Week 43: Deep Learning: Constructing a Neural Network code and solving differential equations#
Morten Hjorth-Jensen, Department of Physics, University of Oslo, Norway
Date: October 20, 2025
Plans for week 43#
Material for the lecture on Monday October 20, 2025.
Reminder from last week, see also lecture notes from week 42 at https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/week42.html as well as those from week 41, see see https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/week41.html.
Building our own Feed-forward Neural Network.
Coding examples using Tensorflow/Keras and Pytorch examples. The Pytorch examples are adapted from Rashcka’s text, see chapters 11-13..
Start discussions on how to use neural networks for solving differential equations (ordinary and partial ones). This topic continues next week as well.
Video of lecture at https://youtu.be/Gi6mzxAT0Ew
Whiteboard notes at CompPhysics/MachineLearning
Exercises and lab session week 43#
Lab sessions on Tuesday and Wednesday.
Work on writing your own neural network code and discussions of project 2. If you didn’t get time to do the exercises from the two last weeks, we recommend doing so as these exercises give you the basic elements of a neural network code.
The exercises this week are tailored to the optional part of project 2, and deal with studying ways to display results from classification problems
Using Automatic differentiation#
In our discussions of ordinary differential equations and neural network codes we will also study the usage of Autograd, see for example https://www.youtube.com/watch?v=fRf4l5qaX1M&ab_channel=AlexSmola in computing gradients for deep learning. For the documentation of Autograd and examples see the Autograd documentation at HIPS/autograd and the lecture slides from week 41, see https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/week41.html.
Back propagation and automatic differentiation#
For more details on the back propagation algorithm and automatic differentiation see
Lecture Monday October 20#
Setting up the back propagation algorithm and algorithm for a feed forward NN, initalizations#
This is a reminder from last week.
The architecture (our model).
Set up your inputs and outputs (scalars, vectors, matrices or higher-order arrays)
Define the number of hidden layers and hidden nodes
Define activation functions for hidden layers and output layers
Define optimizer (plan learning rate, momentum, ADAgrad, RMSprop, ADAM etc) and array of initial learning rates
Define cost function and possible regularization terms with hyperparameters
Initialize weights and biases
Fix number of iterations for the feed forward part and back propagation part
Setting up the back propagation algorithm, part 1#
Let us write this out in the form of an algorithm.
First, we set up the input data \(\boldsymbol{x}\) and the activations \(\boldsymbol{z}_1\) of the input layer and compute the activation function and the pertinent outputs \(\boldsymbol{a}^1\).
Secondly, we perform then the feed forward till we reach the output layer and compute all \(\boldsymbol{z}_l\) of the input layer and compute the activation function and the pertinent outputs \(\boldsymbol{a}^l\) for \(l=1,2,3,\dots,L\).
Notation: The first hidden layer has \(l=1\) as label and the final output layer has \(l=L\).
Setting up the back propagation algorithm, part 2#
Thereafter we compute the ouput error \(\boldsymbol{\delta}^L\) by computing all
Then we compute the back propagate error for each \(l=L-1,L-2,\dots,1\) as
Setting up the Back propagation algorithm, part 3#
Finally, we update the weights and the biases using gradient descent for each \(l=L-1,L-2,\dots,1\) (the first hidden layer) and update the weights and biases according to the rules
with \(\eta\) being the learning rate.
Updating the gradients#
With the back propagate error for each \(l=L-1,L-2,\dots,1\) as
we update the weights and the biases using gradient descent for each \(l=L-1,L-2,\dots,1\) and update the weights and biases according to the rules
Activation functions#
A property that characterizes a neural network, other than its connectivity, is the choice of activation function(s). The following restrictions are imposed on an activation function for an FFNN to fulfill the universal approximation theorem
Non-constant
Bounded
Monotonically-increasing
Continuous
Activation functions, examples#
Typical examples are the logistic Sigmoid
and the hyperbolic tangent function
The RELU function family#
The ReLU activation function suffers from a problem known as the dying ReLUs: during training, some neurons effectively die, meaning they stop outputting anything other than 0.
In some cases, you may find that half of your network’s neurons are dead, especially if you used a large learning rate. During training, if a neuron’s weights get updated such that the weighted sum of the neuron’s inputs is negative, it will start outputting 0. When this happen, the neuron is unlikely to come back to life since the gradient of the ReLU function is 0 when its input is negative.
ELU function#
To solve this problem, nowadays practitioners use a variant of the ReLU function, such as the leaky ReLU discussed above or the so-called exponential linear unit (ELU) function
Which activation function should we use?#
In general it seems that the ELU activation function is better than the leaky ReLU function (and its variants), which is better than ReLU. ReLU performs better than \(\tanh\) which in turn performs better than the logistic function.
If runtime performance is an issue, then you may opt for the leaky ReLU function over the ELU function If you don’t want to tweak yet another hyperparameter, you may just use the default \(\alpha\) of \(0.01\) for the leaky ReLU, and \(1\) for ELU. If you have spare time and computing power, you can use cross-validation or bootstrap to evaluate other activation functions.
More on activation functions, output layers#
In most cases you can use the ReLU activation function in the hidden layers (or one of its variants).
It is a bit faster to compute than other activation functions, and the gradient descent optimization does in general not get stuck.
For the output layer:
For classification the softmax activation function is generally a good choice for classification tasks (when the classes are mutually exclusive).
For regression tasks, you can simply use no activation function at all.
Building neural networks in Tensorflow and Keras#
Now we want to build on the experience gained from our neural network implementation in NumPy and scikit-learn and use it to construct a neural network in Tensorflow. Once we have constructed a neural network in NumPy and Tensorflow, building one in Keras is really quite trivial, though the performance may suffer.
In our previous example we used only one hidden layer, and in this we will use two. From this it should be quite clear how to build one using an arbitrary number of hidden layers, using data structures such as Python lists or NumPy arrays.
Tensorflow#
Tensorflow is an open source library machine learning library developed by the Google Brain team for internal use. It was released under the Apache 2.0 open source license in November 9, 2015.
Tensorflow is a computational framework that allows you to construct machine learning models at different levels of abstraction, from high-level, object-oriented APIs like Keras, down to the C++ kernels that Tensorflow is built upon. The higher levels of abstraction are simpler to use, but less flexible, and our choice of implementation should reflect the problems we are trying to solve.
Tensorflow uses so-called graphs to represent your computation in terms of the dependencies between individual operations, such that you first build a Tensorflow graph to represent your model, and then create a Tensorflow session to run the graph.
In this guide we will analyze the same data as we did in our NumPy and scikit-learn tutorial, gathered from the MNIST database of images. We will give an introduction to the lower level Python Application Program Interfaces (APIs), and see how we use them to build our graph. Then we will build (effectively) the same graph in Keras, to see just how simple solving a machine learning problem can be.
To install tensorflow on Unix/Linux systems, use pip as
pip3 install tensorflow
and/or if you use anaconda, just write (or install from the graphical user interface) (current release of CPU-only TensorFlow)
conda create -n tf tensorflow
conda activate tf
To install the current release of GPU TensorFlow
conda create -n tf-gpu tensorflow-gpu
conda activate tf-gpu
Using Keras#
Keras is a high level neural network
that supports Tensorflow, CTNK and Theano as backends.
If you have Anaconda installed you may run the following command
conda install keras
You can look up the instructions here for more information.
We will to a large extent use keras in this course.
Collect and pre-process data#
Let us look again at the MINST data set.
%matplotlib inline
# import necessary packages
import numpy as np
import matplotlib.pyplot as plt
import tensorflow as tf
from sklearn import datasets
# ensure the same random numbers appear every time
np.random.seed(0)
# display images in notebook
%matplotlib inline
plt.rcParams['figure.figsize'] = (12,12)
# download MNIST dataset
digits = datasets.load_digits()
# define inputs and labels
inputs = digits.images
labels = digits.target
print("inputs = (n_inputs, pixel_width, pixel_height) = " + str(inputs.shape))
print("labels = (n_inputs) = " + str(labels.shape))
# flatten the image
# the value -1 means dimension is inferred from the remaining dimensions: 8x8 = 64
n_inputs = len(inputs)
inputs = inputs.reshape(n_inputs, -1)
print("X = (n_inputs, n_features) = " + str(inputs.shape))
# choose some random images to display
indices = np.arange(n_inputs)
random_indices = np.random.choice(indices, size=5)
for i, image in enumerate(digits.images[random_indices]):
plt.subplot(1, 5, i+1)
plt.axis('off')
plt.imshow(image, cmap=plt.cm.gray_r, interpolation='nearest')
plt.title("Label: %d" % digits.target[random_indices[i]])
plt.show()
from tensorflow.keras.layers import Input
from tensorflow.keras.models import Sequential #This allows appending layers to existing models
from tensorflow.keras.layers import Dense #This allows defining the characteristics of a particular layer
from tensorflow.keras import optimizers #This allows using whichever optimiser we want (sgd,adam,RMSprop)
from tensorflow.keras import regularizers #This allows using whichever regularizer we want (l1,l2,l1_l2)
from tensorflow.keras.utils import to_categorical #This allows using categorical cross entropy as the cost function
from sklearn.model_selection import train_test_split
# one-hot representation of labels
labels = to_categorical(labels)
# split into train and test data
train_size = 0.8
test_size = 1 - train_size
X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=train_size,
test_size=test_size)
epochs = 100
batch_size = 100
n_neurons_layer1 = 100
n_neurons_layer2 = 50
n_categories = 10
eta_vals = np.logspace(-5, 1, 7)
lmbd_vals = np.logspace(-5, 1, 7)
def create_neural_network_keras(n_neurons_layer1, n_neurons_layer2, n_categories, eta, lmbd):
model = Sequential()
model.add(Dense(n_neurons_layer1, activation='sigmoid', kernel_regularizer=regularizers.l2(lmbd)))
model.add(Dense(n_neurons_layer2, activation='sigmoid', kernel_regularizer=regularizers.l2(lmbd)))
model.add(Dense(n_categories, activation='softmax'))
sgd = optimizers.SGD(learning_rate=eta)
model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['accuracy'])
return model
DNN_keras = np.zeros((len(eta_vals), len(lmbd_vals)), dtype=object)
for i, eta in enumerate(eta_vals):
for j, lmbd in enumerate(lmbd_vals):
DNN = create_neural_network_keras(n_neurons_layer1, n_neurons_layer2, n_categories,
eta=eta, lmbd=lmbd)
DNN.fit(X_train, Y_train, epochs=epochs, batch_size=batch_size, verbose=0)
scores = DNN.evaluate(X_test, Y_test)
DNN_keras[i][j] = DNN
print("Learning rate = ", eta)
print("Lambda = ", lmbd)
print("Test accuracy: %.3f" % scores[1])
print()
# optional
# visual representation of grid search
# uses seaborn heatmap, could probably do this in matplotlib
import seaborn as sns
sns.set()
train_accuracy = np.zeros((len(eta_vals), len(lmbd_vals)))
test_accuracy = np.zeros((len(eta_vals), len(lmbd_vals)))
for i in range(len(eta_vals)):
for j in range(len(lmbd_vals)):
DNN = DNN_keras[i][j]
train_accuracy[i][j] = DNN.evaluate(X_train, Y_train)[1]
test_accuracy[i][j] = DNN.evaluate(X_test, Y_test)[1]
fig, ax = plt.subplots(figsize = (10, 10))
sns.heatmap(train_accuracy, annot=True, ax=ax, cmap="viridis")
ax.set_title("Training Accuracy")
ax.set_ylabel("$\eta$")
ax.set_xlabel("$\lambda$")
plt.show()
fig, ax = plt.subplots(figsize = (10, 10))
sns.heatmap(test_accuracy, annot=True, ax=ax, cmap="viridis")
ax.set_title("Test Accuracy")
ax.set_ylabel("$\eta$")
ax.set_xlabel("$\lambda$")
plt.show()
Using Pytorch with the full MNIST data set#
import torch
import torch.nn as nn
import torch.optim as optim
import torchvision
import torchvision.transforms as transforms
# Device configuration: use GPU if available
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# MNIST dataset (downloads if not already present)
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.5,), (0.5,)) # normalize to mean=0.5, std=0.5 (approx. [-1,1] pixel range)
])
train_dataset = torchvision.datasets.MNIST(root='./data', train=True, download=True, transform=transform)
test_dataset = torchvision.datasets.MNIST(root='./data', train=False, download=True, transform=transform)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=64, shuffle=True)
test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=64, shuffle=False)
class NeuralNet(nn.Module):
def __init__(self):
super(NeuralNet, self).__init__()
self.fc1 = nn.Linear(28*28, 100) # first hidden layer (784 -> 100)
self.fc2 = nn.Linear(100, 100) # second hidden layer (100 -> 100)
self.fc3 = nn.Linear(100, 10) # output layer (100 -> 10 classes)
def forward(self, x):
x = x.view(x.size(0), -1) # flatten images into vectors of size 784
x = torch.relu(self.fc1(x)) # hidden layer 1 + ReLU activation
x = torch.relu(self.fc2(x)) # hidden layer 2 + ReLU activation
x = self.fc3(x) # output layer (logits for 10 classes)
return x
model = NeuralNet().to(device)
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-4)
num_epochs = 10
for epoch in range(num_epochs):
model.train() # set model to training mode
running_loss = 0.0
for images, labels in train_loader:
# Move data to device (GPU if available, else CPU)
images, labels = images.to(device), labels.to(device)
optimizer.zero_grad() # reset gradients to zero
outputs = model(images) # forward pass: compute predictions
loss = criterion(outputs, labels) # compute cross-entropy loss
loss.backward() # backpropagate to compute gradients
optimizer.step() # update weights using SGD step
running_loss += loss.item()
# Compute average loss over all batches in this epoch
avg_loss = running_loss / len(train_loader)
print(f"Epoch {epoch+1}/{num_epochs}, Loss: {avg_loss:.4f}")
#Evaluation on the Test Set
model.eval() # set model to evaluation mode
correct = 0
total = 0
with torch.no_grad(): # disable gradient calculation for evaluation
for images, labels in test_loader:
images, labels = images.to(device), labels.to(device)
outputs = model(images)
_, predicted = torch.max(outputs, dim=1) # class with highest score
total += labels.size(0)
correct += (predicted == labels).sum().item()
accuracy = 100 * correct / total
print(f"Test Accuracy: {accuracy:.2f}%")
And a similar example using Tensorflow with Keras#
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers, regularizers
# Check for GPU (TensorFlow will use it automatically if available)
gpus = tf.config.list_physical_devices('GPU')
print(f"GPUs available: {gpus}")
# 1) Load and preprocess MNIST
(x_train, y_train), (x_test, y_test) = keras.datasets.mnist.load_data()
# Normalize to [0, 1]
x_train = (x_train.astype("float32") / 255.0)
x_test = (x_test.astype("float32") / 255.0)
# 2) Build the model: 784 -> 100 -> 100 -> 10
l2_reg = 1e-4 # L2 regularization strength
model = keras.Sequential([
layers.Input(shape=(28, 28)),
layers.Flatten(),
layers.Dense(100, activation="relu",
kernel_regularizer=regularizers.l2(l2_reg)),
layers.Dense(100, activation="relu",
kernel_regularizer=regularizers.l2(l2_reg)),
layers.Dense(10, activation="softmax") # output probabilities for 10 classes
])
# 3) Compile with SGD + weight decay via L2 regularizers
model.compile(
optimizer=keras.optimizers.SGD(learning_rate=0.01),
loss="sparse_categorical_crossentropy",
metrics=["accuracy"],
)
model.summary()
# 4) Train
history = model.fit(
x_train, y_train,
epochs=10,
batch_size=64,
validation_split=0.1, # optional: monitor validation during training
verbose=1
)
# 5) Evaluate on test set
test_loss, test_acc = model.evaluate(x_test, y_test, verbose=0)
print(f"Test accuracy: {test_acc:.4f}, Test loss: {test_loss:.4f}")
Building our own neural network code#
Here we present a flexible object oriented codebase for a feed forward neural network, along with a demonstration of how to use it. Before we get into the details of the neural network, we will first present some implementations of various schedulers, cost functions and activation functions that can be used together with the neural network.
The codes here were developed by Eric Reber and Gregor Kajda during spring 2023.
Learning rate methods#
The code below shows object oriented implementations of the Constant, Momentum, Adagrad, AdagradMomentum, RMS prop and Adam schedulers. All of the classes belong to the shared abstract Scheduler class, and share the update_change() and reset() methods allowing for any of the schedulers to be seamlessly used during the training stage, as will later be shown in the fit() method of the neural network. Update_change() only has one parameter, the gradient (\(δ^l_ja^{l−1}_k\)), and returns the change which will be subtracted from the weights. The reset() function takes no parameters, and resets the desired variables. For Constant and Momentum, reset does nothing.
import autograd.numpy as np
class Scheduler:
"""
Abstract class for Schedulers
"""
def __init__(self, eta):
self.eta = eta
# should be overwritten
def update_change(self, gradient):
raise NotImplementedError
# overwritten if needed
def reset(self):
pass
class Constant(Scheduler):
def __init__(self, eta):
super().__init__(eta)
def update_change(self, gradient):
return self.eta * gradient
def reset(self):
pass
class Momentum(Scheduler):
def __init__(self, eta: float, momentum: float):
super().__init__(eta)
self.momentum = momentum
self.change = 0
def update_change(self, gradient):
self.change = self.momentum * self.change + self.eta * gradient
return self.change
def reset(self):
pass
class Adagrad(Scheduler):
def __init__(self, eta):
super().__init__(eta)
self.G_t = None
def update_change(self, gradient):
delta = 1e-8 # avoid division ny zero
if self.G_t is None:
self.G_t = np.zeros((gradient.shape[0], gradient.shape[0]))
self.G_t += gradient @ gradient.T
G_t_inverse = 1 / (
delta + np.sqrt(np.reshape(np.diagonal(self.G_t), (self.G_t.shape[0], 1)))
)
return self.eta * gradient * G_t_inverse
def reset(self):
self.G_t = None
class AdagradMomentum(Scheduler):
def __init__(self, eta, momentum):
super().__init__(eta)
self.G_t = None
self.momentum = momentum
self.change = 0
def update_change(self, gradient):
delta = 1e-8 # avoid division ny zero
if self.G_t is None:
self.G_t = np.zeros((gradient.shape[0], gradient.shape[0]))
self.G_t += gradient @ gradient.T
G_t_inverse = 1 / (
delta + np.sqrt(np.reshape(np.diagonal(self.G_t), (self.G_t.shape[0], 1)))
)
self.change = self.change * self.momentum + self.eta * gradient * G_t_inverse
return self.change
def reset(self):
self.G_t = None
class RMS_prop(Scheduler):
def __init__(self, eta, rho):
super().__init__(eta)
self.rho = rho
self.second = 0.0
def update_change(self, gradient):
delta = 1e-8 # avoid division ny zero
self.second = self.rho * self.second + (1 - self.rho) * gradient * gradient
return self.eta * gradient / (np.sqrt(self.second + delta))
def reset(self):
self.second = 0.0
class Adam(Scheduler):
def __init__(self, eta, rho, rho2):
super().__init__(eta)
self.rho = rho
self.rho2 = rho2
self.moment = 0
self.second = 0
self.n_epochs = 1
def update_change(self, gradient):
delta = 1e-8 # avoid division ny zero
self.moment = self.rho * self.moment + (1 - self.rho) * gradient
self.second = self.rho2 * self.second + (1 - self.rho2) * gradient * gradient
moment_corrected = self.moment / (1 - self.rho**self.n_epochs)
second_corrected = self.second / (1 - self.rho2**self.n_epochs)
return self.eta * moment_corrected / (np.sqrt(second_corrected + delta))
def reset(self):
self.n_epochs += 1
self.moment = 0
self.second = 0
Usage of the above learning rate schedulers#
To initalize a scheduler, simply create the object and pass in the necessary parameters such as the learning rate and the momentum as shown below. As the Scheduler class is an abstract class it should not called directly, and will raise an error upon usage.
momentum_scheduler = Momentum(eta=1e-3, momentum=0.9)
adam_scheduler = Adam(eta=1e-3, rho=0.9, rho2=0.999)
Here is a small example for how a segment of code using schedulers could look. Switching out the schedulers is simple.
weights = np.ones((3,3))
print(f"Before scheduler:\n{weights=}")
epochs = 10
for e in range(epochs):
gradient = np.random.rand(3, 3)
change = adam_scheduler.update_change(gradient)
weights = weights - change
adam_scheduler.reset()
print(f"\nAfter scheduler:\n{weights=}")
Cost functions#
Here we discuss cost functions that can be used when creating the neural network. Every cost function takes the target vector as its parameter, and returns a function valued only at \(x\) such that it may easily be differentiated.
import autograd.numpy as np
def CostOLS(target):
def func(X):
return (1.0 / target.shape[0]) * np.sum((target - X) ** 2)
return func
def CostLogReg(target):
def func(X):
return -(1.0 / target.shape[0]) * np.sum(
(target * np.log(X + 10e-10)) + ((1 - target) * np.log(1 - X + 10e-10))
)
return func
def CostCrossEntropy(target):
def func(X):
return -(1.0 / target.size) * np.sum(target * np.log(X + 10e-10))
return func
Below we give a short example of how these cost function may be used to obtain results if you wish to test them out on your own using AutoGrad’s automatics differentiation.
from autograd import grad
target = np.array([[1, 2, 3]]).T
a = np.array([[4, 5, 6]]).T
cost_func = CostCrossEntropy
cost_func_derivative = grad(cost_func(target))
valued_at_a = cost_func_derivative(a)
print(f"Derivative of cost function {cost_func.__name__} valued at a:\n{valued_at_a}")
Activation functions#
Finally, before we look at the neural network, we will look at the activation functions which can be specified between the hidden layers and as the output function. Each function can be valued for any given vector or matrix X, and can be differentiated via derivate().
import autograd.numpy as np
from autograd import elementwise_grad
def identity(X):
return X
def sigmoid(X):
try:
return 1.0 / (1 + np.exp(-X))
except FloatingPointError:
return np.where(X > np.zeros(X.shape), np.ones(X.shape), np.zeros(X.shape))
def softmax(X):
X = X - np.max(X, axis=-1, keepdims=True)
delta = 10e-10
return np.exp(X) / (np.sum(np.exp(X), axis=-1, keepdims=True) + delta)
def RELU(X):
return np.where(X > np.zeros(X.shape), X, np.zeros(X.shape))
def LRELU(X):
delta = 10e-4
return np.where(X > np.zeros(X.shape), X, delta * X)
def derivate(func):
if func.__name__ == "RELU":
def func(X):
return np.where(X > 0, 1, 0)
return func
elif func.__name__ == "LRELU":
def func(X):
delta = 10e-4
return np.where(X > 0, 1, delta)
return func
else:
return elementwise_grad(func)
Below follows a short demonstration of how to use an activation function. The derivative of the activation function will be important when calculating the output delta term during backpropagation. Note that derivate() can also be used for cost functions for a more generalized approach.
z = np.array([[4, 5, 6]]).T
print(f"Input to activation function:\n{z}")
act_func = sigmoid
a = act_func(z)
print(f"\nOutput from {act_func.__name__} activation function:\n{a}")
act_func_derivative = derivate(act_func)
valued_at_z = act_func_derivative(a)
print(f"\nDerivative of {act_func.__name__} activation function valued at z:\n{valued_at_z}")
The Neural Network#
Now that we have gotten a good understanding of the implementation of some important components, we can take a look at an object oriented implementation of a feed forward neural network. The feed forward neural network has been implemented as a class named FFNN, which can be initiated as a regressor or classifier dependant on the choice of cost function. The FFNN can have any number of input nodes, hidden layers with any amount of hidden nodes, and any amount of output nodes meaning it can perform multiclass classification as well as binary classification and regression problems. Although there is a lot of code present, it makes for an easy to use and generalizeable interface for creating many types of neural networks as will be demonstrated below.
import math
import autograd.numpy as np
import sys
import warnings
from autograd import grad, elementwise_grad
from random import random, seed
from copy import deepcopy, copy
from typing import Tuple, Callable
from sklearn.utils import resample
warnings.simplefilter("error")
class FFNN:
"""
Description:
------------
Feed Forward Neural Network with interface enabling flexible design of a
nerual networks architecture and the specification of activation function
in the hidden layers and output layer respectively. This model can be used
for both regression and classification problems, depending on the output function.
Attributes:
------------
I dimensions (tuple[int]): A list of positive integers, which specifies the
number of nodes in each of the networks layers. The first integer in the array
defines the number of nodes in the input layer, the second integer defines number
of nodes in the first hidden layer and so on until the last number, which
specifies the number of nodes in the output layer.
II hidden_func (Callable): The activation function for the hidden layers
III output_func (Callable): The activation function for the output layer
IV cost_func (Callable): Our cost function
V seed (int): Sets random seed, makes results reproducible
"""
def __init__(
self,
dimensions: tuple[int],
hidden_func: Callable = sigmoid,
output_func: Callable = lambda x: x,
cost_func: Callable = CostOLS,
seed: int = None,
):
self.dimensions = dimensions
self.hidden_func = hidden_func
self.output_func = output_func
self.cost_func = cost_func
self.seed = seed
self.weights = list()
self.schedulers_weight = list()
self.schedulers_bias = list()
self.a_matrices = list()
self.z_matrices = list()
self.classification = None
self.reset_weights()
self._set_classification()
def fit(
self,
X: np.ndarray,
t: np.ndarray,
scheduler: Scheduler,
batches: int = 1,
epochs: int = 100,
lam: float = 0,
X_val: np.ndarray = None,
t_val: np.ndarray = None,
):
"""
Description:
------------
This function performs the training the neural network by performing the feedforward and backpropagation
algorithm to update the networks weights.
Parameters:
------------
I X (np.ndarray) : training data
II t (np.ndarray) : target data
III scheduler (Scheduler) : specified scheduler (algorithm for optimization of gradient descent)
IV scheduler_args (list[int]) : list of all arguments necessary for scheduler
Optional Parameters:
------------
V batches (int) : number of batches the datasets are split into, default equal to 1
VI epochs (int) : number of iterations used to train the network, default equal to 100
VII lam (float) : regularization hyperparameter lambda
VIII X_val (np.ndarray) : validation set
IX t_val (np.ndarray) : validation target set
Returns:
------------
I scores (dict) : A dictionary containing the performance metrics of the model.
The number of the metrics depends on the parameters passed to the fit-function.
"""
# setup
if self.seed is not None:
np.random.seed(self.seed)
val_set = False
if X_val is not None and t_val is not None:
val_set = True
# creating arrays for score metrics
train_errors = np.empty(epochs)
train_errors.fill(np.nan)
val_errors = np.empty(epochs)
val_errors.fill(np.nan)
train_accs = np.empty(epochs)
train_accs.fill(np.nan)
val_accs = np.empty(epochs)
val_accs.fill(np.nan)
self.schedulers_weight = list()
self.schedulers_bias = list()
batch_size = X.shape[0] // batches
X, t = resample(X, t)
# this function returns a function valued only at X
cost_function_train = self.cost_func(t)
if val_set:
cost_function_val = self.cost_func(t_val)
# create schedulers for each weight matrix
for i in range(len(self.weights)):
self.schedulers_weight.append(copy(scheduler))
self.schedulers_bias.append(copy(scheduler))
print(f"{scheduler.__class__.__name__}: Eta={scheduler.eta}, Lambda={lam}")
try:
for e in range(epochs):
for i in range(batches):
# allows for minibatch gradient descent
if i == batches - 1:
# If the for loop has reached the last batch, take all thats left
X_batch = X[i * batch_size :, :]
t_batch = t[i * batch_size :, :]
else:
X_batch = X[i * batch_size : (i + 1) * batch_size, :]
t_batch = t[i * batch_size : (i + 1) * batch_size, :]
self._feedforward(X_batch)
self._backpropagate(X_batch, t_batch, lam)
# reset schedulers for each epoch (some schedulers pass in this call)
for scheduler in self.schedulers_weight:
scheduler.reset()
for scheduler in self.schedulers_bias:
scheduler.reset()
# computing performance metrics
pred_train = self.predict(X)
train_error = cost_function_train(pred_train)
train_errors[e] = train_error
if val_set:
pred_val = self.predict(X_val)
val_error = cost_function_val(pred_val)
val_errors[e] = val_error
if self.classification:
train_acc = self._accuracy(self.predict(X), t)
train_accs[e] = train_acc
if val_set:
val_acc = self._accuracy(pred_val, t_val)
val_accs[e] = val_acc
# printing progress bar
progression = e / epochs
print_length = self._progress_bar(
progression,
train_error=train_errors[e],
train_acc=train_accs[e],
val_error=val_errors[e],
val_acc=val_accs[e],
)
except KeyboardInterrupt:
# allows for stopping training at any point and seeing the result
pass
# visualization of training progression (similiar to tensorflow progression bar)
sys.stdout.write("\r" + " " * print_length)
sys.stdout.flush()
self._progress_bar(
1,
train_error=train_errors[e],
train_acc=train_accs[e],
val_error=val_errors[e],
val_acc=val_accs[e],
)
sys.stdout.write("")
# return performance metrics for the entire run
scores = dict()
scores["train_errors"] = train_errors
if val_set:
scores["val_errors"] = val_errors
if self.classification:
scores["train_accs"] = train_accs
if val_set:
scores["val_accs"] = val_accs
return scores
def predict(self, X: np.ndarray, *, threshold=0.5):
"""
Description:
------------
Performs prediction after training of the network has been finished.
Parameters:
------------
I X (np.ndarray): The design matrix, with n rows of p features each
Optional Parameters:
------------
II threshold (float) : sets minimal value for a prediction to be predicted as the positive class
in classification problems
Returns:
------------
I z (np.ndarray): A prediction vector (row) for each row in our design matrix
This vector is thresholded if regression=False, meaning that classification results
in a vector of 1s and 0s, while regressions in an array of decimal numbers
"""
predict = self._feedforward(X)
if self.classification:
return np.where(predict > threshold, 1, 0)
else:
return predict
def reset_weights(self):
"""
Description:
------------
Resets/Reinitializes the weights in order to train the network for a new problem.
"""
if self.seed is not None:
np.random.seed(self.seed)
self.weights = list()
for i in range(len(self.dimensions) - 1):
weight_array = np.random.randn(
self.dimensions[i] + 1, self.dimensions[i + 1]
)
weight_array[0, :] = np.random.randn(self.dimensions[i + 1]) * 0.01
self.weights.append(weight_array)
def _feedforward(self, X: np.ndarray):
"""
Description:
------------
Calculates the activation of each layer starting at the input and ending at the output.
Each following activation is calculated from a weighted sum of each of the preceeding
activations (except in the case of the input layer).
Parameters:
------------
I X (np.ndarray): The design matrix, with n rows of p features each
Returns:
------------
I z (np.ndarray): A prediction vector (row) for each row in our design matrix
"""
# reset matrices
self.a_matrices = list()
self.z_matrices = list()
# if X is just a vector, make it into a matrix
if len(X.shape) == 1:
X = X.reshape((1, X.shape[0]))
# Add a coloumn of zeros as the first coloumn of the design matrix, in order
# to add bias to our data
bias = np.ones((X.shape[0], 1)) * 0.01
X = np.hstack([bias, X])
# a^0, the nodes in the input layer (one a^0 for each row in X - where the
# exponent indicates layer number).
a = X
self.a_matrices.append(a)
self.z_matrices.append(a)
# The feed forward algorithm
for i in range(len(self.weights)):
if i < len(self.weights) - 1:
z = a @ self.weights[i]
self.z_matrices.append(z)
a = self.hidden_func(z)
# bias column again added to the data here
bias = np.ones((a.shape[0], 1)) * 0.01
a = np.hstack([bias, a])
self.a_matrices.append(a)
else:
try:
# a^L, the nodes in our output layers
z = a @ self.weights[i]
a = self.output_func(z)
self.a_matrices.append(a)
self.z_matrices.append(z)
except Exception as OverflowError:
print(
"OverflowError in fit() in FFNN\nHOW TO DEBUG ERROR: Consider lowering your learning rate or scheduler specific parameters such as momentum, or check if your input values need scaling"
)
# this will be a^L
return a
def _backpropagate(self, X, t, lam):
"""
Description:
------------
Performs the backpropagation algorithm. In other words, this method
calculates the gradient of all the layers starting at the
output layer, and moving from right to left accumulates the gradient until
the input layer is reached. Each layers respective weights are updated while
the algorithm propagates backwards from the output layer (auto-differentation in reverse mode).
Parameters:
------------
I X (np.ndarray): The design matrix, with n rows of p features each.
II t (np.ndarray): The target vector, with n rows of p targets.
III lam (float32): regularization parameter used to punish the weights in case of overfitting
Returns:
------------
No return value.
"""
out_derivative = derivate(self.output_func)
hidden_derivative = derivate(self.hidden_func)
for i in range(len(self.weights) - 1, -1, -1):
# delta terms for output
if i == len(self.weights) - 1:
# for multi-class classification
if (
self.output_func.__name__ == "softmax"
):
delta_matrix = self.a_matrices[i + 1] - t
# for single class classification
else:
cost_func_derivative = grad(self.cost_func(t))
delta_matrix = out_derivative(
self.z_matrices[i + 1]
) * cost_func_derivative(self.a_matrices[i + 1])
# delta terms for hidden layer
else:
delta_matrix = (
self.weights[i + 1][1:, :] @ delta_matrix.T
).T * hidden_derivative(self.z_matrices[i + 1])
# calculate gradient
gradient_weights = self.a_matrices[i][:, 1:].T @ delta_matrix
gradient_bias = np.sum(delta_matrix, axis=0).reshape(
1, delta_matrix.shape[1]
)
# regularization term
gradient_weights += self.weights[i][1:, :] * lam
# use scheduler
update_matrix = np.vstack(
[
self.schedulers_bias[i].update_change(gradient_bias),
self.schedulers_weight[i].update_change(gradient_weights),
]
)
# update weights and bias
self.weights[i] -= update_matrix
def _accuracy(self, prediction: np.ndarray, target: np.ndarray):
"""
Description:
------------
Calculates accuracy of given prediction to target
Parameters:
------------
I prediction (np.ndarray): vector of predicitons output network
(1s and 0s in case of classification, and real numbers in case of regression)
II target (np.ndarray): vector of true values (What the network ideally should predict)
Returns:
------------
A floating point number representing the percentage of correctly classified instances.
"""
assert prediction.size == target.size
return np.average((target == prediction))
def _set_classification(self):
"""
Description:
------------
Decides if FFNN acts as classifier (True) og regressor (False),
sets self.classification during init()
"""
self.classification = False
if (
self.cost_func.__name__ == "CostLogReg"
or self.cost_func.__name__ == "CostCrossEntropy"
):
self.classification = True
def _progress_bar(self, progression, **kwargs):
"""
Description:
------------
Displays progress of training
"""
print_length = 40
num_equals = int(progression * print_length)
num_not = print_length - num_equals
arrow = ">" if num_equals > 0 else ""
bar = "[" + "=" * (num_equals - 1) + arrow + "-" * num_not + "]"
perc_print = self._format(progression * 100, decimals=5)
line = f" {bar} {perc_print}% "
for key in kwargs:
if not np.isnan(kwargs[key]):
value = self._format(kwargs[key], decimals=4)
line += f"| {key}: {value} "
sys.stdout.write("\r" + line)
sys.stdout.flush()
return len(line)
def _format(self, value, decimals=4):
"""
Description:
------------
Formats decimal numbers for progress bar
"""
if value > 0:
v = value
elif value < 0:
v = -10 * value
else:
v = 1
n = 1 + math.floor(math.log10(v))
if n >= decimals - 1:
return str(round(value))
return f"{value:.{decimals-n-1}f}"
Before we make a model, we will quickly generate a dataset we can use for our linear regression problem as shown below
import autograd.numpy as np
from sklearn.model_selection import train_test_split
def SkrankeFunction(x, y):
return np.ravel(0 + 1*x + 2*y + 3*x**2 + 4*x*y + 5*y**2)
def create_X(x, y, n):
if len(x.shape) > 1:
x = np.ravel(x)
y = np.ravel(y)
N = len(x)
l = int((n + 1) * (n + 2) / 2) # Number of elements in beta
X = np.ones((N, l))
for i in range(1, n + 1):
q = int((i) * (i + 1) / 2)
for k in range(i + 1):
X[:, q + k] = (x ** (i - k)) * (y**k)
return X
step=0.5
x = np.arange(0, 1, step)
y = np.arange(0, 1, step)
x, y = np.meshgrid(x, y)
target = SkrankeFunction(x, y)
target = target.reshape(target.shape[0], 1)
poly_degree=3
X = create_X(x, y, poly_degree)
X_train, X_test, t_train, t_test = train_test_split(X, target)
Now that we have our dataset ready for the regression, we can create our regressor. Note that with the seed parameter, we can make sure our results stay the same every time we run the neural network. For inititialization, we simply specify the dimensions (we wish the amount of input nodes to be equal to the datapoints, and the output to predict one value).
input_nodes = X_train.shape[1]
output_nodes = 1
linear_regression = FFNN((input_nodes, output_nodes), output_func=identity, cost_func=CostOLS, seed=2023)
We then fit our model with our training data using the scheduler of our choice.
linear_regression.reset_weights() # reset weights such that previous runs or reruns don't affect the weights
scheduler = Constant(eta=1e-3)
scores = linear_regression.fit(X_train, t_train, scheduler)
Due to the progress bar we can see the MSE (train_error) throughout the FFNN’s training. Note that the fit() function has some optional parameters with defualt arguments. For example, the regularization hyperparameter can be left ignored if not needed, and equally the FFNN will by default run for 100 epochs. These can easily be changed, such as for example:
linear_regression.reset_weights() # reset weights such that previous runs or reruns don't affect the weights
scores = linear_regression.fit(X_train, t_train, scheduler, lam=1e-4, epochs=1000)
We see that given more epochs to train on, the regressor reaches a lower MSE.
Let us then switch to a binary classification. We use a binary classification dataset, and follow a similar setup to the regression case.
from sklearn.datasets import load_breast_cancer
from sklearn.preprocessing import MinMaxScaler
wisconsin = load_breast_cancer()
X = wisconsin.data
target = wisconsin.target
target = target.reshape(target.shape[0], 1)
X_train, X_val, t_train, t_val = train_test_split(X, target)
scaler = MinMaxScaler()
scaler.fit(X_train)
X_train = scaler.transform(X_train)
X_val = scaler.transform(X_val)
input_nodes = X_train.shape[1]
output_nodes = 1
logistic_regression = FFNN((input_nodes, output_nodes), output_func=sigmoid, cost_func=CostLogReg, seed=2023)
We will now make use of our validation data by passing it into our fit function as a keyword argument
logistic_regression.reset_weights() # reset weights such that previous runs or reruns don't affect the weights
scheduler = Adam(eta=1e-3, rho=0.9, rho2=0.999)
scores = logistic_regression.fit(X_train, t_train, scheduler, epochs=1000, X_val=X_val, t_val=t_val)
Finally, we will create a neural network with 2 hidden layers with activation functions.
input_nodes = X_train.shape[1]
hidden_nodes1 = 100
hidden_nodes2 = 30
output_nodes = 1
dims = (input_nodes, hidden_nodes1, hidden_nodes2, output_nodes)
neural_network = FFNN(dims, hidden_func=RELU, output_func=sigmoid, cost_func=CostLogReg, seed=2023)
neural_network.reset_weights() # reset weights such that previous runs or reruns don't affect the weights
scheduler = Adam(eta=1e-4, rho=0.9, rho2=0.999)
scores = neural_network.fit(X_train, t_train, scheduler, epochs=1000, X_val=X_val, t_val=t_val)
Multiclass classification#
Finally, we will demonstrate the use case of multiclass classification using our FFNN with the famous MNIST dataset, which contain images of digits between the range of 0 to 9.
from sklearn.datasets import load_digits
def onehot(target: np.ndarray):
onehot = np.zeros((target.size, target.max() + 1))
onehot[np.arange(target.size), target] = 1
return onehot
digits = load_digits()
X = digits.data
target = digits.target
target = onehot(target)
input_nodes = 64
hidden_nodes1 = 100
hidden_nodes2 = 30
output_nodes = 10
dims = (input_nodes, hidden_nodes1, hidden_nodes2, output_nodes)
multiclass = FFNN(dims, hidden_func=LRELU, output_func=softmax, cost_func=CostCrossEntropy)
multiclass.reset_weights() # reset weights such that previous runs or reruns don't affect the weights
scheduler = Adam(eta=1e-4, rho=0.9, rho2=0.999)
scores = multiclass.fit(X, target, scheduler, epochs=1000)
Testing the XOR gate and other gates#
Let us now use our code to test the XOR gate.
X = np.array([ [0, 0], [0, 1], [1, 0],[1, 1]],dtype=np.float64)
# The XOR gate
yXOR = np.array( [[ 0], [1] ,[1], [0]])
input_nodes = X.shape[1]
output_nodes = 1
logistic_regression = FFNN((input_nodes, output_nodes), output_func=sigmoid, cost_func=CostLogReg, seed=2023)
logistic_regression.reset_weights() # reset weights such that previous runs or reruns don't affect the weights
scheduler = Adam(eta=1e-1, rho=0.9, rho2=0.999)
scores = logistic_regression.fit(X, yXOR, scheduler, epochs=1000)
Not bad, but the results depend strongly on the learning reate. Try different learning rates.
Solving differential equations with Deep Learning#
The Universal Approximation Theorem states that a neural network can approximate any function at a single hidden layer along with one input and output layer to any given precision.
Book on solving differential equations with ML methods.
An Introduction to Neural Network Methods for Differential Equations, by Yadav and Kumar.
Physics informed neural networks.
Scientific Machine Learning Through Physics–Informed Neural Networks: Where we are and What’s Next, by Cuomo et al
Thanks to Kristine Baluka Hein.
The lectures on differential equations were developed by Kristine Baluka Hein, now PhD student at IFI. A great thanks to Kristine.
Ordinary Differential Equations first#
An ordinary differential equation (ODE) is an equation involving functions having one variable.
In general, an ordinary differential equation looks like
where \(g(x)\) is the function to find, and \(g^{(n)}(x)\) is the \(n\)-th derivative of \(g(x)\).
The \(f\left(x, g(x), g'(x), g''(x), \, \dots \, , g^{(n)}(x)\right)\) is just a way to write that there is an expression involving \(x\) and \(g(x), \ g'(x), \ g''(x), \, \dots \, , \text{ and } g^{(n)}(x)\) on the left side of the equality sign in (1). The highest order of derivative, that is the value of \(n\), determines to the order of the equation. The equation is referred to as a \(n\)-th order ODE. Along with (1), some additional conditions of the function \(g(x)\) are typically given for the solution to be unique.
The trial solution#
Let the trial solution \(g_t(x)\) be
where \(h_1(x)\) is a function that makes \(g_t(x)\) satisfy a given set of conditions, \(N(x,P)\) a neural network with weights and biases described by \(P\) and \(h_2(x, N(x,P))\) some expression involving the neural network. The role of the function \(h_2(x, N(x,P))\), is to ensure that the output from \(N(x,P)\) is zero when \(g_t(x)\) is evaluated at the values of \(x\) where the given conditions must be satisfied. The function \(h_1(x)\) should alone make \(g_t(x)\) satisfy the conditions.
But what about the network \(N(x,P)\)?
As described previously, an optimization method could be used to minimize the parameters of a neural network, that being its weights and biases, through backward propagation.
Minimization process#
For the minimization to be defined, we need to have a cost function at hand to minimize.
It is given that \(f\left(x, \, g(x), \, g'(x), \, g''(x), \, \dots \, , \, g^{(n)}(x)\right)\) should be equal to zero in (1). We can choose to consider the mean squared error as the cost function for an input \(x\). Since we are looking at one input, the cost function is just \(f\) squared. The cost function \(c\left(x, P \right)\) can therefore be expressed as
If \(N\) inputs are given as a vector \(\boldsymbol{x}\) with elements \(x_i\) for \(i = 1,\dots,N\), the cost function becomes
The neural net should then find the parameters \(P\) that minimizes the cost function in (3) for a set of \(N\) training samples \(x_i\).
Minimizing the cost function using gradient descent and automatic differentiation#
To perform the minimization using gradient descent, the gradient of \(C\left(\boldsymbol{x}, P\right)\) is needed. It might happen so that finding an analytical expression of the gradient of \(C(\boldsymbol{x}, P)\) from (3) gets too messy, depending on which cost function one desires to use.
Luckily, there exists libraries that makes the job for us through automatic differentiation. Automatic differentiation is a method of finding the derivatives numerically with very high precision.
Example: Exponential decay#
An exponential decay of a quantity \(g(x)\) is described by the equation
with \(g(0) = g_0\) for some chosen initial value \(g_0\).
The analytical solution of (4) is
Having an analytical solution at hand, it is possible to use it to compare how well a neural network finds a solution of (4).
The function to solve for#
The program will use a neural network to solve
where \(g(0) = g_0\) with \(\gamma\) and \(g_0\) being some chosen values.
In this example, \(\gamma = 2\) and \(g_0 = 10\).
The trial solution#
To begin with, a trial solution \(g_t(t)\) must be chosen. A general trial solution for ordinary differential equations could be
with \(h_1(x)\) ensuring that \(g_t(x)\) satisfies some conditions and \(h_2(x,N(x, P))\) an expression involving \(x\) and the output from the neural network \(N(x,P)\) with \(P \) being the collection of the weights and biases for each layer. For now, it is assumed that the network consists of one input layer, one hidden layer, and one output layer.
Setup of Network#
In this network, there are no weights and bias at the input layer, so \(P = \{ P_{\text{hidden}}, P_{\text{output}} \}\). If there are \(N_{\text{hidden} }\) neurons in the hidden layer, then \(P_{\text{hidden}}\) is a \(N_{\text{hidden} } \times (1 + N_{\text{input}})\) matrix, given that there are \(N_{\text{input}}\) neurons in the input layer.
The first column in \(P_{\text{hidden} }\) represents the bias for each neuron in the hidden layer and the second column represents the weights for each neuron in the hidden layer from the input layer. If there are \(N_{\text{output} }\) neurons in the output layer, then \(P_{\text{output}} \) is a \(N_{\text{output} } \times (1 + N_{\text{hidden} })\) matrix.
Its first column represents the bias of each neuron and the remaining columns represents the weights to each neuron.
It is given that \(g(0) = g_0\). The trial solution must fulfill this condition to be a proper solution of (6). A possible way to ensure that \(g_t(0, P) = g_0\), is to let \(F(N(x,P)) = x \cdot N(x,P)\) and \(A(x) = g_0\). This gives the following trial solution:
Reformulating the problem#
We wish that our neural network manages to minimize a given cost function.
A reformulation of out equation, (6), must therefore be done, such that it describes the problem a neural network can solve for.
The neural network must find the set of weights and biases \(P\) such that the trial solution in (7) satisfies (6).
The trial solution
has been chosen such that it already solves the condition \(g(0) = g_0\). What remains, is to find \(P\) such that
is fulfilled as best as possible.
More technicalities#
The left hand side and right hand side of (8) must be computed separately, and then the neural network must choose weights and biases, contained in \(P\), such that the sides are equal as best as possible. This means that the absolute or squared difference between the sides must be as close to zero, ideally equal to zero. In this case, the difference squared shows to be an appropriate measurement of how erroneous the trial solution is with respect to \(P\) of the neural network.
This gives the following cost function our neural network must solve for:
(the notation \(\min_{P}\{ f(x, P) \}\) means that we desire to find \(P\) that yields the minimum of \(f(x, P)\))
or, in terms of weights and biases for the hidden and output layer in our network:
for an input value \(x\).
More details#
If the neural network evaluates \(g_t(x, P)\) at more values for \(x\), say \(N\) values \(x_i\) for \(i = 1, \dots, N\), then the total error to minimize becomes
Letting \(\boldsymbol{x}\) be a vector with elements \(x_i\) and \(C(\boldsymbol{x}, P) = \frac{1}{N} \sum_i \big(g_t'(x_i, P) - ( -\gamma g_t(x_i, P) \big)^2\) denote the cost function, the minimization problem that our network must solve, becomes
In terms of \(P_{\text{hidden} }\) and \(P_{\text{output} }\), this could also be expressed as
A possible implementation of a neural network#
For simplicity, it is assumed that the input is an array \(\boldsymbol{x} = (x_1, \dots, x_N)\) with \(N\) elements. It is at these points the neural network should find \(P\) such that it fulfills (9).
First, the neural network must feed forward the inputs. This means that \(\boldsymbol{x}s\) must be passed through an input layer, a hidden layer and a output layer. The input layer in this case, does not need to process the data any further. The input layer will consist of \(N_{\text{input} }\) neurons, passing its element to each neuron in the hidden layer. The number of neurons in the hidden layer will be \(N_{\text{hidden} }\).
Technicalities#
For the \(i\)-th in the hidden layer with weight \(w_i^{\text{hidden} }\) and bias \(b_i^{\text{hidden} }\), the weighting from the \(j\)-th neuron at the input layer is:
Final technicalities I#
The result after weighting the inputs at the \(i\)-th hidden neuron can be written as a vector:
Final technicalities II#
The vector \(\boldsymbol{p}_{i, \text{hidden}}^T\) constitutes each row in \(P_{\text{hidden} }\), which contains the weights for the neural network to minimize according to (9).
After having found \(\boldsymbol{z}_{i}^{\text{hidden}} \) for every \(i\)-th neuron within the hidden layer, the vector will be sent to an activation function \(a_i(\boldsymbol{z})\).
In this example, the sigmoid function has been chosen to be the activation function for each hidden neuron:
It is possible to use other activations functions for the hidden layer also.
The output \(\boldsymbol{x}_i^{\text{hidden}}\) from each \(i\)-th hidden neuron is:
The outputs \(\boldsymbol{x}_i^{\text{hidden} } \) are then sent to the output layer.
The output layer consists of one neuron in this case, and combines the output from each of the neurons in the hidden layers. The output layer combines the results from the hidden layer using some weights \(w_i^{\text{output}}\) and biases \(b_i^{\text{output}}\). In this case, it is assumes that the number of neurons in the output layer is one.
Final technicalities III#
The procedure of weighting the output neuron \(j\) in the hidden layer to the \(i\)-th neuron in the output layer is similar as for the hidden layer described previously.
Final technicalities IV#
Expressing \(z_{1,j}^{\text{output}}\) as a vector gives the following way of weighting the inputs from the hidden layer:
In this case we seek a continuous range of values since we are approximating a function. This means that after computing \(\boldsymbol{z}_{1}^{\text{output}}\) the neural network has finished its feed forward step, and \(\boldsymbol{z}_{1}^{\text{output}}\) is the final output of the network.
Back propagation#
The next step is to decide how the parameters should be changed such that they minimize the cost function.
The chosen cost function for this problem is
In order to minimize the cost function, an optimization method must be chosen.
Here, gradient descent with a constant step size has been chosen.
Gradient descent#
The idea of the gradient descent algorithm is to update parameters in a direction where the cost function decreases goes to a minimum.
In general, the update of some parameters \(\boldsymbol{\omega}\) given a cost function defined by some weights \(\boldsymbol{\omega}\), \(C(\boldsymbol{x}, \boldsymbol{\omega})\), goes as follows:
for a number of iterations or until \( \big|\big| \boldsymbol{\omega}_{\text{new} } - \boldsymbol{\omega} \big|\big|\) becomes smaller than some given tolerance.
The value of \(\lambda\) decides how large steps the algorithm must take in the direction of \( \nabla_{\boldsymbol{\omega}} C(\boldsymbol{x}, \boldsymbol{\omega})\). The notation \(\nabla_{\boldsymbol{\omega}}\) express the gradient with respect to the elements in \(\boldsymbol{\omega}\).
In our case, we have to minimize the cost function \(C(\boldsymbol{x}, P)\) with respect to the two sets of weights and biases, that is for the hidden layer \(P_{\text{hidden} }\) and for the output layer \(P_{\text{output} }\) .
This means that \(P_{\text{hidden} }\) and \(P_{\text{output} }\) is updated by
The code for solving the ODE#
import autograd.numpy as np
from autograd import grad, elementwise_grad
import autograd.numpy.random as npr
from matplotlib import pyplot as plt
def sigmoid(z):
return 1/(1 + np.exp(-z))
# Assuming one input, hidden, and output layer
def neural_network(params, x):
# Find the weights (including and biases) for the hidden and output layer.
# Assume that params is a list of parameters for each layer.
# The biases are the first element for each array in params,
# and the weights are the remaning elements in each array in params.
w_hidden = params[0]
w_output = params[1]
# Assumes input x being an one-dimensional array
num_values = np.size(x)
x = x.reshape(-1, num_values)
# Assume that the input layer does nothing to the input x
x_input = x
## Hidden layer:
# Add a row of ones to include bias
x_input = np.concatenate((np.ones((1,num_values)), x_input ), axis = 0)
z_hidden = np.matmul(w_hidden, x_input)
x_hidden = sigmoid(z_hidden)
## Output layer:
# Include bias:
x_hidden = np.concatenate((np.ones((1,num_values)), x_hidden ), axis = 0)
z_output = np.matmul(w_output, x_hidden)
x_output = z_output
return x_output
# The trial solution using the deep neural network:
def g_trial(x,params, g0 = 10):
return g0 + x*neural_network(params,x)
# The right side of the ODE:
def g(x, g_trial, gamma = 2):
return -gamma*g_trial
# The cost function:
def cost_function(P, x):
# Evaluate the trial function with the current parameters P
g_t = g_trial(x,P)
# Find the derivative w.r.t x of the neural network
d_net_out = elementwise_grad(neural_network,1)(P,x)
# Find the derivative w.r.t x of the trial function
d_g_t = elementwise_grad(g_trial,0)(x,P)
# The right side of the ODE
func = g(x, g_t)
err_sqr = (d_g_t - func)**2
cost_sum = np.sum(err_sqr)
return cost_sum / np.size(err_sqr)
# Solve the exponential decay ODE using neural network with one input, hidden, and output layer
def solve_ode_neural_network(x, num_neurons_hidden, num_iter, lmb):
## Set up initial weights and biases
# For the hidden layer
p0 = npr.randn(num_neurons_hidden, 2 )
# For the output layer
p1 = npr.randn(1, num_neurons_hidden + 1 ) # +1 since bias is included
P = [p0, p1]
print('Initial cost: %g'%cost_function(P, x))
## Start finding the optimal weights using gradient descent
# Find the Python function that represents the gradient of the cost function
# w.r.t the 0-th input argument -- that is the weights and biases in the hidden and output layer
cost_function_grad = grad(cost_function,0)
# Let the update be done num_iter times
for i in range(num_iter):
# Evaluate the gradient at the current weights and biases in P.
# The cost_grad consist now of two arrays;
# one for the gradient w.r.t P_hidden and
# one for the gradient w.r.t P_output
cost_grad = cost_function_grad(P, x)
P[0] = P[0] - lmb * cost_grad[0]
P[1] = P[1] - lmb * cost_grad[1]
print('Final cost: %g'%cost_function(P, x))
return P
def g_analytic(x, gamma = 2, g0 = 10):
return g0*np.exp(-gamma*x)
# Solve the given problem
if __name__ == '__main__':
# Set seed such that the weight are initialized
# with same weights and biases for every run.
npr.seed(15)
## Decide the vales of arguments to the function to solve
N = 10
x = np.linspace(0, 1, N)
## Set up the initial parameters
num_hidden_neurons = 10
num_iter = 10000
lmb = 0.001
# Use the network
P = solve_ode_neural_network(x, num_hidden_neurons, num_iter, lmb)
# Print the deviation from the trial solution and true solution
res = g_trial(x,P)
res_analytical = g_analytic(x)
print('Max absolute difference: %g'%np.max(np.abs(res - res_analytical)))
# Plot the results
plt.figure(figsize=(10,10))
plt.title('Performance of neural network solving an ODE compared to the analytical solution')
plt.plot(x, res_analytical)
plt.plot(x, res[0,:])
plt.legend(['analytical','nn'])
plt.xlabel('x')
plt.ylabel('g(x)')
plt.show()
Example: Population growth#
A logistic model of population growth assumes that a population converges toward an equilibrium. The population growth can be modeled by
where \(g(t)\) is the population density at time \(t\), \(\alpha > 0\) the growth rate and \(A > 0\) is the maximum population number in the environment. Also, at \(t = 0\) the population has the size \(g(0) = g_0\), where \(g_0\) is some chosen constant.
In this example, similar network as for the exponential decay using Autograd has been used to solve the equation. However, as the implementation might suffer from e.g numerical instability and high execution time (this might be more apparent in the examples solving PDEs), using a library like TensorFlow is recommended. Here, we stay with a more simple approach and implement for comparison, the simple forward Euler method.
Setting up the problem#
Here, we will model a population \(g(t)\) in an environment having carrying capacity \(A\). The population follows the model
where \(g(0) = g_0\).
In this example, we let \(\alpha = 2\), \(A = 1\), and \(g_0 = 1.2\).
The trial solution#
We will get a slightly different trial solution, as the boundary conditions are different compared to the case for exponential decay.
A possible trial solution satisfying the condition \(g(0) = g_0\) could be
with \(N(t,P)\) being the output from the neural network with weights and biases for each layer collected in the set \(P\).
The analytical solution is
The program using Autograd#
The network will be the similar as for the exponential decay example, but with some small modifications for our problem.
import autograd.numpy as np
from autograd import grad, elementwise_grad
import autograd.numpy.random as npr
from matplotlib import pyplot as plt
def sigmoid(z):
return 1/(1 + np.exp(-z))
# Function to get the parameters.
# Done such that one can easily change the paramaters after one's liking.
def get_parameters():
alpha = 2
A = 1
g0 = 1.2
return alpha, A, g0
def deep_neural_network(deep_params, x):
# N_hidden is the number of hidden layers
# deep_params is a list, len() should be used
N_hidden = len(deep_params) - 1 # -1 since params consists of
# parameters to all the hidden
# layers AND the output layer.
# Assumes input x being an one-dimensional array
num_values = np.size(x)
x = x.reshape(-1, num_values)
# Assume that the input layer does nothing to the input x
x_input = x
# Due to multiple hidden layers, define a variable referencing to the
# output of the previous layer:
x_prev = x_input
## Hidden layers:
for l in range(N_hidden):
# From the list of parameters P; find the correct weigths and bias for this layer
w_hidden = deep_params[l]
# Add a row of ones to include bias
x_prev = np.concatenate((np.ones((1,num_values)), x_prev ), axis = 0)
z_hidden = np.matmul(w_hidden, x_prev)
x_hidden = sigmoid(z_hidden)
# Update x_prev such that next layer can use the output from this layer
x_prev = x_hidden
## Output layer:
# Get the weights and bias for this layer
w_output = deep_params[-1]
# Include bias:
x_prev = np.concatenate((np.ones((1,num_values)), x_prev), axis = 0)
z_output = np.matmul(w_output, x_prev)
x_output = z_output
return x_output
def cost_function_deep(P, x):
# Evaluate the trial function with the current parameters P
g_t = g_trial_deep(x,P)
# Find the derivative w.r.t x of the trial function
d_g_t = elementwise_grad(g_trial_deep,0)(x,P)
# The right side of the ODE
func = f(x, g_t)
err_sqr = (d_g_t - func)**2
cost_sum = np.sum(err_sqr)
return cost_sum / np.size(err_sqr)
# The right side of the ODE:
def f(x, g_trial):
alpha,A, g0 = get_parameters()
return alpha*g_trial*(A - g_trial)
# The trial solution using the deep neural network:
def g_trial_deep(x, params):
alpha,A, g0 = get_parameters()
return g0 + x*deep_neural_network(params,x)
# The analytical solution:
def g_analytic(t):
alpha,A, g0 = get_parameters()
return A*g0/(g0 + (A - g0)*np.exp(-alpha*A*t))
def solve_ode_deep_neural_network(x, num_neurons, num_iter, lmb):
# num_hidden_neurons is now a list of number of neurons within each hidden layer
# Find the number of hidden layers:
N_hidden = np.size(num_neurons)
## Set up initial weigths and biases
# Initialize the list of parameters:
P = [None]*(N_hidden + 1) # + 1 to include the output layer
P[0] = npr.randn(num_neurons[0], 2 )
for l in range(1,N_hidden):
P[l] = npr.randn(num_neurons[l], num_neurons[l-1] + 1) # +1 to include bias
# For the output layer
P[-1] = npr.randn(1, num_neurons[-1] + 1 ) # +1 since bias is included
print('Initial cost: %g'%cost_function_deep(P, x))
## Start finding the optimal weigths using gradient descent
# Find the Python function that represents the gradient of the cost function
# w.r.t the 0-th input argument -- that is the weights and biases in the hidden and output layer
cost_function_deep_grad = grad(cost_function_deep,0)
# Let the update be done num_iter times
for i in range(num_iter):
# Evaluate the gradient at the current weights and biases in P.
# The cost_grad consist now of N_hidden + 1 arrays; the gradient w.r.t the weights and biases
# in the hidden layers and output layers evaluated at x.
cost_deep_grad = cost_function_deep_grad(P, x)
for l in range(N_hidden+1):
P[l] = P[l] - lmb * cost_deep_grad[l]
print('Final cost: %g'%cost_function_deep(P, x))
return P
if __name__ == '__main__':
npr.seed(4155)
## Decide the vales of arguments to the function to solve
Nt = 10
T = 1
t = np.linspace(0,T, Nt)
## Set up the initial parameters
num_hidden_neurons = [100, 50, 25]
num_iter = 1000
lmb = 1e-3
P = solve_ode_deep_neural_network(t, num_hidden_neurons, num_iter, lmb)
g_dnn_ag = g_trial_deep(t,P)
g_analytical = g_analytic(t)
# Find the maximum absolute difference between the solutons:
diff_ag = np.max(np.abs(g_dnn_ag - g_analytical))
print("The max absolute difference between the solutions is: %g"%diff_ag)
plt.figure(figsize=(10,10))
plt.title('Performance of neural network solving an ODE compared to the analytical solution')
plt.plot(t, g_analytical)
plt.plot(t, g_dnn_ag[0,:])
plt.legend(['analytical','nn'])
plt.xlabel('t')
plt.ylabel('g(t)')
plt.show()
Using forward Euler to solve the ODE#
A straightforward way of solving an ODE numerically, is to use Euler’s method.
Euler’s method uses Taylor series to approximate the value at a function \(f\) at a step \(\Delta x\) from \(x\):
In our case, using Euler’s method to approximate the value of \(g\) at a step \(\Delta t\) from \(t\) yields
along with the condition that \(g(0) = g_0\).
Let \(t_i = i \cdot \Delta t\) where \(\Delta t = \frac{T}{N_t-1}\) where \(T\) is the final time our solver must solve for and \(N_t\) the number of values for \(t \in [0, T]\) for \(i = 0, \dots, N_t-1\).
For \(i \geq 1\), we have that
Now, if \(g_i = g(t_i)\) then
for \(i \geq 1\) and \(g_0 = g(t_0) = g(0) = g_0\).
Equation (12) could be implemented in the following way, extending the program that uses the network using Autograd:
# Assume that all function definitions from the example program using Autograd
# are located here.
if __name__ == '__main__':
npr.seed(4155)
## Decide the vales of arguments to the function to solve
Nt = 10
T = 1
t = np.linspace(0,T, Nt)
## Set up the initial parameters
num_hidden_neurons = [100,50,25]
num_iter = 1000
lmb = 1e-3
P = solve_ode_deep_neural_network(t, num_hidden_neurons, num_iter, lmb)
g_dnn_ag = g_trial_deep(t,P)
g_analytical = g_analytic(t)
# Find the maximum absolute difference between the solutons:
diff_ag = np.max(np.abs(g_dnn_ag - g_analytical))
print("The max absolute difference between the solutions is: %g"%diff_ag)
plt.figure(figsize=(10,10))
plt.title('Performance of neural network solving an ODE compared to the analytical solution')
plt.plot(t, g_analytical)
plt.plot(t, g_dnn_ag[0,:])
plt.legend(['analytical','nn'])
plt.xlabel('t')
plt.ylabel('g(t)')
## Find an approximation to the funtion using forward Euler
alpha, A, g0 = get_parameters()
dt = T/(Nt - 1)
# Perform forward Euler to solve the ODE
g_euler = np.zeros(Nt)
g_euler[0] = g0
for i in range(1,Nt):
g_euler[i] = g_euler[i-1] + dt*(alpha*g_euler[i-1]*(A - g_euler[i-1]))
# Print the errors done by each method
diff1 = np.max(np.abs(g_euler - g_analytical))
diff2 = np.max(np.abs(g_dnn_ag[0,:] - g_analytical))
print('Max absolute difference between Euler method and analytical: %g'%diff1)
print('Max absolute difference between deep neural network and analytical: %g'%diff2)
# Plot results
plt.figure(figsize=(10,10))
plt.plot(t,g_euler)
plt.plot(t,g_analytical)
plt.plot(t,g_dnn_ag[0,:])
plt.legend(['euler','analytical','dnn'])
plt.xlabel('Time t')
plt.ylabel('g(t)')
plt.show()
Example: Solving the one dimensional Poisson equation#
The Poisson equation for \(g(x)\) in one dimension is
where \(f(x)\) is a given function for \(x \in (0,1)\).
The conditions that \(g(x)\) is chosen to fulfill, are
This equation can be solved numerically using programs where e.g Autograd and TensorFlow are used. The results from the networks can then be compared to the analytical solution. In addition, it could be interesting to see how a typical method for numerically solving second order ODEs compares to the neural networks.
The specific equation to solve for#
Here, the function \(g(x)\) to solve for follows the equation
where \(f(x)\) is a given function, along with the chosen conditions
In this example, we consider the case when \(f(x) = (3x + x^2)\exp(x)\).
For this case, a possible trial solution satisfying the conditions could be
The analytical solution for this problem is
Solving the equation using Autograd#
import autograd.numpy as np
from autograd import grad, elementwise_grad
import autograd.numpy.random as npr
from matplotlib import pyplot as plt
def sigmoid(z):
return 1/(1 + np.exp(-z))
def deep_neural_network(deep_params, x):
# N_hidden is the number of hidden layers
# deep_params is a list, len() should be used
N_hidden = len(deep_params) - 1 # -1 since params consists of
# parameters to all the hidden
# layers AND the output layer.
# Assumes input x being an one-dimensional array
num_values = np.size(x)
x = x.reshape(-1, num_values)
# Assume that the input layer does nothing to the input x
x_input = x
# Due to multiple hidden layers, define a variable referencing to the
# output of the previous layer:
x_prev = x_input
## Hidden layers:
for l in range(N_hidden):
# From the list of parameters P; find the correct weigths and bias for this layer
w_hidden = deep_params[l]
# Add a row of ones to include bias
x_prev = np.concatenate((np.ones((1,num_values)), x_prev ), axis = 0)
z_hidden = np.matmul(w_hidden, x_prev)
x_hidden = sigmoid(z_hidden)
# Update x_prev such that next layer can use the output from this layer
x_prev = x_hidden
## Output layer:
# Get the weights and bias for this layer
w_output = deep_params[-1]
# Include bias:
x_prev = np.concatenate((np.ones((1,num_values)), x_prev), axis = 0)
z_output = np.matmul(w_output, x_prev)
x_output = z_output
return x_output
def solve_ode_deep_neural_network(x, num_neurons, num_iter, lmb):
# num_hidden_neurons is now a list of number of neurons within each hidden layer
# Find the number of hidden layers:
N_hidden = np.size(num_neurons)
## Set up initial weigths and biases
# Initialize the list of parameters:
P = [None]*(N_hidden + 1) # + 1 to include the output layer
P[0] = npr.randn(num_neurons[0], 2 )
for l in range(1,N_hidden):
P[l] = npr.randn(num_neurons[l], num_neurons[l-1] + 1) # +1 to include bias
# For the output layer
P[-1] = npr.randn(1, num_neurons[-1] + 1 ) # +1 since bias is included
print('Initial cost: %g'%cost_function_deep(P, x))
## Start finding the optimal weigths using gradient descent
# Find the Python function that represents the gradient of the cost function
# w.r.t the 0-th input argument -- that is the weights and biases in the hidden and output layer
cost_function_deep_grad = grad(cost_function_deep,0)
# Let the update be done num_iter times
for i in range(num_iter):
# Evaluate the gradient at the current weights and biases in P.
# The cost_grad consist now of N_hidden + 1 arrays; the gradient w.r.t the weights and biases
# in the hidden layers and output layers evaluated at x.
cost_deep_grad = cost_function_deep_grad(P, x)
for l in range(N_hidden+1):
P[l] = P[l] - lmb * cost_deep_grad[l]
print('Final cost: %g'%cost_function_deep(P, x))
return P
## Set up the cost function specified for this Poisson equation:
# The right side of the ODE
def f(x):
return (3*x + x**2)*np.exp(x)
def cost_function_deep(P, x):
# Evaluate the trial function with the current parameters P
g_t = g_trial_deep(x,P)
# Find the derivative w.r.t x of the trial function
d2_g_t = elementwise_grad(elementwise_grad(g_trial_deep,0))(x,P)
right_side = f(x)
err_sqr = (-d2_g_t - right_side)**2
cost_sum = np.sum(err_sqr)
return cost_sum/np.size(err_sqr)
# The trial solution:
def g_trial_deep(x,P):
return x*(1-x)*deep_neural_network(P,x)
# The analytic solution;
def g_analytic(x):
return x*(1-x)*np.exp(x)
if __name__ == '__main__':
npr.seed(4155)
## Decide the vales of arguments to the function to solve
Nx = 10
x = np.linspace(0,1, Nx)
## Set up the initial parameters
num_hidden_neurons = [200,100]
num_iter = 1000
lmb = 1e-3
P = solve_ode_deep_neural_network(x, num_hidden_neurons, num_iter, lmb)
g_dnn_ag = g_trial_deep(x,P)
g_analytical = g_analytic(x)
# Find the maximum absolute difference between the solutons:
max_diff = np.max(np.abs(g_dnn_ag - g_analytical))
print("The max absolute difference between the solutions is: %g"%max_diff)
plt.figure(figsize=(10,10))
plt.title('Performance of neural network solving an ODE compared to the analytical solution')
plt.plot(x, g_analytical)
plt.plot(x, g_dnn_ag[0,:])
plt.legend(['analytical','nn'])
plt.xlabel('x')
plt.ylabel('g(x)')
plt.show()
Comparing with a numerical scheme#
The Poisson equation is possible to solve using Taylor series to approximate the second derivative.
Using Taylor series, the second derivative can be expressed as
where \(\Delta x\) is a small step size and \(E_{\Delta x}(x)\) being the error term.
Looking away from the error terms gives an approximation to the second derivative:
If \(x_i = i \Delta x = x_{i-1} + \Delta x\) and \(g_i = g(x_i)\) for \(i = 1,\dots N_x - 2\) with \(N_x\) being the number of values for \(x\), (15) becomes
Since we know from our problem that
along with the conditions \(g(0) = g(1) = 0\), the following scheme can be used to find an approximate solution for \(g(x)\) numerically:
for \(i = 1, \dots, N_x - 2\) where \(g_0 = g_{N_x - 1} = 0\) and \(f(x_i) = (3x_i + x_i^2)\exp(x_i)\), which is given for our specific problem.
The equation can be rewritten into a matrix equation:
which makes it possible to solve for the vector \(\boldsymbol{g}\).
Setting up the code#
We can then compare the result from this numerical scheme with the output from our network using Autograd:
import autograd.numpy as np
from autograd import grad, elementwise_grad
import autograd.numpy.random as npr
from matplotlib import pyplot as plt
def sigmoid(z):
return 1/(1 + np.exp(-z))
def deep_neural_network(deep_params, x):
# N_hidden is the number of hidden layers
# deep_params is a list, len() should be used
N_hidden = len(deep_params) - 1 # -1 since params consists of
# parameters to all the hidden
# layers AND the output layer.
# Assumes input x being an one-dimensional array
num_values = np.size(x)
x = x.reshape(-1, num_values)
# Assume that the input layer does nothing to the input x
x_input = x
# Due to multiple hidden layers, define a variable referencing to the
# output of the previous layer:
x_prev = x_input
## Hidden layers:
for l in range(N_hidden):
# From the list of parameters P; find the correct weigths and bias for this layer
w_hidden = deep_params[l]
# Add a row of ones to include bias
x_prev = np.concatenate((np.ones((1,num_values)), x_prev ), axis = 0)
z_hidden = np.matmul(w_hidden, x_prev)
x_hidden = sigmoid(z_hidden)
# Update x_prev such that next layer can use the output from this layer
x_prev = x_hidden
## Output layer:
# Get the weights and bias for this layer
w_output = deep_params[-1]
# Include bias:
x_prev = np.concatenate((np.ones((1,num_values)), x_prev), axis = 0)
z_output = np.matmul(w_output, x_prev)
x_output = z_output
return x_output
def solve_ode_deep_neural_network(x, num_neurons, num_iter, lmb):
# num_hidden_neurons is now a list of number of neurons within each hidden layer
# Find the number of hidden layers:
N_hidden = np.size(num_neurons)
## Set up initial weigths and biases
# Initialize the list of parameters:
P = [None]*(N_hidden + 1) # + 1 to include the output layer
P[0] = npr.randn(num_neurons[0], 2 )
for l in range(1,N_hidden):
P[l] = npr.randn(num_neurons[l], num_neurons[l-1] + 1) # +1 to include bias
# For the output layer
P[-1] = npr.randn(1, num_neurons[-1] + 1 ) # +1 since bias is included
print('Initial cost: %g'%cost_function_deep(P, x))
## Start finding the optimal weigths using gradient descent
# Find the Python function that represents the gradient of the cost function
# w.r.t the 0-th input argument -- that is the weights and biases in the hidden and output layer
cost_function_deep_grad = grad(cost_function_deep,0)
# Let the update be done num_iter times
for i in range(num_iter):
# Evaluate the gradient at the current weights and biases in P.
# The cost_grad consist now of N_hidden + 1 arrays; the gradient w.r.t the weights and biases
# in the hidden layers and output layers evaluated at x.
cost_deep_grad = cost_function_deep_grad(P, x)
for l in range(N_hidden+1):
P[l] = P[l] - lmb * cost_deep_grad[l]
print('Final cost: %g'%cost_function_deep(P, x))
return P
## Set up the cost function specified for this Poisson equation:
# The right side of the ODE
def f(x):
return (3*x + x**2)*np.exp(x)
def cost_function_deep(P, x):
# Evaluate the trial function with the current parameters P
g_t = g_trial_deep(x,P)
# Find the derivative w.r.t x of the trial function
d2_g_t = elementwise_grad(elementwise_grad(g_trial_deep,0))(x,P)
right_side = f(x)
err_sqr = (-d2_g_t - right_side)**2
cost_sum = np.sum(err_sqr)
return cost_sum/np.size(err_sqr)
# The trial solution:
def g_trial_deep(x,P):
return x*(1-x)*deep_neural_network(P,x)
# The analytic solution;
def g_analytic(x):
return x*(1-x)*np.exp(x)
if __name__ == '__main__':
npr.seed(4155)
## Decide the vales of arguments to the function to solve
Nx = 10
x = np.linspace(0,1, Nx)
## Set up the initial parameters
num_hidden_neurons = [200,100]
num_iter = 1000
lmb = 1e-3
P = solve_ode_deep_neural_network(x, num_hidden_neurons, num_iter, lmb)
g_dnn_ag = g_trial_deep(x,P)
g_analytical = g_analytic(x)
# Find the maximum absolute difference between the solutons:
plt.figure(figsize=(10,10))
plt.title('Performance of neural network solving an ODE compared to the analytical solution')
plt.plot(x, g_analytical)
plt.plot(x, g_dnn_ag[0,:])
plt.legend(['analytical','nn'])
plt.xlabel('x')
plt.ylabel('g(x)')
## Perform the computation using the numerical scheme
dx = 1/(Nx - 1)
# Set up the matrix A
A = np.zeros((Nx-2,Nx-2))
A[0,0] = 2
A[0,1] = -1
for i in range(1,Nx-3):
A[i,i-1] = -1
A[i,i] = 2
A[i,i+1] = -1
A[Nx - 3, Nx - 4] = -1
A[Nx - 3, Nx - 3] = 2
# Set up the vector f
f_vec = dx**2 * f(x[1:-1])
# Solve the equation
g_res = np.linalg.solve(A,f_vec)
g_vec = np.zeros(Nx)
g_vec[1:-1] = g_res
# Print the differences between each method
max_diff1 = np.max(np.abs(g_dnn_ag - g_analytical))
max_diff2 = np.max(np.abs(g_vec - g_analytical))
print("The max absolute difference between the analytical solution and DNN Autograd: %g"%max_diff1)
print("The max absolute difference between the analytical solution and numerical scheme: %g"%max_diff2)
# Plot the results
plt.figure(figsize=(10,10))
plt.plot(x,g_vec)
plt.plot(x,g_analytical)
plt.plot(x,g_dnn_ag[0,:])
plt.legend(['numerical scheme','analytical','dnn'])
plt.show()
Partial Differential Equations#
A partial differential equation (PDE) has a solution here the function is defined by multiple variables. The equation may involve all kinds of combinations of which variables the function is differentiated with respect to.
In general, a partial differential equation for a function \(g(x_1,\dots,x_N)\) with \(N\) variables may be expressed as
where \(f\) is an expression involving all kinds of possible mixed derivatives of \(g(x_1,\dots,x_N)\) up to an order \(n\). In order for the solution to be unique, some additional conditions must also be given.
Type of problem#
The problem our network must solve for, is similar to the ODE case. We must have a trial solution \(g_t\) at hand.
For instance, the trial solution could be expressed as
where \(h_1(x_1,\dots,x_N)\) is a function that ensures \(g_t(x_1,\dots,x_N)\) satisfies some given conditions. The neural network \(N(x_1,\dots,x_N,P)\) has weights and biases described by \(P\) and \(h_2(x_1,\dots,x_N,N(x_1,\dots,x_N,P))\) is an expression using the output from the neural network in some way.
The role of the function \(h_2(x_1,\dots,x_N,N(x_1,\dots,x_N,P))\), is to ensure that the output of \(N(x_1,\dots,x_N,P)\) is zero when \(g_t(x_1,\dots,x_N)\) is evaluated at the values of \(x_1,\dots,x_N\) where the given conditions must be satisfied. The function \(h_1(x_1,\dots,x_N)\) should alone make \(g_t(x_1,\dots,x_N)\) satisfy the conditions.
Network requirements#
The network tries then the minimize the cost function following the same ideas as described for the ODE case, but now with more than one variables to consider. The concept still remains the same; find a set of parameters \(P\) such that the expression \(f\) in (17) is as close to zero as possible.
As for the ODE case, the cost function is the mean squared error that the network must try to minimize. The cost function for the network to minimize is
More details#
If we let \(\boldsymbol{x} = \big( x_1, \dots, x_N \big)\) be an array containing the values for \(x_1, \dots, x_N\) respectively, the cost function can be reformulated into the following:
If we also have \(M\) different sets of values for \(x_1, \dots, x_N\), that is \(\boldsymbol{x}_i = \big(x_1^{(i)}, \dots, x_N^{(i)}\big)\) for \(i = 1,\dots,M\) being the rows in matrix \(X\), the cost function can be generalized into
Example: The diffusion equation#
In one spatial dimension, the equation reads
where a possible choice of conditions are
with \(u(x)\) being some given function.
Defining the problem#
For this case, we want to find \(g(x,t)\) such that
and
with \(u(x) = \sin(\pi x)\).
First, let us set up the deep neural network. The deep neural network will follow the same structure as discussed in the examples solving the ODEs. First, we will look into how Autograd could be used in a network tailored to solve for bivariate functions.
Setting up the network using Autograd#
The only change to do here, is to extend our network such that functions of multiple parameters are correctly handled. In this case we have two variables in our function to solve for, that is time \(t\) and position \(x\). The variables will be represented by a one-dimensional array in the program. The program will evaluate the network at each possible pair \((x,t)\), given an array for the desired \(x\)-values and \(t\)-values to approximate the solution at.
def sigmoid(z):
return 1/(1 + np.exp(-z))
def deep_neural_network(deep_params, x):
# x is now a point and a 1D numpy array; make it a column vector
num_coordinates = np.size(x,0)
x = x.reshape(num_coordinates,-1)
num_points = np.size(x,1)
# N_hidden is the number of hidden layers
N_hidden = len(deep_params) - 1 # -1 since params consist of parameters to all the hidden layers AND the output layer
# Assume that the input layer does nothing to the input x
x_input = x
x_prev = x_input
## Hidden layers:
for l in range(N_hidden):
# From the list of parameters P; find the correct weigths and bias for this layer
w_hidden = deep_params[l]
# Add a row of ones to include bias
x_prev = np.concatenate((np.ones((1,num_points)), x_prev ), axis = 0)
z_hidden = np.matmul(w_hidden, x_prev)
x_hidden = sigmoid(z_hidden)
# Update x_prev such that next layer can use the output from this layer
x_prev = x_hidden
## Output layer:
# Get the weights and bias for this layer
w_output = deep_params[-1]
# Include bias:
x_prev = np.concatenate((np.ones((1,num_points)), x_prev), axis = 0)
z_output = np.matmul(w_output, x_prev)
x_output = z_output
return x_output[0][0]
Setting up the network using Autograd; The trial solution#
The cost function must then iterate through the given arrays containing values for \(x\) and \(t\), defines a point \((x,t)\) the deep neural network and the trial solution is evaluated at, and then finds the Jacobian of the trial solution.
A possible trial solution for this PDE is
with \(A(x,t)\) being a function ensuring that \(g_t(x,t)\) satisfies our given conditions, and \(N(x,t,P)\) being the output from the deep neural network using weights and biases for each layer from \(P\).
To fulfill the conditions, \(A(x,t)\) could be:
since \((0) = u(1) = 0\) and \(u(x) = \sin(\pi x)\).
Why the jacobian?#
The Jacobian is used because the program must find the derivative of the trial solution with respect to \(x\) and \(t\).
This gives the necessity of computing the Jacobian matrix, as we want to evaluate the gradient with respect to \(x\) and \(t\) (note that the Jacobian of a scalar-valued multivariate function is simply its gradient).
In Autograd, the differentiation is by default done with respect to the first input argument of your Python function. Since the points is an array representing \(x\) and \(t\), the Jacobian is calculated using the values of \(x\) and \(t\).
To find the second derivative with respect to \(x\) and \(t\), the Jacobian can be found for the second time. The result is a Hessian matrix, which is the matrix containing all the possible second order mixed derivatives of \(g(x,t)\).
# Set up the trial function:
def u(x):
return np.sin(np.pi*x)
def g_trial(point,P):
x,t = point
return (1-t)*u(x) + x*(1-x)*t*deep_neural_network(P,point)
# The right side of the ODE:
def f(point):
return 0.
# The cost function:
def cost_function(P, x, t):
cost_sum = 0
g_t_jacobian_func = jacobian(g_trial)
g_t_hessian_func = hessian(g_trial)
for x_ in x:
for t_ in t:
point = np.array([x_,t_])
g_t = g_trial(point,P)
g_t_jacobian = g_t_jacobian_func(point,P)
g_t_hessian = g_t_hessian_func(point,P)
g_t_dt = g_t_jacobian[1]
g_t_d2x = g_t_hessian[0][0]
func = f(point)
err_sqr = ( (g_t_dt - g_t_d2x) - func)**2
cost_sum += err_sqr
return cost_sum
Setting up the network using Autograd; The full program#
Having set up the network, along with the trial solution and cost function, we can now see how the deep neural network performs by comparing the results to the analytical solution.
The analytical solution of our problem is
A possible way to implement a neural network solving the PDE, is given below. Be aware, though, that it is fairly slow for the parameters used. A better result is possible, but requires more iterations, and thus longer time to complete.
Indeed, the program below is not optimal in its implementation, but rather serves as an example on how to implement and use a neural network to solve a PDE. Using TensorFlow results in a much better execution time. Try it!
import autograd.numpy as np
from autograd import jacobian,hessian,grad
import autograd.numpy.random as npr
from matplotlib import cm
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import axes3d
## Set up the network
def sigmoid(z):
return 1/(1 + np.exp(-z))
def deep_neural_network(deep_params, x):
# x is now a point and a 1D numpy array; make it a column vector
num_coordinates = np.size(x,0)
x = x.reshape(num_coordinates,-1)
num_points = np.size(x,1)
# N_hidden is the number of hidden layers
N_hidden = len(deep_params) - 1 # -1 since params consist of parameters to all the hidden layers AND the output layer
# Assume that the input layer does nothing to the input x
x_input = x
x_prev = x_input
## Hidden layers:
for l in range(N_hidden):
# From the list of parameters P; find the correct weigths and bias for this layer
w_hidden = deep_params[l]
# Add a row of ones to include bias
x_prev = np.concatenate((np.ones((1,num_points)), x_prev ), axis = 0)
z_hidden = np.matmul(w_hidden, x_prev)
x_hidden = sigmoid(z_hidden)
# Update x_prev such that next layer can use the output from this layer
x_prev = x_hidden
## Output layer:
# Get the weights and bias for this layer
w_output = deep_params[-1]
# Include bias:
x_prev = np.concatenate((np.ones((1,num_points)), x_prev), axis = 0)
z_output = np.matmul(w_output, x_prev)
x_output = z_output
return x_output[0][0]
## Define the trial solution and cost function
def u(x):
return np.sin(np.pi*x)
def g_trial(point,P):
x,t = point
return (1-t)*u(x) + x*(1-x)*t*deep_neural_network(P,point)
# The right side of the ODE:
def f(point):
return 0.
# The cost function:
def cost_function(P, x, t):
cost_sum = 0
g_t_jacobian_func = jacobian(g_trial)
g_t_hessian_func = hessian(g_trial)
for x_ in x:
for t_ in t:
point = np.array([x_,t_])
g_t = g_trial(point,P)
g_t_jacobian = g_t_jacobian_func(point,P)
g_t_hessian = g_t_hessian_func(point,P)
g_t_dt = g_t_jacobian[1]
g_t_d2x = g_t_hessian[0][0]
func = f(point)
err_sqr = ( (g_t_dt - g_t_d2x) - func)**2
cost_sum += err_sqr
return cost_sum /( np.size(x)*np.size(t) )
## For comparison, define the analytical solution
def g_analytic(point):
x,t = point
return np.exp(-np.pi**2*t)*np.sin(np.pi*x)
## Set up a function for training the network to solve for the equation
def solve_pde_deep_neural_network(x,t, num_neurons, num_iter, lmb):
## Set up initial weigths and biases
N_hidden = np.size(num_neurons)
## Set up initial weigths and biases
# Initialize the list of parameters:
P = [None]*(N_hidden + 1) # + 1 to include the output layer
P[0] = npr.randn(num_neurons[0], 2 + 1 ) # 2 since we have two points, +1 to include bias
for l in range(1,N_hidden):
P[l] = npr.randn(num_neurons[l], num_neurons[l-1] + 1) # +1 to include bias
# For the output layer
P[-1] = npr.randn(1, num_neurons[-1] + 1 ) # +1 since bias is included
print('Initial cost: ',cost_function(P, x, t))
cost_function_grad = grad(cost_function,0)
# Let the update be done num_iter times
for i in range(num_iter):
cost_grad = cost_function_grad(P, x , t)
for l in range(N_hidden+1):
P[l] = P[l] - lmb * cost_grad[l]
print('Final cost: ',cost_function(P, x, t))
return P
if __name__ == '__main__':
### Use the neural network:
npr.seed(15)
## Decide the vales of arguments to the function to solve
Nx = 10; Nt = 10
x = np.linspace(0, 1, Nx)
t = np.linspace(0,1,Nt)
## Set up the parameters for the network
num_hidden_neurons = [100, 25]
num_iter = 250
lmb = 0.01
P = solve_pde_deep_neural_network(x,t, num_hidden_neurons, num_iter, lmb)
## Store the results
g_dnn_ag = np.zeros((Nx, Nt))
G_analytical = np.zeros((Nx, Nt))
for i,x_ in enumerate(x):
for j, t_ in enumerate(t):
point = np.array([x_, t_])
g_dnn_ag[i,j] = g_trial(point,P)
G_analytical[i,j] = g_analytic(point)
# Find the map difference between the analytical and the computed solution
diff_ag = np.abs(g_dnn_ag - G_analytical)
print('Max absolute difference between the analytical solution and the network: %g'%np.max(diff_ag))
## Plot the solutions in two dimensions, that being in position and time
T,X = np.meshgrid(t,x)
fig = plt.figure(figsize=(10,10))
ax = fig.add_suplot(projection='3d')
ax.set_title('Solution from the deep neural network w/ %d layer'%len(num_hidden_neurons))
s = ax.plot_surface(T,X,g_dnn_ag,linewidth=0,antialiased=False,cmap=cm.viridis)
ax.set_xlabel('Time $t$')
ax.set_ylabel('Position $x$');
fig = plt.figure(figsize=(10,10))
ax = fig.add_suplot(projection='3d')
ax.set_title('Analytical solution')
s = ax.plot_surface(T,X,G_analytical,linewidth=0,antialiased=False,cmap=cm.viridis)
ax.set_xlabel('Time $t$')
ax.set_ylabel('Position $x$');
fig = plt.figure(figsize=(10,10))
ax = fig.add_suplot(projection='3d')
ax.set_title('Difference')
s = ax.plot_surface(T,X,diff_ag,linewidth=0,antialiased=False,cmap=cm.viridis)
ax.set_xlabel('Time $t$')
ax.set_ylabel('Position $x$');
## Take some slices of the 3D plots just to see the solutions at particular times
indx1 = 0
indx2 = int(Nt/2)
indx3 = Nt-1
t1 = t[indx1]
t2 = t[indx2]
t3 = t[indx3]
# Slice the results from the DNN
res1 = g_dnn_ag[:,indx1]
res2 = g_dnn_ag[:,indx2]
res3 = g_dnn_ag[:,indx3]
# Slice the analytical results
res_analytical1 = G_analytical[:,indx1]
res_analytical2 = G_analytical[:,indx2]
res_analytical3 = G_analytical[:,indx3]
# Plot the slices
plt.figure(figsize=(10,10))
plt.title("Computed solutions at time = %g"%t1)
plt.plot(x, res1)
plt.plot(x,res_analytical1)
plt.legend(['dnn','analytical'])
plt.figure(figsize=(10,10))
plt.title("Computed solutions at time = %g"%t2)
plt.plot(x, res2)
plt.plot(x,res_analytical2)
plt.legend(['dnn','analytical'])
plt.figure(figsize=(10,10))
plt.title("Computed solutions at time = %g"%t3)
plt.plot(x, res3)
plt.plot(x,res_analytical3)
plt.legend(['dnn','analytical'])
plt.show()
Example: Solving the wave equation with Neural Networks#
The wave equation is
with \(c\) being the specified wave speed.
Here, the chosen conditions are
where \(\frac{\partial g(x,t)}{\partial t} \Big |_{t = 0}\) means the derivative of \(g(x,t)\) with respect to \(t\) is evaluated at \(t = 0\), and \(u(x)\) and \(v(x)\) being given functions.
The problem to solve for#
The wave equation to solve for, is
where \(c\) is the given wave speed. The chosen conditions for this equation are
In this example, let \(c = 1\) and \(u(x) = \sin(\pi x)\) and \(v(x) = -\pi\sin(\pi x)\).
The trial solution#
Setting up the network is done in similar matter as for the example of solving the diffusion equation. The only things we have to change, is the trial solution such that it satisfies the conditions from (20) and the cost function.
The trial solution becomes slightly different since we have other conditions than in the example of solving the diffusion equation. Here, a possible trial solution \(g_t(x,t)\) is
where
Note that this trial solution satisfies the conditions only if \(u(0) = v(0) = u(1) = v(1) = 0\), which is the case in this example.
The analytical solution#
The analytical solution for our specific problem, is
Solving the wave equation - the full program using Autograd#
import autograd.numpy as np
from autograd import hessian,grad
import autograd.numpy.random as npr
from matplotlib import cm
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import axes3d
## Set up the trial function:
def u(x):
return np.sin(np.pi*x)
def v(x):
return -np.pi*np.sin(np.pi*x)
def h1(point):
x,t = point
return (1 - t**2)*u(x) + t*v(x)
def g_trial(point,P):
x,t = point
return h1(point) + x*(1-x)*t**2*deep_neural_network(P,point)
## Define the cost function
def cost_function(P, x, t):
cost_sum = 0
g_t_hessian_func = hessian(g_trial)
for x_ in x:
for t_ in t:
point = np.array([x_,t_])
g_t_hessian = g_t_hessian_func(point,P)
g_t_d2x = g_t_hessian[0][0]
g_t_d2t = g_t_hessian[1][1]
err_sqr = ( (g_t_d2t - g_t_d2x) )**2
cost_sum += err_sqr
return cost_sum / (np.size(t) * np.size(x))
## The neural network
def sigmoid(z):
return 1/(1 + np.exp(-z))
def deep_neural_network(deep_params, x):
# x is now a point and a 1D numpy array; make it a column vector
num_coordinates = np.size(x,0)
x = x.reshape(num_coordinates,-1)
num_points = np.size(x,1)
# N_hidden is the number of hidden layers
N_hidden = len(deep_params) - 1 # -1 since params consist of parameters to all the hidden layers AND the output layer
# Assume that the input layer does nothing to the input x
x_input = x
x_prev = x_input
## Hidden layers:
for l in range(N_hidden):
# From the list of parameters P; find the correct weigths and bias for this layer
w_hidden = deep_params[l]
# Add a row of ones to include bias
x_prev = np.concatenate((np.ones((1,num_points)), x_prev ), axis = 0)
z_hidden = np.matmul(w_hidden, x_prev)
x_hidden = sigmoid(z_hidden)
# Update x_prev such that next layer can use the output from this layer
x_prev = x_hidden
## Output layer:
# Get the weights and bias for this layer
w_output = deep_params[-1]
# Include bias:
x_prev = np.concatenate((np.ones((1,num_points)), x_prev), axis = 0)
z_output = np.matmul(w_output, x_prev)
x_output = z_output
return x_output[0][0]
## The analytical solution
def g_analytic(point):
x,t = point
return np.sin(np.pi*x)*np.cos(np.pi*t) - np.sin(np.pi*x)*np.sin(np.pi*t)
def solve_pde_deep_neural_network(x,t, num_neurons, num_iter, lmb):
## Set up initial weigths and biases
N_hidden = np.size(num_neurons)
## Set up initial weigths and biases
# Initialize the list of parameters:
P = [None]*(N_hidden + 1) # + 1 to include the output layer
P[0] = npr.randn(num_neurons[0], 2 + 1 ) # 2 since we have two points, +1 to include bias
for l in range(1,N_hidden):
P[l] = npr.randn(num_neurons[l], num_neurons[l-1] + 1) # +1 to include bias
# For the output layer
P[-1] = npr.randn(1, num_neurons[-1] + 1 ) # +1 since bias is included
print('Initial cost: ',cost_function(P, x, t))
cost_function_grad = grad(cost_function,0)
# Let the update be done num_iter times
for i in range(num_iter):
cost_grad = cost_function_grad(P, x , t)
for l in range(N_hidden+1):
P[l] = P[l] - lmb * cost_grad[l]
print('Final cost: ',cost_function(P, x, t))
return P
if __name__ == '__main__':
### Use the neural network:
npr.seed(15)
## Decide the vales of arguments to the function to solve
Nx = 10; Nt = 10
x = np.linspace(0, 1, Nx)
t = np.linspace(0,1,Nt)
## Set up the parameters for the network
num_hidden_neurons = [50,20]
num_iter = 1000
lmb = 0.01
P = solve_pde_deep_neural_network(x,t, num_hidden_neurons, num_iter, lmb)
## Store the results
res = np.zeros((Nx, Nt))
res_analytical = np.zeros((Nx, Nt))
for i,x_ in enumerate(x):
for j, t_ in enumerate(t):
point = np.array([x_, t_])
res[i,j] = g_trial(point,P)
res_analytical[i,j] = g_analytic(point)
diff = np.abs(res - res_analytical)
print("Max difference between analytical and solution from nn: %g"%np.max(diff))
## Plot the solutions in two dimensions, that being in position and time
T,X = np.meshgrid(t,x)
fig = plt.figure(figsize=(10,10))
ax = fig.add_suplot(projection='3d')
ax.set_title('Solution from the deep neural network w/ %d layer'%len(num_hidden_neurons))
s = ax.plot_surface(T,X,res,linewidth=0,antialiased=False,cmap=cm.viridis)
ax.set_xlabel('Time $t$')
ax.set_ylabel('Position $x$');
fig = plt.figure(figsize=(10,10))
ax = fig.add_suplot(projection='3d')
ax.set_title('Analytical solution')
s = ax.plot_surface(T,X,res_analytical,linewidth=0,antialiased=False,cmap=cm.viridis)
ax.set_xlabel('Time $t$')
ax.set_ylabel('Position $x$');
fig = plt.figure(figsize=(10,10))
ax = fig.add_suplot(projection='3d')
ax.set_title('Difference')
s = ax.plot_surface(T,X,diff,linewidth=0,antialiased=False,cmap=cm.viridis)
ax.set_xlabel('Time $t$')
ax.set_ylabel('Position $x$');
## Take some slices of the 3D plots just to see the solutions at particular times
indx1 = 0
indx2 = int(Nt/2)
indx3 = Nt-1
t1 = t[indx1]
t2 = t[indx2]
t3 = t[indx3]
# Slice the results from the DNN
res1 = res[:,indx1]
res2 = res[:,indx2]
res3 = res[:,indx3]
# Slice the analytical results
res_analytical1 = res_analytical[:,indx1]
res_analytical2 = res_analytical[:,indx2]
res_analytical3 = res_analytical[:,indx3]
# Plot the slices
plt.figure(figsize=(10,10))
plt.title("Computed solutions at time = %g"%t1)
plt.plot(x, res1)
plt.plot(x,res_analytical1)
plt.legend(['dnn','analytical'])
plt.figure(figsize=(10,10))
plt.title("Computed solutions at time = %g"%t2)
plt.plot(x, res2)
plt.plot(x,res_analytical2)
plt.legend(['dnn','analytical'])
plt.figure(figsize=(10,10))
plt.title("Computed solutions at time = %g"%t3)
plt.plot(x, res3)
plt.plot(x,res_analytical3)
plt.legend(['dnn','analytical'])
plt.show()