
         """
-        The log marginal likelihood of the model, :math:`p(\mathbf{y})`, this is the objective function of the model being optimised
+        The log marginal likelihood of the model, :math:`p(\\mathbf{y})`, this is the objective function of the model being optimised
         """
         return self._log_marginal_likelihood
 
@@ -296,9 +296,9 @@ class GP(Model):
         diagonal of the covariance is returned.
 
         .. math::
-            p(f*|X*, X, Y) = \int^{\inf}_{\inf} p(f*|f,X*)p(f|X,Y) df
-                        = N(f*| K_{x*x}(K_{xx} + \Sigma)^{-1}Y, K_{x*x*} - K_{xx*}(K_{xx} + \Sigma)^{-1}K_{xx*}
-            \Sigma := \texttt{Likelihood.variance / Approximate likelihood covariance}
+            p(f*|X*, X, Y) = \\int^{\\inf}_{\\inf} p(f*|f,X*)p(f|X,Y) df
+                        = N(f*| K_{x*x}(K_{xx} + \\Sigma)^{-1}Y, K_{x*x*} - K_{xx*}(K_{xx} + \\Sigma)^{-1}K_{xx*}
+            \\Sigma := \texttt{Likelihood.variance / Approximate likelihood covariance}
         """
         mu, var = self.posterior._raw_predict(kern=self.kern if kern is None else kern, Xnew=Xnew, pred_var=self._predictive_variable, full_cov=full_cov)
         if self.mean_function is not None:
@@ -702,7 +702,7 @@ class GP(Model):
         Calculation of the log predictive density
 
         .. math:
-            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\mu_{*}\\sigma^{2}_{*})
+            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\\mu_{*}\\sigma^{2}_{*})
 
         :param x_test: test locations (x_{*})
         :type x_test: (Nx1) array
@@ -718,7 +718,7 @@ class GP(Model):
         Calculation of the log predictive density by sampling
 
         .. math:
-            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\mu_{*}\\sigma^{2}_{*})
+            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\\mu_{*}\\sigma^{2}_{*})
 
         :param x_test: test locations (x_{*})
         :type x_test: (Nx1) array
@@ -734,24 +734,24 @@ class GP(Model):
 
     def _raw_posterior_covariance_between_points(self, X1, X2):
         """
-        Computes the posterior covariance between points. Does not account for 
+        Computes the posterior covariance between points. Does not account for
         normalization or likelihood
 
         :param X1: some input observations
         :param X2: other input observations
 
-        :returns: 
+        :returns:
             cov: raw posterior covariance: k(X1,X2) - k(X1,X) G^{-1} K(X,X2)
         """
         return self.posterior.covariance_between_points(self.kern, self.X, X1, X2)
 
 
-    def posterior_covariance_between_points(self, X1, X2, Y_metadata=None, 
-                                            likelihood=None, 
+    def posterior_covariance_between_points(self, X1, X2, Y_metadata=None,
+                                            likelihood=None,
                                             include_likelihood=True):
         """
-        Computes the posterior covariance between points. Includes likelihood 
-        variance as well as normalization so that evaluation at (x,x) is consistent 
+        Computes the posterior covariance between points. Includes likelihood
+        variance as well as normalization so that evaluation at (x,x) is consistent
         with model.predict
 
         :param X1: some input observations
@@ -762,8 +762,8 @@ class GP(Model):
                                    the predicted underlying latent function f.
         :type include_likelihood: bool
 
-        :returns: 
-            cov: posterior covariance, a Numpy array, Nnew x Nnew if 
+        :returns:
+            cov: posterior covariance, a Numpy array, Nnew x Nnew if
             self.output_dim == 1, and Nnew x Nnew x self.output_dim otherwise.
         """
 
@@ -774,7 +774,7 @@ class GP(Model):
             mean, _ = self._raw_predict(X1, full_cov=True)
             if likelihood is None:
                 likelihood = self.likelihood
-            _, cov = likelihood.predictive_values(mean, cov, full_cov=True, 
+            _, cov = likelihood.predictive_values(mean, cov, full_cov=True,
                                                   Y_metadata=Y_metadata)
 
         if self.normalizer is not None:
diff --git a/GPy/core/symbolic.py b/GPy/core/symbolic.py
index c4261e24..c11312e6 100644
--- a/GPy/core/symbolic.py
+++ b/GPy/core/symbolic.py
@@ -44,7 +44,7 @@ class Symbolic_core():
         self._set_derivatives(derivatives)
         self._set_parameters(parameters)
         # Convert the expressions to a list for common sub expression elimination
-        # We should find the following type of expressions: 'function', 'derivative', 'second_derivative', 'third_derivative'. 
+        # We should find the following type of expressions: 'function', 'derivative', 'second_derivative', 'third_derivative'.
         self.update_expression_list()
 
         # Apply any global stabilisation operations to expressions.
@@ -86,7 +86,7 @@ class Symbolic_core():
         # object except as cached. For covariance functions this is X
         # and Z, for likelihoods F and for mapping functions X.
         self.cacheable_vars = [] # list of everything that's cacheable
-        for var in cacheable:            
+        for var in cacheable:
             self.variables[var] = [e for e in vars if e.name.split('_')[0]==var.lower()]
             self.cacheable_vars += self.variables[var]
         for var in cacheable:
@@ -105,7 +105,7 @@ class Symbolic_core():
             for derivative in derivatives:
                 derivative_arguments += self.variables[derivative]
 
-            # Do symbolic work to compute derivatives.        
+            # Do symbolic work to compute derivatives.
             for key, func in self.expressions.items():
                 # if func['function'].is_Matrix:
                 #     rows = func['function'].shape[0]
@@ -126,7 +126,7 @@ class Symbolic_core():
                 if theta.name in parameters:
                     val = parameters[theta.name]
             # Add parameter.
-            
+
             self.link_parameters(Param(theta.name, val, None))
             #self._set_attribute(theta.name, )
 
@@ -174,7 +174,7 @@ class Symbolic_core():
             code = self.code[function]['derivative'][theta.name]
             gradient[theta.name] = (partial*eval(code, self.namespace)).sum()
         return gradient
-        
+
     def eval_gradients_X(self, function, partial, **kwargs):
         if 'X' in kwargs:
             gradients_X = np.zeros_like(kwargs['X'])
@@ -194,7 +194,7 @@ class Symbolic_core():
         for variable, code in self.variable_sort(self.code['parameters_changed']):
             lcode += self._print_code(variable) + ' = ' + self._print_code(code) + '\n'
         return lcode
-    
+
     def code_update_cache(self):
         lcode = ''
         for var in self.cacheable:
@@ -208,7 +208,7 @@ class Symbolic_core():
             for i, theta in enumerate(self.variables[var]):
                 lcode+= "\t" + var + '= np.atleast_2d(' + var + ')\n'
                 lcode+= "\t" + self._print_code(theta.name) + ' = ' + var + '[:, ' + str(i) + "]" + reorder + "\n"
-    
+
         for variable, code in self.variable_sort(self.code['update_cache']):
             lcode+= self._print_code(variable) + ' = ' + self._print_code(code) + "\n"
 
@@ -250,7 +250,7 @@ class Symbolic_core():
         """Make sure namespace gets updated when setting attributes."""
         setattr(self, name, value)
         self.namespace.update({name: getattr(self, name)})
-        
+
 
     def update_expression_list(self):
         """Extract a list of expressions from the dictionary of expressions."""
@@ -260,9 +260,9 @@ class Symbolic_core():
         for fname, fexpressions in self.expressions.items():
             for type, texpressions in fexpressions.items():
                 if type == 'function':
-                    self.expression_list.append(texpressions)            
+                    self.expression_list.append(texpressions)
                     self.expression_keys.append([fname, type])
-                    self.expression_order.append(1) 
+                    self.expression_order.append(1)
                 elif type[-10:] == 'derivative':
                     for dtype, expression in texpressions.items():
                         self.expression_list.append(expression)
@@ -274,9 +274,9 @@ class Symbolic_core():
                         elif type[:-10] == 'third_':
                             self.expression_order.append(5) #sym.count_ops(self.expressions[type][dtype]))
                 else:
-                    self.expression_list.append(fexpressions[type])            
+                    self.expression_list.append(fexpressions[type])
                     self.expression_keys.append([fname, type])
-                    self.expression_order.append(2) 
+                    self.expression_order.append(2)
 
         # This step may be unecessary.
         # Not 100% sure if the sub expression elimination is order sensitive. This step orders the list with the 'function' code first and derivatives after.
@@ -313,7 +313,7 @@ class Symbolic_core():
             sym_var = sym.var(cache_prefix + str(i))
             self.variables[cache_prefix].append(sym_var)
             replace_dict[expr.name] = sym_var
-            
+
         for i, expr in enumerate(params_change_list):
             sym_var = sym.var(sub_prefix + str(i))
             self.variables[sub_prefix].append(sym_var)
@@ -329,7 +329,7 @@ class Symbolic_core():
         for keys in self.expression_keys:
             for replace, void in common_sub_expressions:
                 setInDict(self.expressions, keys, getFromDict(self.expressions, keys).subs(replace, replace_dict[replace.name]))
-        
+
         self.expressions['parameters_changed'] = {}
         self.expressions['update_cache'] = {}
         for var, expr in common_sub_expressions:
@@ -339,7 +339,7 @@ class Symbolic_core():
                 self.expressions['update_cache'][replace_dict[var.name].name] = expr
             else:
                 self.expressions['parameters_changed'][replace_dict[var.name].name] = expr
-            
+
 
     def _gen_code(self):
         """Generate code for the list of expressions provided using the common sub-expression eliminator to separate out portions that are computed multiple times."""
@@ -357,8 +357,8 @@ class Symbolic_core():
             return code
 
         self.code = match_key(self.expressions)
-                            
- 
+
+
     def _expr2code(self, arg_list, expr):
         """Convert the given symbolic expression into code."""
         code = lambdastr(arg_list, expr)
@@ -379,7 +379,7 @@ class Symbolic_core():
     def _display_expression(self, keys, user_substitutes={}):
         """Helper function for human friendly display of the symbolic components."""
         # Create some pretty maths symbols for the display.
-        sigma, alpha, nu, omega, l, variance = sym.var('\sigma, \alpha, \nu, \omega, \ell, \sigma^2')
+        sigma, alpha, nu, omega, l, variance = sym.var(r'\sigma, \alpha, \nu, \omega, \ell, \sigma^2')
         substitutes = {'scale': sigma, 'shape': alpha, 'lengthscale': l, 'variance': variance}
         substitutes.update(user_substitutes)
 
@@ -416,5 +416,5 @@ class Symbolic_core():
                 return int(digits[0])
             else:
                 return x[0]
-            
+
         return sorted(var_dict.items(), key=sort_key, reverse=reverse)
diff --git a/GPy/inference/latent_function_inference/laplace.py b/GPy/inference/latent_function_inference/laplace.py
index a77c7631..e943077b 100644
--- a/GPy/inference/latent_function_inference/laplace.py
+++ b/GPy/inference/latent_function_inference/laplace.py
@@ -27,7 +27,7 @@ class Laplace(LatentFunctionInference):
         """
         Laplace Approximation
 
-        Find the moments \hat{f} and the hessian at this point
+        Find the moments \\hat{f} and the hessian at this point
         (using Newton-Raphson) of the unnormalised posterior
 
         """
diff --git a/GPy/inference/latent_function_inference/posterior.py b/GPy/inference/latent_function_inference/posterior.py
index 47f8c615..e3cea7ea 100644
--- a/GPy/inference/latent_function_inference/posterior.py
+++ b/GPy/inference/latent_function_inference/posterior.py
@@ -178,8 +178,8 @@ class Posterior(object):
         """
         The inverse of the woodbury matrix, in the gaussian likelihood case it is defined as
         $$
-        (K_{xx} + \Sigma_{xx})^{-1}
-        \Sigma_{xx} := \texttt{Likelihood.variance / Approximate likelihood covariance}
+        (K_{xx} + \\Sigma_{xx})^{-1}
+        \\Sigma_{xx} := \texttt{Likelihood.variance / Approximate likelihood covariance}
         $$
         """
         if self._woodbury_inv is None:
@@ -200,8 +200,8 @@ class Posterior(object):
         """
         Woodbury vector in the gaussian likelihood case only is defined as
         $$
-        (K_{xx} + \Sigma)^{-1}Y
-        \Sigma := \texttt{Likelihood.variance / Approximate likelihood covariance}
+        (K_{xx} + \\Sigma)^{-1}Y
+        \\Sigma := \texttt{Likelihood.variance / Approximate likelihood covariance}
         $$
         """
         if self._woodbury_vector is None:
diff --git a/GPy/kern/src/coregionalize.py b/GPy/kern/src/coregionalize.py
index 7f92d4f7..aaaeae72 100644
--- a/GPy/kern/src/coregionalize.py
+++ b/GPy/kern/src/coregionalize.py
@@ -19,7 +19,7 @@ except ImportError:
 
 
 class Coregionalize(Kern):
-    """
+    r"""
     Covariance function for intrinsic/linear coregionalization models
 
     This covariance has the form:
diff --git a/GPy/kern/src/eq_ode1.py b/GPy/kern/src/eq_ode1.py
index 4361ec23..69b5ec54 100644
--- a/GPy/kern/src/eq_ode1.py
+++ b/GPy/kern/src/eq_ode1.py
@@ -15,7 +15,7 @@ class EQ_ODE1(Kern):
 
     This outputs of this kernel have the form
     .. math::
-       \frac{\text{d}y_j}{\text{d}t} = \sum_{i=1}^R w_{j,i} u_i(t-\delta_j) - d_jy_j(t)
+       \frac{\text{d}y_j}{\text{d}t} = \\sum_{i=1}^R w_{j,i} u_i(t-\\delta_j) - d_jy_j(t)
 
     where :math:`R` is the rank of the system, :math:`w_{j,i}` is the sensitivity of the :math:`j`th output to the :math:`i`th latent function, :math:`d_j` is the decay rate of the :math:`j`th output and :math:`u_i(t)` are independent latent Gaussian processes goverened by an exponentiated quadratic covariance.
 
diff --git a/GPy/kern/src/eq_ode2.py b/GPy/kern/src/eq_ode2.py
index 27b15b87..e2f251d7 100644
--- a/GPy/kern/src/eq_ode2.py
+++ b/GPy/kern/src/eq_ode2.py
@@ -15,7 +15,7 @@ class EQ_ODE2(Kern):
 
     This outputs of this kernel have the form
     .. math::
-       \frac{\text{d}^2y_j(t)}{\text{d}^2t} + C_j\frac{\text{d}y_j(t)}{\text{d}t} + B_jy_j(t) = \sum_{i=1}^R w_{j,i} u_i(t)
+       \frac{\text{d}^2y_j(t)}{\text{d}^2t} + C_j\frac{\text{d}y_j(t)}{\text{d}t} + B_jy_j(t) = \\sum_{i=1}^R w_{j,i} u_i(t)
 
     where :math:`R` is the rank of the system, :math:`w_{j,i}` is the sensitivity of the :math:`j`th output to the :math:`i`th latent function, :math:`d_j` is the decay rate of the :math:`j`th output and :math:`f_i(t)` and :math:`g_i(t)` are independent latent Gaussian processes goverened by an exponentiated quadratic covariance.
 
diff --git a/GPy/kern/src/grid_kerns.py b/GPy/kern/src/grid_kerns.py
index 9f153c27..5acec525 100644
--- a/GPy/kern/src/grid_kerns.py
+++ b/GPy/kern/src/grid_kerns.py
@@ -45,7 +45,7 @@ class GridRBF(GridKern):
 
     .. math::
 
-       k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r^2 \\bigg)
+       k(r) = \\sigma^2 \\exp \\bigg(- \\frac{1}{2} r^2 \\bigg)
 
     """
     _support_GPU = True
diff --git a/GPy/kern/src/kern.py b/GPy/kern/src/kern.py
index b8597c73..92b41e25 100644
--- a/GPy/kern/src/kern.py
+++ b/GPy/kern/src/kern.py
@@ -146,25 +146,25 @@ class Kern(Parameterized):
     def psi0(self, Z, variational_posterior):
         """
         .. math::
-            \psi_0 = \sum_{i=0}^{n}E_{q(X)}[k(X_i, X_i)]
+            \\psi_0 = \\sum_{i=0}^{n}E_{q(X)}[k(X_i, X_i)]
         """
         return self.psicomp.psicomputations(self, Z, variational_posterior)[0]
     def psi1(self, Z, variational_posterior):
         """
         .. math::
-            \psi_1^{n,m} = E_{q(X)}[k(X_n, Z_m)]
+            \\psi_1^{n,m} = E_{q(X)}[k(X_n, Z_m)]
         """
         return self.psicomp.psicomputations(self, Z, variational_posterior)[1]
     def psi2(self, Z, variational_posterior):
         """
         .. math::
-            \psi_2^{m,m'} = \sum_{i=0}^{n}E_{q(X)}[ k(Z_m, X_i) k(X_i, Z_{m'})]
+            \\psi_2^{m,m'} = \\sum_{i=0}^{n}E_{q(X)}[ k(Z_m, X_i) k(X_i, Z_{m'})]
         """
         return self.psicomp.psicomputations(self, Z, variational_posterior, return_psi2_n=False)[2]
     def psi2n(self, Z, variational_posterior):
         """
         .. math::
-            \psi_2^{n,m,m'} = E_{q(X)}[ k(Z_m, X_n) k(X_n, Z_{m'})]
+            \\psi_2^{n,m,m'} = E_{q(X)}[ k(Z_m, X_n) k(X_n, Z_{m'})]
 
         Thus, we do not sum out n, compared to psi2
         """
@@ -173,7 +173,7 @@ class Kern(Parameterized):
         """
         .. math::
 
-            \\frac{\partial L}{\partial X} = \\frac{\partial L}{\partial K}\\frac{\partial K}{\partial X}
+            \\frac{\\partial L}{\\partial X} = \\frac{\\partial L}{\\partial K}\\frac{\\partial K}{\\partial X}
         """
         raise NotImplementedError
     def gradients_X_X2(self, dL_dK, X, X2):
@@ -182,7 +182,7 @@ class Kern(Parameterized):
         """
         .. math::
 
-            \\frac{\partial^2 L}{\partial X\partial X_2} = \\frac{\partial L}{\partial K}\\frac{\partial^2 K}{\partial X\partial X_2}
+            \\frac{\\partial^2 L}{\\partial X\\partial X_2} = \\frac{\\partial L}{\\partial K}\\frac{\\partial^2 K}{\\partial X\\partial X_2}
         """
         raise NotImplementedError("This is the second derivative of K wrt X and X2, and not implemented for this kernel")
     def gradients_XX_diag(self, dL_dKdiag, X, cov=True):
@@ -203,7 +203,7 @@ class Kern(Parameterized):
     def update_gradients_full(self, dL_dK, X, X2):
         """Set the gradients of all parameters when doing full (N) inference."""
         raise NotImplementedError
-    
+
     def reset_gradients(self):
         raise NotImplementedError
 
@@ -216,9 +216,9 @@ class Kern(Parameterized):
 
         .. math::
 
-            \\frac{\partial L}{\partial \\theta_i} & = \\frac{\partial L}{\partial \psi_0}\\frac{\partial \psi_0}{\partial \\theta_i}\\
-                & \quad + \\frac{\partial L}{\partial \psi_1}\\frac{\partial \psi_1}{\partial \\theta_i}\\
-                & \quad + \\frac{\partial L}{\partial \psi_2}\\frac{\partial \psi_2}{\partial \\theta_i}
+            \\frac{\\partial L}{\\partial \\theta_i} & = \\frac{\\partial L}{\\partial \\psi_0}\\frac{\\partial \\psi_0}{\\partial \\theta_i}\\
+                & \\quad + \\frac{\\partial L}{\\partial \\psi_1}\\frac{\\partial \\psi_1}{\\partial \\theta_i}\\
+                & \\quad + \\frac{\\partial L}{\\partial \\psi_2}\\frac{\\partial \\psi_2}{\\partial \\theta_i}
 
         Thus, we push the different derivatives through the gradients of the psi
         statistics. Be sure to set the gradients for all kernel
diff --git a/GPy/kern/src/linear.py b/GPy/kern/src/linear.py
index bf75f201..6106a671 100644
--- a/GPy/kern/src/linear.py
+++ b/GPy/kern/src/linear.py
@@ -16,15 +16,15 @@ class Linear(Kern):
 
     .. math::
 
-       k(x,y) = \sum_{i=1}^{\\text{input_dim}} \sigma^2_i x_iy_i
+       k(x,y) = \\sum_{i=1}^{\\text{input_dim}} \\sigma^2_i x_iy_i
 
     :param input_dim: the number of input dimensions
     :type input_dim: int
-    :param variances: the vector of variances :math:`\sigma^2_i`
+    :param variances: the vector of variances :math:`\\sigma^2_i`
     :type variances: array or list of the appropriate size (or float if there
                      is only one variance parameter)
     :param ARD: Auto Relevance Determination. If False, the kernel has only one
-                variance parameter \sigma^2, otherwise there is one variance
+                variance parameter \\sigma^2, otherwise there is one variance
                 parameter per dimension.
     :type ARD: Boolean
     :rtype: kernel object
@@ -121,7 +121,7 @@ class Linear(Kern):
         the returned array is of shape [NxNxQxQ].
 
         ..math:
-            \frac{\partial^2 K}{\partial X2 ^2} = - \frac{\partial^2 K}{\partial X\partial X2}
+            \frac{\\partial^2 K}{\\partial X2 ^2} = - \frac{\\partial^2 K}{\\partial X\\partial X2}
 
         ..returns:
             dL2_dXdX2:  [NxMxQxQ] for X [NxQ] and X2[MxQ] (X2 is X if, X2 is None)
diff --git a/GPy/kern/src/mlp.py b/GPy/kern/src/mlp.py
index dc69f5fd..2b087df0 100644
--- a/GPy/kern/src/mlp.py
+++ b/GPy/kern/src/mlp.py
@@ -20,12 +20,12 @@ class MLP(Kern):
 
     :param input_dim: the number of input dimensions
     :type input_dim: int
-    :param variance: the variance :math:`\sigma^2`
+    :param variance: the variance :math:`\\sigma^2`
     :type variance: float
-    :param weight_variance: the vector of the variances of the prior over input weights in the neural network :math:`\sigma^2_w`
+    :param weight_variance: the vector of the variances of the prior over input weights in the neural network :math:`\\sigma^2_w`
     :type weight_variance: array or list of the appropriate size (or float if there is only one weight variance parameter)
-    :param bias_variance: the variance of the prior over bias parameters :math:`\sigma^2_b`
-    :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one weight variance parameter \sigma^2_w), otherwise there is one weight variance parameter per dimension.
+    :param bias_variance: the variance of the prior over bias parameters :math:`\\sigma^2_b`
+    :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one weight variance parameter \\sigma^2_w), otherwise there is one weight variance parameter per dimension.
     :type ARD: Boolean
     :rtype: Kernpart object
 
diff --git a/GPy/kern/src/rbf.py b/GPy/kern/src/rbf.py
index 9ed81b9c..28e7a977 100644
--- a/GPy/kern/src/rbf.py
+++ b/GPy/kern/src/rbf.py
@@ -16,7 +16,7 @@ class RBF(Stationary):
 
     .. math::
 
-       k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r^2 \\bigg)
+       k(r) = \\sigma^2 \\exp \\bigg(- \\frac{1}{2} r^2 \\bigg)
 
     """
     _support_GPU = True
diff --git a/GPy/kern/src/sde_brownian.py b/GPy/kern/src/sde_brownian.py
index b76761a0..8c0d6efb 100644
--- a/GPy/kern/src/sde_brownian.py
+++ b/GPy/kern/src/sde_brownian.py
@@ -12,48 +12,48 @@ import numpy as np
 
 class sde_Brownian(Brownian):
     """
-    
+
     Class provide extra functionality to transfer this covariance function into
     SDE form.
-    
+
     Linear kernel:
 
     .. math::
 
-       k(x,y) = \sigma^2 min(x,y)
+       k(x,y) = \\sigma^2 min(x,y)
 
     """
-    
+
     def sde_update_gradient_full(self, gradients):
         """
         Update gradient in the order in which parameters are represented in the
         kernel
         """
-    
+
         self.variance.gradient = gradients[0]
-        
-    def sde(self): 
-        """ 
-        Return the state space representation of the covariance. 
-        """ 
-        
+
+    def sde(self):
+        """
+        Return the state space representation of the covariance.
+        """
+
         variance = float(self.variance.values) # this is initial variancve in Bayesian linear regression
-        
+
         F = np.array( ((0,1.0),(0,0) ))
         L = np.array( ((1.0,),(0,)) )
         Qc = np.array( ((variance,),) )
         H = np.array( ((1.0,0),) )
-        
+
         Pinf   = np.array( ( (0, -0.5*variance ), (-0.5*variance, 0) ) )
-        #P0 = Pinf.copy() 
-        P0 = np.zeros((2,2))   
+        #P0 = Pinf.copy()
+        P0 = np.zeros((2,2))
         #Pinf   = np.array( ( (t0, 1.0), (1.0, 1.0/t0) ) ) * variance
         dF = np.zeros((2,2,1))
         dQc    = np.ones( (1,1,1) )
-        
+
         dPinf = np.zeros((2,2,1))
         dPinf[:,:,0] = np.array( ( (0, -0.5), (-0.5, 0) ) )
-        #dP0 = dPinf.copy() 
+        #dP0 = dPinf.copy()
         dP0 = np.zeros((2,2,1))
-  
+
         return (F, L, Qc, H, Pinf, P0, dF, dQc, dPinf, dP0)
diff --git a/GPy/kern/src/sde_linear.py b/GPy/kern/src/sde_linear.py
index 943e3bd7..59017068 100644
--- a/GPy/kern/src/sde_linear.py
+++ b/GPy/kern/src/sde_linear.py
@@ -11,15 +11,15 @@ import numpy as np
 
 class sde_Linear(Linear):
     """
-    
+
     Class provide extra functionality to transfer this covariance function into
     SDE form.
-    
+
     Linear kernel:
 
     .. math::
 
-       k(x,y) = \sum_{i=1}^{input dim} \sigma^2_i x_iy_i
+       k(x,y) = \\sum_{i=1}^{input dim} \\sigma^2_i x_iy_i
 
     """
     def __init__(self, input_dim, X, variances=None, ARD=False, active_dims=None, name='linear'):
@@ -27,40 +27,40 @@ class sde_Linear(Linear):
         Modify the init method, because one extra parameter is required. X - points
         on the X axis.
         """
-        
+
         super(sde_Linear, self).__init__(input_dim, variances, ARD, active_dims, name)
-        
+
         self.t0 = np.min(X)
-        
-    
+
+
     def sde_update_gradient_full(self, gradients):
         """
         Update gradient in the order in which parameters are represented in the
         kernel
         """
-    
+
         self.variances.gradient = gradients[0]
-        
-    def sde(self): 
-        """ 
-        Return the state space representation of the covariance. 
-        """ 
-        
+
+    def sde(self):
+        """
+        Return the state space representation of the covariance.
+        """
+
         variance = float(self.variances.values) # this is initial variancve in Bayesian linear regression
         t0 = float(self.t0)
-        
+
         F = np.array( ((0,1.0),(0,0) ))
         L = np.array( ((0,),(1.0,)) )
         Qc = np.zeros((1,1))
         H = np.array( ((1.0,0),) )
-        
+
         Pinf   = np.zeros((2,2))
-        P0 = np.array( ( (t0**2, t0), (t0, 1) ) ) * variance        
+        P0 = np.array( ( (t0**2, t0), (t0, 1) ) ) * variance
         dF = np.zeros((2,2,1))
         dQc    = np.zeros( (1,1,1) )
-        
+
         dPinf = np.zeros((2,2,1))
         dP0 = np.zeros((2,2,1))
         dP0[:,:,0]  = P0 / variance
-  
+
         return (F, L, Qc, H, Pinf, P0, dF, dQc, dPinf, dP0)
diff --git a/GPy/kern/src/sde_matern.py b/GPy/kern/src/sde_matern.py
index fe302753..f896957e 100644
--- a/GPy/kern/src/sde_matern.py
+++ b/GPy/kern/src/sde_matern.py
@@ -19,7 +19,7 @@ class sde_Matern32(Matern32):
 
     .. math::
 
-       k(r) = \sigma^2 (1 + \sqrt{3} r) \exp(- \sqrt{3} r) \\ \\ \\ \\  \text{ where  } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
+       k(r) = \\sigma^2 (1 + \\sqrt{3} r) \\exp(- \\sqrt{3} r) \\ \\ \\ \\  \text{ where  } r = \\sqrt{\\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
     def sde_update_gradient_full(self, gradients):
@@ -79,7 +79,7 @@ class sde_Matern52(Matern52):
 
     .. math::
 
-       k(r) = \sigma^2 (1 + \sqrt{5} r + \frac{5}{3}r^2) \exp(- \sqrt{5} r) \\ \\ \\ \\  \text{ where  } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
+       k(r) = \\sigma^2 (1 + \\sqrt{5} r + \frac{5}{3}r^2) \\exp(- \\sqrt{5} r) \\ \\ \\ \\  \text{ where  } r = \\sqrt{\\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
     def sde_update_gradient_full(self, gradients):
diff --git a/GPy/kern/src/sde_standard_periodic.py b/GPy/kern/src/sde_standard_periodic.py
index 2963eb66..cf6ab30d 100644
--- a/GPy/kern/src/sde_standard_periodic.py
+++ b/GPy/kern/src/sde_standard_periodic.py
@@ -24,8 +24,8 @@ class sde_StdPeriodic(StdPeriodic):
 
     .. math::
 
-       k(x,y) = \theta_1 \exp \left[  - \frac{1}{2} {}\sum_{i=1}^{input\_dim}
-       \left( \frac{\sin(\frac{\pi}{\lambda_i} (x_i - y_i) )}{l_i} \right)^2 \right] }
+       k(x,y) = \theta_1 \\exp \\left[  - \frac{1}{2} {}\\sum_{i=1}^{input\\_dim}
+       \\left( \frac{\\sin(\frac{\\pi}{\\lambda_i} (x_i - y_i) )}{l_i} \right)^2 \right] }
 
     """
 
@@ -177,7 +177,7 @@ def seriescoeff(m=6, lengthScale=1.0, magnSigma2=1.0, true_covariance=False):
     Calculate the coefficients q_j^2 for the covariance function
     approximation:
 
-        k(\tau) =  \sum_{j=0}^{+\infty} q_j^2 \cos(j\omega_0 \tau)
+        k(\tau) =  \\sum_{j=0}^{+\\infty} q_j^2 \\cos(j\\omega_0 \tau)
 
     Reference is:
 
diff --git a/GPy/kern/src/sde_static.py b/GPy/kern/src/sde_static.py
index 6a30c693..93a13b3e 100644
--- a/GPy/kern/src/sde_static.py
+++ b/GPy/kern/src/sde_static.py
@@ -20,7 +20,7 @@ class sde_White(White):
 
     .. math::
 
-       k(x,y) = \alpha*\delta(x-y)
+       k(x,y) = \alpha*\\delta(x-y)
 
     """
     
diff --git a/GPy/kern/src/sde_stationary.py b/GPy/kern/src/sde_stationary.py
index 61516799..58bb7424 100644
--- a/GPy/kern/src/sde_stationary.py
+++ b/GPy/kern/src/sde_stationary.py
@@ -29,7 +29,7 @@ class sde_RBF(RBF):
 
     .. math::
 
-        k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r^2 \\bigg) \\ \\ \\ \\  \text{ where  } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
+        k(r) = \\sigma^2 \\exp \\bigg(- \\frac{1}{2} r^2 \\bigg) \\ \\ \\ \\  \text{ where  } r = \\sqrt{\\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
 
@@ -204,7 +204,7 @@ class sde_Exponential(Exponential):
 
     .. math::
 
-       k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r \\bigg) \\ \\ \\ \\  \text{ where  } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
+       k(r) = \\sigma^2 \\exp \\bigg(- \\frac{1}{2} r \\bigg) \\ \\ \\ \\  \text{ where  } r = \\sqrt{\\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
 
@@ -259,7 +259,7 @@ class sde_RatQuad(RatQuad):
 
     .. math::
 
-       k(r) = \sigma^2 \\bigg( 1 + \\frac{r^2}{2} \\bigg)^{- \alpha} \\ \\ \\ \\  \text{ where  } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
+       k(r) = \\sigma^2 \\bigg( 1 + \\frac{r^2}{2} \\bigg)^{- \alpha} \\ \\ \\ \\  \text{ where  } r = \\sqrt{\\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
 
diff --git a/GPy/kern/src/standard_periodic.py b/GPy/kern/src/standard_periodic.py
index c8b91563..4baf86ee 100644
--- a/GPy/kern/src/standard_periodic.py
+++ b/GPy/kern/src/standard_periodic.py
@@ -24,19 +24,19 @@ class StdPeriodic(Kern):
 
     .. math::
 
-       k(x,y) = \theta_1 \exp \left[  - \frac{1}{2} \sum_{i=1}^{input\_dim}
-       \left( \frac{\sin(\frac{\pi}{T_i} (x_i - y_i) )}{l_i} \right)^2 \right] }
+       k(x,y) = \theta_1 \\exp \\left[  - \frac{1}{2} \\sum_{i=1}^{input\\_dim}
+       \\left( \frac{\\sin(\frac{\\pi}{T_i} (x_i - y_i) )}{l_i} \right)^2 \right] }
 
     :param input_dim: the number of input dimensions
     :type input_dim: int
     :param variance: the variance :math:`\theta_1` in the formula above
     :type variance: float
-    :param period: the vector of periods :math:`\T_i`. If None then 1.0 is assumed.
+    :param period: the vector of periods :math:`\\T_i`. If None then 1.0 is assumed.
     :type period: array or list of the appropriate size (or float if there is only one period parameter)
-    :param lengthscale: the vector of lengthscale :math:`\l_i`. If None then 1.0 is assumed.
+    :param lengthscale: the vector of lengthscale :math:`\\l_i`. If None then 1.0 is assumed.
     :type lengthscale: array or list of the appropriate size (or float if there is only one lengthscale parameter)
     :param ARD1: Auto Relevance Determination with respect to period.
-        If equal to "False" one single period parameter :math:`\T_i` for
+        If equal to "False" one single period parameter :math:`\\T_i` for
         each dimension is assumed, otherwise there is one lengthscale
         parameter per dimension.
     :type ARD1: Boolean
diff --git a/GPy/kern/src/stationary.py b/GPy/kern/src/stationary.py
index e3931008..bb8c6bcb 100644
--- a/GPy/kern/src/stationary.py
+++ b/GPy/kern/src/stationary.py
@@ -35,7 +35,7 @@ class Stationary(Kern):
 
     .. math::
 
-        r(x, x') = \\sqrt{ \\sum_{q=1}^Q \\frac{(x_q - x'_q)^2}{\ell_q^2} }.
+        r(x, x') = \\sqrt{ \\sum_{q=1}^Q \\frac{(x_q - x'_q)^2}{\\ell_q^2} }.
 
     By default, there's only one lengthscale: seaprate lengthscales for each
     dimension can be enables by setting ARD=True.
@@ -153,7 +153,7 @@ class Stationary(Kern):
         Efficiently compute the scaled distance, r.
 
         ..math::
-            r = \sqrt( \sum_{q=1}^Q (x_q - x'q)^2/l_q^2 )
+            r = \\sqrt( \\sum_{q=1}^Q (x_q - x'q)^2/l_q^2 )
 
         Note that if thre is only one lengthscale, l comes outside the sum. In
         this case we compute the unscaled distance first (in a separate
@@ -259,7 +259,7 @@ class Stationary(Kern):
         the returned array is of shape [NxNxQxQ].
 
         ..math:
-            \frac{\partial^2 K}{\partial X2 ^2} = - \frac{\partial^2 K}{\partial X\partial X2}
+            \frac{\\partial^2 K}{\\partial X2 ^2} = - \frac{\\partial^2 K}{\\partial X\\partial X2}
 
         ..returns:
             dL2_dXdX2:  [NxMxQxQ] in the cov=True case, or [NxMxQ] in the cov=False case,
@@ -295,7 +295,7 @@ class Stationary(Kern):
         Given the derivative of the objective dL_dK, compute the second derivative of K wrt X:
 
         ..math:
-          \frac{\partial^2 K}{\partial X\partial X}
+          \frac{\\partial^2 K}{\\partial X\\partial X}
 
         ..returns:
             dL2_dXdX: [NxQxQ]
@@ -423,7 +423,7 @@ class OU(Stationary):
 
     .. math::
 
-       k(r) = \\sigma^2 \exp(- r) \\ \\ \\ \\  \\text{ where  } r = \sqrt{\sum_{i=1}^{\text{input_dim}} \\frac{(x_i-y_i)^2}{\ell_i^2} }
+       k(r) = \\sigma^2 \\exp(- r) \\ \\ \\ \\  \\text{ where  } r = \\sqrt{\\sum_{i=1}^{\text{input_dim}} \\frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
 
@@ -460,7 +460,7 @@ class Matern32(Stationary):
 
     .. math::
 
-       k(r) = \\sigma^2 (1 + \\sqrt{3} r) \exp(- \sqrt{3} r) \\ \\ \\ \\  \\text{ where  } r = \sqrt{\sum_{i=1}^{\\text{input_dim}} \\frac{(x_i-y_i)^2}{\ell_i^2} }
+       k(r) = \\sigma^2 (1 + \\sqrt{3} r) \\exp(- \\sqrt{3} r) \\ \\ \\ \\  \\text{ where  } r = \\sqrt{\\sum_{i=1}^{\\text{input_dim}} \\frac{(x_i-y_i)^2}{\\ell_i^2} }
 
     """
 
@@ -559,7 +559,7 @@ class Matern52(Stationary):
 
     .. math::
 
-       k(r) = \sigma^2 (1 + \sqrt{5} r + \\frac53 r^2) \exp(- \sqrt{5} r)
+       k(r) = \\sigma^2 (1 + \\sqrt{5} r + \\frac53 r^2) \\exp(- \\sqrt{5} r)
     """
     def __init__(self, input_dim, variance=1., lengthscale=None, ARD=False, active_dims=None, name='Mat52'):
         super(Matern52, self).__init__(input_dim, variance, lengthscale, ARD, active_dims, name)
@@ -626,7 +626,7 @@ class ExpQuad(Stationary):
 
     .. math::
 
-       k(r) = \sigma^2 \exp(- 0.5 r^2)
+       k(r) = \\sigma^2 \\exp(- 0.5 r^2)
 
     notes::
      - This is exactly the same as the RBF covariance function, but the
@@ -664,10 +664,10 @@ class ExpQuad(Stationary):
 class Cosine(Stationary):
     """
     Cosine Covariance function
-    
+
     .. math::
 
-        k(r) = \sigma^2 \cos(r)
+        k(r) = \\sigma^2 \\cos(r)
 
     """
     def __init__(self, input_dim, variance=1., lengthscale=None, ARD=False, active_dims=None, name='Cosine'):
@@ -682,18 +682,18 @@ class Cosine(Stationary):
 class ExpQuadCosine(Stationary):
     """
     Exponentiated quadratic multiplied by cosine covariance function (spectral mixture kernel).
-    
+
     .. math::
 
-        k(r) = \sigma^2 \exp(-2\pi^2r^2)\cos(2\pi r/T)
+        k(r) = \\sigma^2 \\exp(-2\\pi^2r^2)\\cos(2\\pi r/T)
 
     """
-    
+
     def __init__(self, input_dim, variance=1., lengthscale=None, period=1., ARD=False, active_dims=None, name='ExpQuadCosine'):
         super(ExpQuadCosine, self).__init__(input_dim, variance, lengthscale, ARD, active_dims, name)
         self.period = Param('period', period, Logexp())
         self.link_parameters(self.period)
-        
+
     def K_of_r(self, r):
         return self.variance * np.exp(-2*np.pi**2*r**2)*np.cos(2*np.pi*r/self.period)
 
@@ -712,18 +712,18 @@ class ExpQuadCosine(Stationary):
         super(ExpQuadCosine, self).update_gradients_diag(dL_dKdiag, X)
         self.period.gradient = 0.
 
-    
-    
+
+
 class Sinc(Stationary):
     """
     Sinc Covariance function
-    
+
     .. math::
 
-        k(r) = \sigma^2 \sinc(\pi r)
+        k(r) = \\sigma^2 \\sinc(\\pi r)
 
     """
-    
+
     def __init__(self, input_dim, variance=1., lengthscale=None, ARD=False, active_dims=None, name='Sinc'):
         super(Sinc, self).__init__(input_dim, variance, lengthscale, ARD, active_dims, name)
 
@@ -734,7 +734,7 @@ class Sinc(Stationary):
         # small angle approximation to avoid divide by zero errors.
         return np.where(r<1e-5, -self.variance*4/3*np.pi*np.pi*r, self.variance/r * (np.cos(2*np.pi*r)-np.sinc(2*r)))
 
-    
+
 
 class RatQuad(Stationary):
     """
@@ -742,7 +742,7 @@ class RatQuad(Stationary):
 
     .. math::
 
-       k(r) = \sigma^2 \\bigg( 1 + \\frac{r^2}{2} \\bigg)^{- \\alpha}
+       k(r) = \\sigma^2 \\bigg( 1 + \\frac{r^2}{2} \\bigg)^{- \\alpha}
 
     """
 
diff --git a/GPy/kern/src/todo/gibbs.py b/GPy/kern/src/todo/gibbs.py
index 68241245..1626af1b 100644
--- a/GPy/kern/src/todo/gibbs.py
+++ b/GPy/kern/src/todo/gibbs.py
@@ -9,13 +9,13 @@ import GPy
 
 class Gibbs(Kernpart):
     """
-    Gibbs non-stationary covariance function. 
+    Gibbs non-stationary covariance function.
 
     .. math::
-       
+
        r = sqrt((x_i - x_j)'*(x_i - x_j))
-       
-       k(x_i, x_j) = \sigma^2*Z*exp(-r^2/(l(x)*l(x) + l(x')*l(x')))
+
+       k(x_i, x_j) = \\sigma^2*Z*exp(-r^2/(l(x)*l(x) + l(x')*l(x')))
 
        Z = (2*l(x)*l(x')/(l(x)*l(x) + l(x')*l(x')^{q/2}
 
@@ -25,18 +25,18 @@ class Gibbs(Kernpart):
         with input location. This leads to an additional term in front of
         the kernel.
 
-        The parameters are :math:`\sigma^2`, the process variance, and
+        The parameters are :math:`\\sigma^2`, the process variance, and
         the parameters of l(x) which is a function that can be
         specified by the user, by default an multi-layer peceptron is
         used.
 
         :param input_dim: the number of input dimensions
-        :type input_dim: int 
-        :param variance: the variance :math:`\sigma^2`
+        :type input_dim: int
+        :param variance: the variance :math:`\\sigma^2`
         :type variance: float
         :param mapping: the mapping that gives the lengthscale across the input space (by default GPy.mappings.MLP is used with 20 hidden nodes).
         :type mapping: GPy.core.Mapping
-        :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one weight variance parameter \sigma^2_w), otherwise there is one weight variance parameter per dimension.
+        :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one weight variance parameter \\sigma^2_w), otherwise there is one weight variance parameter per dimension.
         :type ARD: Boolean
         :rtype: Kernpart object
 
@@ -113,7 +113,7 @@ class Gibbs(Kernpart):
             target += 2.*self.mapping.df_dX(self._dL_dl[:, None], X)
         else:
             target += self.mapping.df_dX(self._dL_dl[:, None], X)
-    
+
     def dKdiag_dX(self, dL_dKdiag, X, target):
         """Gradient of diagonal of covariance with respect to X."""
         pass
@@ -123,7 +123,7 @@ class Gibbs(Kernpart):
         target[0] += np.sum(dL_dKdiag)
 
 
-    
+
     def _K_computations(self, X, X2=None):
         """Pre-computations for the covariance function (used both when computing the covariance and its gradients). Here self._dK_dvar and self._K_dist2 are updated."""
         self._lengthscales=self.mapping.f(X)
@@ -146,7 +146,7 @@ class Gibbs(Kernpart):
         """Pre-computations for the gradients of the covaraince function. Here the gradient of the covariance with respect to all the individual lengthscales is computed.
         :param dL_dK: the gradient of the objective with respect to the covariance function.
         :type dL_dK: ndarray"""
-        
+
         self._dL_dl = (dL_dK*self.variance*self._K_dvar*(self.input_dim/2.*(self._lengthscales_two.T**4 - self._lengthscales**4) + 2*self._lengthscales2*self._K_dist2)/(self._w2*self._w2*self._lengthscales)).sum(1)
         if self._lengthscales_two is self._lengthscales:
             self._dL_dl_two = None
diff --git a/GPy/kern/src/todo/hetero.py b/GPy/kern/src/todo/hetero.py
index 507f6251..63a1600d 100644
--- a/GPy/kern/src/todo/hetero.py
+++ b/GPy/kern/src/todo/hetero.py
@@ -19,11 +19,11 @@ class Hetero(Kernpart):
 
     .. math::
 
-       k(x_i, x_j) = \delta_{i,j} \sigma^2(x_i)
+       k(x_i, x_j) = \\delta_{i,j} \\sigma^2(x_i)
 
-       where :math:`\sigma^2(x)` is a function giving the variance  as a function of input space and :math:`\delta_{i,j}` is the Kronecker delta function.
+       where :math:`\\sigma^2(x)` is a function giving the variance  as a function of input space and :math:`\\delta_{i,j}` is the Kronecker delta function.
 
-    The parameters are the parameters of \sigma^2(x) which is a
+    The parameters are the parameters of \\sigma^2(x) which is a
     function that can be specified by the user, by default an
     multi-layer peceptron is used.
 
diff --git a/GPy/kern/src/todo/poly.py b/GPy/kern/src/todo/poly.py
index 0deb11f4..43945032 100644
--- a/GPy/kern/src/todo/poly.py
+++ b/GPy/kern/src/todo/poly.py
@@ -11,28 +11,28 @@ class POLY(Kernpart):
     Polynomial kernel parameter initialisation.  Included for completeness, but generally not recommended, is the polynomial kernel:
 
     .. math::
-        k(x, y) = \sigma^2\*(\sigma_w^2 x'y+\sigma_b^b)^d
+        k(x, y) = \\sigma^2\\*(\\sigma_w^2 x'y+\\sigma_b^b)^d
 
-    The kernel parameters are :math:`\sigma^2` (variance), :math:`\sigma^2_w`
-    (weight_variance), :math:`\sigma^2_b` (bias_variance) and d
+    The kernel parameters are :math:`\\sigma^2` (variance), :math:`\\sigma^2_w`
+    (weight_variance), :math:`\\sigma^2_b` (bias_variance) and d
     (degree). Only gradients of the first three are provided for
     kernel optimisation, it is assumed that polynomial degree would
     be set by hand.
 
     The kernel is not recommended as it is badly behaved when the
-    :math:`\sigma^2_w\*x'\*y + \sigma^2_b` has a magnitude greater than one. For completeness
+    :math:`\\sigma^2_w\\*x'\\*y + \\sigma^2_b` has a magnitude greater than one. For completeness
     there is an automatic relevance determination version of this
     kernel provided (NOTE YET IMPLEMENTED!).
     :param input_dim: the number of input dimensions
-    :type input_dim: int 
-    :param variance: the variance :math:`\sigma^2`
+    :type input_dim: int
+    :param variance: the variance :math:`\\sigma^2`
     :type variance: float
-    :param weight_variance: the vector of the variances of the prior over input weights in the neural network :math:`\sigma^2_w`
+    :param weight_variance: the vector of the variances of the prior over input weights in the neural network :math:`\\sigma^2_w`
     :type weight_variance: array or list of the appropriate size (or float if there is only one weight variance parameter)
-    :param bias_variance: the variance of the prior over bias parameters :math:`\sigma^2_b`
+    :param bias_variance: the variance of the prior over bias parameters :math:`\\sigma^2_b`
     :param degree: the degree of the polynomial.
     :type degree: int
-    :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one weight variance parameter :math:`\sigma^2_w`), otherwise there is one weight variance parameter per dimension.
+    :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one weight variance parameter :math:`\\sigma^2_w`), otherwise there is one weight variance parameter per dimension.
     :type ARD: Boolean
     :rtype: Kernpart object
 
@@ -93,7 +93,7 @@ class POLY(Kernpart):
         base_cov_grad = base*dL_dK
 
 
-            
+
         target[0] += np.sum(self._K_dvar*dL_dK)
         target[1] += (self._K_inner_prod*base_cov_grad).sum()
         target[2] += base_cov_grad.sum()
@@ -107,14 +107,14 @@ class POLY(Kernpart):
             target += 2*self.weight_variance*self.degree*self.variance*(((X[None,:, :])) *(arg**(self.degree-1))[:, :, None]*dL_dK[:, :, None]).sum(1)
         else:
             target += self.weight_variance*self.degree*self.variance*(((X2[None,:, :])) *(arg**(self.degree-1))[:, :, None]*dL_dK[:, :, None]).sum(1)
-            
+
     def dKdiag_dX(self, dL_dKdiag, X, target):
         """Gradient of diagonal of covariance with respect to X"""
         self._K_diag_computations(X)
         arg = self._K_diag_poly_arg
         target += 2.*self.weight_variance*self.degree*self.variance*X*dL_dKdiag[:, None]*(arg**(self.degree-1))[:, None]
-    
-    
+
+
     def _K_computations(self, X, X2):
         if self.ARD:
             pass
@@ -133,6 +133,6 @@ class POLY(Kernpart):
             self._K_diag_poly_arg = (X*X).sum(1)*self.weight_variance + self.bias_variance
         self._K_diag_dvar = self._K_diag_poly_arg**self.degree
 
-  
+
 
 
diff --git a/GPy/kern/src/todo/rbf_inv.py b/GPy/kern/src/todo/rbf_inv.py
index 8405ae84..56ee7b79 100644
--- a/GPy/kern/src/todo/rbf_inv.py
+++ b/GPy/kern/src/todo/rbf_inv.py
@@ -15,9 +15,9 @@ class RBFInv(RBF):
 
     .. math::
 
-       k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r^2 \\bigg) \ \ \ \ \  \\text{ where  } r^2 = \sum_{i=1}^d \\frac{ (x_i-x^\prime_i)^2}{\ell_i^2}
+       k(r) = \\sigma^2 \\exp \\bigg(- \\frac{1}{2} r^2 \\bigg) \\ \\ \\ \\ \\  \\text{ where  } r^2 = \\sum_{i=1}^d \\frac{ (x_i-x^\\prime_i)^2}{\\ell_i^2}
 
-    where \ell_i is the lengthscale, \sigma^2 the variance and d the dimensionality of the input.
+    where \\ell_i is the lengthscale, \\sigma^2 the variance and d the dimensionality of the input.
 
     :param input_dim: the number of input dimensions
     :type input_dim: int
@@ -25,7 +25,7 @@ class RBFInv(RBF):
     :type variance: float
     :param lengthscale: the vector of lengthscale of the kernel
     :type lengthscale: array or list of the appropriate size (or float if there is only one lengthscale parameter)
-    :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one single lengthscale parameter \ell), otherwise there is one lengthscale parameter per dimension.
+    :param ARD: Auto Relevance Determination. If equal to "False", the kernel is isotropic (ie. one single lengthscale parameter \\ell), otherwise there is one lengthscale parameter per dimension.
     :type ARD: Boolean
     :rtype: kernel object
 
diff --git a/GPy/kern/src/trunclinear.py b/GPy/kern/src/trunclinear.py
index bb94ae73..917ba413 100644
--- a/GPy/kern/src/trunclinear.py
+++ b/GPy/kern/src/trunclinear.py
@@ -14,15 +14,15 @@ class TruncLinear(Kern):
 
     .. math::
 
-       k(x,y) = \sum_{i=1}^input_dim \sigma^2_i \max(0, x_iy_i - \sigma_q)
+       k(x,y) = \\sum_{i=1}^input_dim \\sigma^2_i \\max(0, x_iy_i - \\sigma_q)
 
     :param input_dim: the number of input dimensions
     :type input_dim: int
-    :param variances: the vector of variances :math:`\sigma^2_i`
+    :param variances: the vector of variances :math:`\\sigma^2_i`
     :type variances: array or list of the appropriate size (or float if there
                      is only one variance parameter)
     :param ARD: Auto Relevance Determination. If False, the kernel has only one
-                variance parameter \sigma^2, otherwise there is one variance
+                variance parameter \\sigma^2, otherwise there is one variance
                 parameter per dimension.
     :type ARD: Boolean
     :rtype: kernel object
@@ -113,15 +113,15 @@ class TruncLinear_inf(Kern):
 
     .. math::
 
-       k(x,y) = \sum_{i=1}^input_dim \sigma^2_i \max(0, x_iy_i - \sigma_q)
+       k(x,y) = \\sum_{i=1}^input_dim \\sigma^2_i \\max(0, x_iy_i - \\sigma_q)
 
     :param input_dim: the number of input dimensions
     :type input_dim: int
-    :param variances: the vector of variances :math:`\sigma^2_i`
+    :param variances: the vector of variances :math:`\\sigma^2_i`
     :type variances: array or list of the appropriate size (or float if there
                      is only one variance parameter)
     :param ARD: Auto Relevance Determination. If False, the kernel has only one
-                variance parameter \sigma^2, otherwise there is one variance
+                variance parameter \\sigma^2, otherwise there is one variance
                 parameter per dimension.
     :type ARD: Boolean
     :rtype: kernel object
diff --git a/GPy/likelihoods/exponential.py b/GPy/likelihoods/exponential.py
index ecf0977e..ef071fdf 100644
--- a/GPy/likelihoods/exponential.py
+++ b/GPy/likelihoods/exponential.py
@@ -14,7 +14,7 @@ class Exponential(Likelihood):
     Y is expected to take values in {0,1,2,...}
     -----
     $$
-    L(x) = \exp(\lambda) * \lambda**Y_i / Y_i!
+    L(x) = \\exp(\\lambda) * \\lambda**Y_i / Y_i!
     $$
     """
     def __init__(self,gp_link=None):
@@ -46,7 +46,7 @@ class Exponential(Likelihood):
         Log Likelihood Function given link(f)
 
         .. math::
-            \\ln p(y_{i}|\lambda(f_{i})) = \\ln \\lambda(f_{i}) - y_{i}\\lambda(f_{i})
+            \\ln p(y_{i}|\\lambda(f_{i})) = \\ln \\lambda(f_{i}) - y_{i}\\lambda(f_{i})
 
         :param link_f: latent variables (link(f))
         :type link_f: Nx1 array
@@ -65,7 +65,7 @@ class Exponential(Likelihood):
         Gradient of the log likelihood function at y, given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d \\ln p(y_{i}|\lambda(f_{i}))}{d\\lambda(f)} = \\frac{1}{\\lambda(f)} - y_{i}
+            \\frac{d \\ln p(y_{i}|\\lambda(f_{i}))}{d\\lambda(f)} = \\frac{1}{\\lambda(f)} - y_{i}
 
         :param link_f: latent variables (f)
         :type link_f: Nx1 array
@@ -87,7 +87,7 @@ class Exponential(Likelihood):
         The hessian will be 0 unless i == j
 
         .. math::
-            \\frac{d^{2} \\ln p(y_{i}|\lambda(f_{i}))}{d^{2}\\lambda(f)} = -\\frac{1}{\\lambda(f_{i})^{2}}
+            \\frac{d^{2} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{2}\\lambda(f)} = -\\frac{1}{\\lambda(f_{i})^{2}}
 
         :param link_f: latent variables link(f)
         :type link_f: Nx1 array
@@ -110,7 +110,7 @@ class Exponential(Likelihood):
         Third order derivative log-likelihood function at y given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d^{3} \\ln p(y_{i}|\lambda(f_{i}))}{d^{3}\\lambda(f)} = \\frac{2}{\\lambda(f_{i})^{3}}
+            \\frac{d^{3} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{3}\\lambda(f)} = \\frac{2}{\\lambda(f_{i})^{3}}
 
         :param link_f: latent variables link(f)
         :type link_f: Nx1 array
diff --git a/GPy/likelihoods/gamma.py b/GPy/likelihoods/gamma.py
index 79aba4a5..18ad580c 100644
--- a/GPy/likelihoods/gamma.py
+++ b/GPy/likelihoods/gamma.py
@@ -54,7 +54,7 @@ class Gamma(Likelihood):
         Log Likelihood Function given link(f)
 
         .. math::
-            \\ln p(y_{i}|\lambda(f_{i})) = \\alpha_{i}\\log \\beta - \\log \\Gamma(\\alpha_{i}) + (\\alpha_{i} - 1)\\log y_{i} - \\beta y_{i}\\\\
+            \\ln p(y_{i}|\\lambda(f_{i})) = \\alpha_{i}\\log \\beta - \\log \\Gamma(\\alpha_{i}) + (\\alpha_{i} - 1)\\log y_{i} - \\beta y_{i}\\\\
             \\alpha_{i} = \\beta y_{i}
 
         :param link_f: latent variables (link(f))
@@ -101,7 +101,7 @@ class Gamma(Likelihood):
         The hessian will be 0 unless i == j
 
         .. math::
-            \\frac{d^{2} \\ln p(y_{i}|\lambda(f_{i}))}{d^{2}\\lambda(f)} = -\\beta^{2}\\frac{d\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
+            \\frac{d^{2} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{2}\\lambda(f)} = -\\beta^{2}\\frac{d\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
             \\alpha_{i} = \\beta y_{i}
 
         :param link_f: latent variables link(f)
@@ -126,7 +126,7 @@ class Gamma(Likelihood):
         Third order derivative log-likelihood function at y given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d^{3} \\ln p(y_{i}|\lambda(f_{i}))}{d^{3}\\lambda(f)} = -\\beta^{3}\\frac{d^{2}\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
+            \\frac{d^{3} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{3}\\lambda(f)} = -\\beta^{3}\\frac{d^{2}\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
             \\alpha_{i} = \\beta y_{i}
 
         :param link_f: latent variables link(f)
diff --git a/GPy/likelihoods/likelihood.py b/GPy/likelihoods/likelihood.py
index 0f17a883..bbfc81ad 100644
--- a/GPy/likelihoods/likelihood.py
+++ b/GPy/likelihoods/likelihood.py
@@ -130,7 +130,7 @@ class Likelihood(Parameterized):
         Calculation of the log predictive density
 
         .. math:
-            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\mu_{*}\\sigma^{2}_{*})
+            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\\mu_{*}\\sigma^{2}_{*})
 
         :param y_test: test observations (y_{*})
         :type y_test: (Nx1) array
@@ -199,7 +199,7 @@ class Likelihood(Parameterized):
 
         .. math:
             log p(y_{*}|D) = log 1/num_samples prod^{S}_{s=1} p(y_{*}|f_{*s})
-            f_{*s} ~ p(f_{*}|\mu_{*}\\sigma^{2}_{*})
+            f_{*s} ~ p(f_{*}|\\mu_{*}\\sigma^{2}_{*})
 
         :param y_test: test observations (y_{*})
         :type y_test: (Nx1) array
diff --git a/GPy/likelihoods/link_functions.py b/GPy/likelihoods/link_functions.py
index 0554bd93..dfb8f3e4 100644
--- a/GPy/likelihoods/link_functions.py
+++ b/GPy/likelihoods/link_functions.py
@@ -145,7 +145,7 @@ class ScaledProbit(Probit):
     """
     def __init__(self, nu=1.):
         self.nu = float(nu)
-        
+
     def transf(self,f):
         return std_norm_cdf(f*self.nu)
 
@@ -157,7 +157,7 @@ class ScaledProbit(Probit):
 
     def d3transf_df3(self,f):
         return (safe_square(f*self.nu)-1.)*std_norm_pdf(f*self.nu)*(self.nu**3)
-    
+
     def to_dict(self):
         """
         Convert the object into a json serializable dictionary.
@@ -180,7 +180,7 @@ class Cloglog(GPTransformation):
 
         or
 
-        f = \log (-\log(1-p))
+        f = \\log (-\\log(1-p))
 
     """
     def transf(self,f):
diff --git a/GPy/likelihoods/poisson.py b/GPy/likelihoods/poisson.py
index d3eef7a4..5f2fe39e 100644
--- a/GPy/likelihoods/poisson.py
+++ b/GPy/likelihoods/poisson.py
@@ -54,7 +54,7 @@ class Poisson(Likelihood):
         Log Likelihood Function given link(f)
 
         .. math::
-            \\ln p(y_{i}|\lambda(f_{i})) = -\\lambda(f_{i}) + y_{i}\\log \\lambda(f_{i}) - \\log y_{i}!
+            \\ln p(y_{i}|\\lambda(f_{i})) = -\\lambda(f_{i}) + y_{i}\\log \\lambda(f_{i}) - \\log y_{i}!
 
         :param link_f: latent variables (link(f))
         :type link_f: Nx1 array
@@ -72,7 +72,7 @@ class Poisson(Likelihood):
         Gradient of the log likelihood function at y, given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d \\ln p(y_{i}|\lambda(f_{i}))}{d\\lambda(f)} = \\frac{y_{i}}{\\lambda(f_{i})} - 1
+            \\frac{d \\ln p(y_{i}|\\lambda(f_{i}))}{d\\lambda(f)} = \\frac{y_{i}}{\\lambda(f_{i})} - 1
 
         :param link_f: latent variables (f)
         :type link_f: Nx1 array
@@ -92,7 +92,7 @@ class Poisson(Likelihood):
         The hessian will be 0 unless i == j
 
         .. math::
-            \\frac{d^{2} \\ln p(y_{i}|\lambda(f_{i}))}{d^{2}\\lambda(f)} = \\frac{-y_{i}}{\\lambda(f_{i})^{2}}
+            \\frac{d^{2} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{2}\\lambda(f)} = \\frac{-y_{i}}{\\lambda(f_{i})^{2}}
 
         :param link_f: latent variables link(f)
         :type link_f: Nx1 array
@@ -113,7 +113,7 @@ class Poisson(Likelihood):
         Third order derivative log-likelihood function at y given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d^{3} \\ln p(y_{i}|\lambda(f_{i}))}{d^{3}\\lambda(f)} = \\frac{2y_{i}}{\\lambda(f_{i})^{3}}
+            \\frac{d^{3} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{3}\\lambda(f)} = \\frac{2y_{i}}{\\lambda(f_{i})^{3}}
 
         :param link_f: latent variables link(f)
         :type link_f: Nx1 array
diff --git a/GPy/likelihoods/student_t.py b/GPy/likelihoods/student_t.py
index 6c97a5d8..4d3202fc 100644
--- a/GPy/likelihoods/student_t.py
+++ b/GPy/likelihoods/student_t.py
@@ -78,7 +78,7 @@ class StudentT(Likelihood):
         Log Likelihood Function given link(f)
 
         .. math::
-            \\ln p(y_{i}|\lambda(f_{i})) = \\ln \\Gamma\\left(\\frac{v+1}{2}\\right) - \\ln \\Gamma\\left(\\frac{v}{2}\\right) - \\ln \\sqrt{v \\pi\\sigma^{2}} - \\frac{v+1}{2}\\ln \\left(1 + \\frac{1}{v}\\left(\\frac{(y_{i} - \lambda(f_{i}))^{2}}{\\sigma^{2}}\\right)\\right)
+            \\ln p(y_{i}|\\lambda(f_{i})) = \\ln \\Gamma\\left(\\frac{v+1}{2}\\right) - \\ln \\Gamma\\left(\\frac{v}{2}\\right) - \\ln \\sqrt{v \\pi\\sigma^{2}} - \\frac{v+1}{2}\\ln \\left(1 + \\frac{1}{v}\\left(\\frac{(y_{i} - \\lambda(f_{i}))^{2}}{\\sigma^{2}}\\right)\\right)
 
         :param inv_link_f: latent variables (link(f))
         :type inv_link_f: Nx1 array
@@ -107,7 +107,7 @@ class StudentT(Likelihood):
         Gradient of the log likelihood function at y, given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d \\ln p(y_{i}|\lambda(f_{i}))}{d\\lambda(f)} = \\frac{(v+1)(y_{i}-\lambda(f_{i}))}{(y_{i}-\lambda(f_{i}))^{2} + \\sigma^{2}v}
+            \\frac{d \\ln p(y_{i}|\\lambda(f_{i}))}{d\\lambda(f)} = \\frac{(v+1)(y_{i}-\\lambda(f_{i}))}{(y_{i}-\\lambda(f_{i}))^{2} + \\sigma^{2}v}
 
         :param inv_link_f: latent variables (f)
         :type inv_link_f: Nx1 array
@@ -129,7 +129,7 @@ class StudentT(Likelihood):
         The hessian will be 0 unless i == j
 
         .. math::
-            \\frac{d^{2} \\ln p(y_{i}|\lambda(f_{i}))}{d^{2}\\lambda(f)} = \\frac{(v+1)((y_{i}-\lambda(f_{i}))^{2} - \\sigma^{2}v)}{((y_{i}-\lambda(f_{i}))^{2} + \\sigma^{2}v)^{2}}
+            \\frac{d^{2} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{2}\\lambda(f)} = \\frac{(v+1)((y_{i}-\\lambda(f_{i}))^{2} - \\sigma^{2}v)}{((y_{i}-\\lambda(f_{i}))^{2} + \\sigma^{2}v)^{2}}
 
         :param inv_link_f: latent variables inv_link(f)
         :type inv_link_f: Nx1 array
@@ -154,7 +154,7 @@ class StudentT(Likelihood):
         Third order derivative log-likelihood function at y given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d^{3} \\ln p(y_{i}|\lambda(f_{i}))}{d^{3}\\lambda(f)} = \\frac{-2(v+1)((y_{i} - \lambda(f_{i}))^3 - 3(y_{i} - \lambda(f_{i})) \\sigma^{2} v))}{((y_{i} - \lambda(f_{i})) + \\sigma^{2} v)^3}
+            \\frac{d^{3} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{3}\\lambda(f)} = \\frac{-2(v+1)((y_{i} - \\lambda(f_{i}))^3 - 3(y_{i} - \\lambda(f_{i})) \\sigma^{2} v))}{((y_{i} - \\lambda(f_{i})) + \\sigma^{2} v)^3}
 
         :param inv_link_f: latent variables link(f)
         :type inv_link_f: Nx1 array
@@ -175,7 +175,7 @@ class StudentT(Likelihood):
         Gradient of the log-likelihood function at y given f, w.r.t variance parameter (t_noise)
 
         .. math::
-            \\frac{d \\ln p(y_{i}|\lambda(f_{i}))}{d\\sigma^{2}} = \\frac{v((y_{i} - \lambda(f_{i}))^{2} - \\sigma^{2})}{2\\sigma^{2}(\\sigma^{2}v + (y_{i} - \lambda(f_{i}))^{2})}
+            \\frac{d \\ln p(y_{i}|\\lambda(f_{i}))}{d\\sigma^{2}} = \\frac{v((y_{i} - \\lambda(f_{i}))^{2} - \\sigma^{2})}{2\\sigma^{2}(\\sigma^{2}v + (y_{i} - \\lambda(f_{i}))^{2})}
 
         :param inv_link_f: latent variables link(f)
         :type inv_link_f: Nx1 array
@@ -199,7 +199,7 @@ class StudentT(Likelihood):
         Derivative of the dlogpdf_dlink w.r.t variance parameter (t_noise)
 
         .. math::
-            \\frac{d}{d\\sigma^{2}}(\\frac{d \\ln p(y_{i}|\lambda(f_{i}))}{df}) = \\frac{-2\\sigma v(v + 1)(y_{i}-\lambda(f_{i}))}{(y_{i}-\lambda(f_{i}))^2 + \\sigma^2 v)^2}
+            \\frac{d}{d\\sigma^{2}}(\\frac{d \\ln p(y_{i}|\\lambda(f_{i}))}{df}) = \\frac{-2\\sigma v(v + 1)(y_{i}-\\lambda(f_{i}))}{(y_{i}-\\lambda(f_{i}))^2 + \\sigma^2 v)^2}
 
         :param inv_link_f: latent variables inv_link_f
         :type inv_link_f: Nx1 array
@@ -220,7 +220,7 @@ class StudentT(Likelihood):
         Gradient of the hessian (d2logpdf_dlink2) w.r.t variance parameter (t_noise)
 
         .. math::
-            \\frac{d}{d\\sigma^{2}}(\\frac{d^{2} \\ln p(y_{i}|\lambda(f_{i}))}{d^{2}f}) = \\frac{v(v+1)(\\sigma^{2}v - 3(y_{i} - \lambda(f_{i}))^{2})}{(\\sigma^{2}v + (y_{i} - \lambda(f_{i}))^{2})^{3}}
+            \\frac{d}{d\\sigma^{2}}(\\frac{d^{2} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{2}f}) = \\frac{v(v+1)(\\sigma^{2}v - 3(y_{i} - \\lambda(f_{i}))^{2})}{(\\sigma^{2}v + (y_{i} - \\lambda(f_{i}))^{2})^{3}}
 
         :param inv_link_f: latent variables link(f)
         :type inv_link_f: Nx1 array
diff --git a/GPy/likelihoods/weibull.py b/GPy/likelihoods/weibull.py
index ba9eb540..61da4c66 100644
--- a/GPy/likelihoods/weibull.py
+++ b/GPy/likelihoods/weibull.py
@@ -54,7 +54,7 @@ class Weibull(Likelihood):
         Log Likelihood Function given link(f)
 
         .. math::
-            \\ln p(y_{i}|\lambda(f_{i})) = \\alpha_{i}\\log \\beta - \\log \\Gamma(\\alpha_{i}) + (\\alpha_{i} - 1)\\log y_{i} - \\beta y_{i}\\\\
+            \\ln p(y_{i}|\\lambda(f_{i})) = \\alpha_{i}\\log \\beta - \\log \\Gamma(\\alpha_{i}) + (\\alpha_{i} - 1)\\log y_{i} - \\beta y_{i}\\\\
             \\alpha_{i} = \\beta y_{i}
 
         :param link_f: latent variables (link(f))
@@ -117,7 +117,7 @@ class Weibull(Likelihood):
         The hessian will be 0 unless i == j
 
         .. math::
-            \\frac{d^{2} \\ln p(y_{i}|\lambda(f_{i}))}{d^{2}\\lambda(f)} = -\\beta^{2}\\frac{d\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
+            \\frac{d^{2} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{2}\\lambda(f)} = -\\beta^{2}\\frac{d\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
             \\alpha_{i} = \\beta y_{i}
 
         :param link_f: latent variables link(f)
@@ -150,7 +150,7 @@ class Weibull(Likelihood):
         Third order derivative log-likelihood function at y given link(f) w.r.t link(f)
 
         .. math::
-            \\frac{d^{3} \\ln p(y_{i}|\lambda(f_{i}))}{d^{3}\\lambda(f)} = -\\beta^{3}\\frac{d^{2}\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
+            \\frac{d^{3} \\ln p(y_{i}|\\lambda(f_{i}))}{d^{3}\\lambda(f)} = -\\beta^{3}\\frac{d^{2}\\Psi(\\alpha_{i})}{d\\alpha_{i}}\\\\
             \\alpha_{i} = \\beta y_{i}
 
         :param link_f: latent variables link(f)
diff --git a/GPy/mappings/additive.py b/GPy/mappings/additive.py
index 0fb72ca4..b75118e4 100644
--- a/GPy/mappings/additive.py
+++ b/GPy/mappings/additive.py
@@ -10,7 +10,7 @@ class Additive(Mapping):
 
     .. math::
 
-       f(\mathbf{x}*) = f_1(\mathbf{x}*) + f_2(\mathbf(x)*)
+       f(\\mathbf{x}*) = f_1(\\mathbf{x}*) + f_2(\\mathbf(x)*)
 
     :param mapping1: first mapping to add together.
     :type mapping1: GPy.mappings.Mapping
diff --git a/GPy/mappings/compound.py b/GPy/mappings/compound.py
index 9078910f..9c7dd381 100644
--- a/GPy/mappings/compound.py
+++ b/GPy/mappings/compound.py
@@ -9,7 +9,7 @@ class Compound(Mapping):
 
     .. math::
 
-       f(\mathbf{x}) = f_2(f_1(\mathbf{x}))
+       f(\\mathbf{x}) = f_2(f_1(\\mathbf{x}))
 
     :param mapping1: first mapping
     :type mapping1: GPy.mappings.Mapping
diff --git a/GPy/mappings/constant.py b/GPy/mappings/constant.py
index 281d72d2..85a5f695 100644
--- a/GPy/mappings/constant.py
+++ b/GPy/mappings/constant.py
@@ -9,7 +9,7 @@ class Constant(Mapping):
 
     .. math::
 
-       F(\mathbf{x}) = c
+       F(\\mathbf{x}) = c
 
 
     :param input_dim: dimension of input.
diff --git a/GPy/mappings/kernel.py b/GPy/mappings/kernel.py
index 18b77126..bde2c6a2 100644
--- a/GPy/mappings/kernel.py
+++ b/GPy/mappings/kernel.py
@@ -12,20 +12,20 @@ class Kernel(Mapping):
 
     .. math::
 
-       f(\mathbf{x}) = \sum_i \alpha_i k(\mathbf{z}_i, \mathbf{x})
+       f(\\mathbf{x}) = \\sum_i \alpha_i k(\\mathbf{z}_i, \\mathbf{x})
 
     or for multple outputs
 
     .. math::
 
-       f_i(\mathbf{x}) = \sum_j \alpha_{i,j} k(\mathbf{z}_i, \mathbf{x})
+       f_i(\\mathbf{x}) = \\sum_j \alpha_{i,j} k(\\mathbf{z}_i, \\mathbf{x})
 
 
     :param input_dim: dimension of input.
     :type input_dim: int
     :param output_dim: dimension of output.
     :type output_dim: int
-    :param Z: input observations containing :math:`\mathbf{Z}`
+    :param Z: input observations containing :math:`\\mathbf{Z}`
     :type Z: ndarray
     :param kernel: a GPy kernel, defaults to GPy.kern.RBF
     :type kernel: GPy.kern.kern
diff --git a/GPy/mappings/linear.py b/GPy/mappings/linear.py
index 82b7a0e0..c708f7c0 100644
--- a/GPy/mappings/linear.py
+++ b/GPy/mappings/linear.py
@@ -12,7 +12,7 @@ class Linear(Mapping):
 
     .. math::
 
-       F(\mathbf{x}) = \mathbf{A} \mathbf{x})
+       F(\\mathbf{x}) = \\mathbf{A} \\mathbf{x})
 
 
     :param input_dim: dimension of input.
diff --git a/GPy/models/gp_kronecker_gaussian_regression.py b/GPy/models/gp_kronecker_gaussian_regression.py
index 7898a97a..c1cd38bf 100644
--- a/GPy/models/gp_kronecker_gaussian_regression.py
+++ b/GPy/models/gp_kronecker_gaussian_regression.py
@@ -22,7 +22,7 @@ class GPKroneckerGaussianRegression(Model):
 
     .. rubric:: References
 
-    .. [stegle_et_al_2011] Stegle, O.; Lippert, C.; Mooij, J.M.; Lawrence, N.D.; Borgwardt, K.:Efficient inference in matrix-variate Gaussian models with \iid observation noise. In: Advances in Neural Information Processing Systems, 2011, Pages 630-638
+    .. [stegle_et_al_2011] Stegle, O.; Lippert, C.; Mooij, J.M.; Lawrence, N.D.; Borgwardt, K.:Efficient inference in matrix-variate Gaussian models with \\iid observation noise. In: Advances in Neural Information Processing Systems, 2011, Pages 630-638
 
     """
     def __init__(self, X1, X2, Y, kern1, kern2, noise_var=1., name='KGPR'):
diff --git a/GPy/models/tp_regression.py b/GPy/models/tp_regression.py
index 56ecf4b5..048f6316 100644
--- a/GPy/models/tp_regression.py
+++ b/GPy/models/tp_regression.py
@@ -171,7 +171,7 @@ class TPRegression(Model):
 
     def log_likelihood(self):
         """
-        The log marginal likelihood of the model, :math:`p(\mathbf{y})`, this is the objective function of the model being optimised
+        The log marginal likelihood of the model, :math:`p(\\mathbf{y})`, this is the objective function of the model being optimised
         """
         return self._log_marginal_likelihood or self.inference()[1]
 
@@ -184,8 +184,8 @@ class TPRegression(Model):
         diagonal of the covariance is returned.
 
         .. math::
-            p(f*|X*, X, Y) = \int^{\inf}_{\inf} p(f*|f,X*)p(f|X,Y) df
-                        = MVN\left(\nu + N,f*| K_{x*x}(K_{xx})^{-1}Y,
+            p(f*|X*, X, Y) = \\int^{\\inf}_{\\inf} p(f*|f,X*)p(f|X,Y) df
+                        = MVN\\left(\nu + N,f*| K_{x*x}(K_{xx})^{-1}Y,
                         \frac{\nu + \beta - 2}{\nu + N - 2}K_{x*x*} - K_{xx*}(K_{xx})^{-1}K_{xx*}\right)
             \nu := \texttt{Degrees of freedom}
         """
diff --git a/GPy/models/warped_gp.py b/GPy/models/warped_gp.py
index a24401fe..0c0aaaac 100644
--- a/GPy/models/warped_gp.py
+++ b/GPy/models/warped_gp.py
@@ -146,7 +146,7 @@ class WarpedGP(GP):
         the jacobian of the warping function here.
 
         .. math:
-            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\mu_{*}\\sigma^{2}_{*})
+            p(y_{*}|D) = p(y_{*}|f_{*})p(f_{*}|\\mu_{*}\\sigma^{2}_{*})
 
         :param x_test: test locations (x_{*})
         :type x_test: (Nx1) array
diff --git a/GPy/testing/state_space_main_tests.py b/GPy/testing/state_space_main_tests.py
index 898eaa90..29b30ff3 100644
--- a/GPy/testing/state_space_main_tests.py
+++ b/GPy/testing/state_space_main_tests.py
@@ -159,7 +159,7 @@ def generate_brownian_data(
 ):
     """
     Generate brownian data - data from Brownian motion.
-    First point is always 0, and \Beta(0) = 0  - standard conditions for Brownian motion.
+    First point is always 0, and \\Beta(0) = 0  - standard conditions for Brownian motion.
 
     Input:
     --------------------------------
diff --git a/GPy/testing/test_model.py b/GPy/testing/test_model.py
index 954bffe9..aba6b2e8 100644
--- a/GPy/testing/test_model.py
+++ b/GPy/testing/test_model.py
@@ -684,7 +684,7 @@ class TestMisc:
         warp_m = GPy.models.WarpedGP(
             X, Y
         )  # , kernel=warp_k)#, warping_function=warp_f)
-        warp_m[".*\.d"].constrain_fixed(1.0)
+        warp_m[r".*\.d"].constrain_fixed(1.0)
         warp_m.optimize_restarts(
             parallel=False, robust=False, num_restarts=5, max_iters=max_iters
         )
diff --git a/GPy/util/datasets.py b/GPy/util/datasets.py
index c406ed7b..a8e2ea3e 100644
--- a/GPy/util/datasets.py
+++ b/GPy/util/datasets.py
@@ -537,7 +537,7 @@ http://nbviewer.ipython.org/github/sahuguet/notebooks/blob/master/GoogleTrends%2
         # In the notebook they did some data cleaning: remove Javascript header+footer, and translate new Date(....,..,..) into YYYY-MM-DD.
         header = """// Data table response\ngoogle.visualization.Query.setResponse("""
         data = data[len(header):-2]
-        data = re.sub('new Date\((\d+),(\d+),(\d+)\)', (lambda m: '"%s-%02d-%02d"' % (m.group(1).strip(), 1+int(m.group(2)), int(m.group(3)))), data)
+        data = re.sub(r'new Date\((\d+),(\d+),(\d+)\)', (lambda m: '"%s-%02d-%02d"' % (m.group(1).strip(), 1+int(m.group(2)), int(m.group(3)))), data)
         timeseries = json.loads(data)
         columns = [k['label'] for k in timeseries['table']['cols']]
         rows = map(lambda x: [k['v'] for k in x['c']], timeseries['table']['rows'])
@@ -782,7 +782,7 @@ def hapmap3(data_set='hapmap3'):
 
           /  1, iff SNPij==(B1,B1)
     Aij = |  0, iff SNPij==(B1,B2)
-          \ -1, iff SNPij==(B2,B2)
+          \\ -1, iff SNPij==(B2,B2)
 
     The SNP data and the meta information (such as iid, sex and phenotype) are
     stored in the dataframe datadf, index is the Individual ID,
@@ -1011,7 +1011,7 @@ def singlecell_rna_seq_deng(dataset='singlecell_deng'):
     sample_info.columns = c
 
     # get the labels right:
-    rep = re.compile('\(.*\)')
+    rep = re.compile(r'\(.*\)')
     def filter_dev_stage(row):
         if isnull(row):
             row = "2-cell stage embryo"
@@ -1050,7 +1050,7 @@ def singlecell_rna_seq_deng(dataset='singlecell_deng'):
                 #gene_info[file_info.name[:-18]] = inner.Refseq_IDs
 
     # Strip GSM number off data index
-    rep = re.compile('GSM\d+_')
+    rep = re.compile(r'GSM\d+_')
 
     from pandas import MultiIndex
     columns = MultiIndex.from_tuples([row.split('_', 1) for row in data.columns])
diff --git a/GPy/util/netpbmfile.py b/GPy/util/netpbmfile.py
index 030bd574..dc09c64b 100644
--- a/GPy/util/netpbmfile.py
+++ b/GPy/util/netpbmfile.py
@@ -180,24 +180,24 @@ class NetpbmFile(object):
         """Read PAM header and initialize instance."""
         regroups = re.search(
             b"(^P7[\n\r]+(?:(?:[\n\r]+)|(?:#.*)|"
-            b"(HEIGHT\s+\d+)|(WIDTH\s+\d+)|(DEPTH\s+\d+)|(MAXVAL\s+\d+)|"
-            b"(?:TUPLTYPE\s+\w+))*ENDHDR\n)", data).groups()
+            rb"(HEIGHT\s+\d+)|(WIDTH\s+\d+)|(DEPTH\s+\d+)|(MAXVAL\s+\d+)|"
+            rb"(?:TUPLTYPE\s+\w+))*ENDHDR\n)", data).groups()
         self.header = regroups[0]
         self.magicnum = b'P7'
         for group in regroups[1:]:
             key, value = group.split()
             setattr(self, unicode(key).lower(), int(value))
-        matches = re.findall(b"(TUPLTYPE\s+\w+)", self.header)
+        matches = re.findall(rb"(TUPLTYPE\s+\w+)", self.header)
         self.tupltypes = [s.split(None, 1)[1] for s in matches]
 
     def _read_pnm_header(self, data):
         """Read PNM header and initialize instance."""
         bpm = data[1:2] in b"14"
         regroups = re.search(b"".join((
-            b"(^(P[123456]|P7 332)\s+(?:#.*[\r\n])*",
-            b"\s*(\d+)\s+(?:#.*[\r\n])*",
-            b"\s*(\d+)\s+(?:#.*[\r\n])*" * (not bpm),
-            b"\s*(\d+)\s(?:\s*#.*[\r\n]\s)*)")), data).groups() + (1, ) * bpm
+            rb"(^(P[123456]|P7 332)\s+(?:#.*[\r\n])*",
+            rb"\s*(\d+)\s+(?:#.*[\r\n])*",
+            rb"\s*(\d+)\s+(?:#.*[\r\n])*" * (not bpm),
+            rb"\s*(\d+)\s(?:\s*#.*[\r\n]\s)*)")), data).groups() + (1, ) * bpm
         self.header = regroups[0]
         self.magicnum = regroups[1]
         self.width = int(regroups[2])
diff --git a/doc/source/conf.py b/doc/source/conf.py
index 5fe9be77..a6a7c9cf 100644
--- a/doc/source/conf.py
+++ b/doc/source/conf.py
@@ -150,7 +150,7 @@ with open('../../GPy/__version__.py', 'r') as f:
     version = f.read()
     release = version
 
-print version
+print(version)
 
 # version = '0.8.8'
 # The full version, including alpha/beta/rc tags.