SlideShare a Scribd company logo
1 of 68
Prepared by: Eng. Javier Daza Piragauta
EXPERT SYSTEMS
AND ARTIFICIAL
INTELLIGENCE
REACTIVE SYSTEMS –
NEURAL NETWORKS
October 13, 2023
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
REACTIVE SYSTEMS
October 13, 2023
Reactive Systems are a class of software systems designed to react to events and
changes in their environment in a timely and efficient manner. These systems are typically
used in real-time and critical applications where quick and reliable responses are crucial.
Key characteristics of reactive systems include:
• Responsiveness: Reactive systems respond rapidly to input and events from their
environment, minimizing latency.
• Scalability: They can scale horizontally or vertically to handle varying workloads.
• Resilience: These systems are resilient to failures and can recover from unexpected
situations.
• Actor Model: Reactive systems often rely on the actor model, where independent
components (actors) interact through message passing.
Reactive systems find applications in a variety of domains, including online trading
systems, real-time gaming systems, flight controllers, and infrastructure monitoring.
NEURAL NETWORKS
October 13, 2023
 Neural Networks:
Neural Networks are a type of artificial intelligence model inspired by the functioning of the human brain.
They consist of layers of interconnected artificial neurons that process input data, perform computations,
and generate outputs. Neural networks are used for machine learning and data processing tasks,
including:
• Supervised Learning: Neural networks can learn from labeled data, making them suitable for tasks
like image classification, text translation, or numeric value prediction.
• Unsupervised Learning: They are also used for tasks such as clustering, dimensionality reduction,
and pattern detection in unlabeled data.
• Deep Learning: Deep Neural Networks (DNNs) are a subtype of neural networks with multiple
hidden layers that can handle highly complex tasks like image processing and natural language
understanding.
Neural networks have revolutionized many fields of computer science and artificial intelligence,
including speech recognition, computer vision, autonomous driving, and online content recommendation.
In summary, Reactive Systems focus on rapid and efficient reaction to events in real-time
applications, while Neural Networks are artificial intelligence models capable of learning and performing
complex tasks based on data processing and are widely used in the machine learning domain. Both
concepts have significant applications and play crucial roles in different areas of technology and science.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Neural Networks - I Neural Networks - II
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
Structure of an artificial neural system
October 13, 2023
An artificial neural system, also known as an artificial neural
network (ANN) or simply a neural network, is a computational model
inspired by the structure and function of biological neural networks
in the human brain.
The structure of an artificial neural system consists of several
interconnected layers, each with its unique role in processing
information.
Here is a breakdown of the typical structure of an artificial neural
network:
Structure of an artificial neural system
October 13, 2023
1. Input Layer:
The input layer is the first layer of the neural network. It receives
external data or features as input. Each neuron in the input layer
represents a feature or attribute of the input data. The number of
neurons in this layer is determined by the dimensionality of the
input data.
2. Hidden Layers:
Between the input and output layers, there may be one or more
hidden layers. These layers are called "hidden" because they are not
directly connected to the external data or the final output. The
hidden layers perform complex computations and feature extraction.
The number of hidden layers and neurons in each layer is determined
by the architecture of the network, and it can vary widely.
Structure of an artificial neural system
October 13, 2023
3. Neurons (Nodes):
Each layer consists of interconnected neurons, also known as
nodes or units. Neurons in one layer are connected to neurons in the
subsequent layer through weighted connections. Neurons process
information by summing up the weighted inputs, applying an
activation function, and passing the result to the next layer.
4. Weights and Connections:
Connections between neurons are associated with weights.
These weights determine the strength of the connection and play a
crucial role in learning. During training, the network adjusts the
weights to improve its performance in making predictions or
classifications.
Structure of an artificial neural system
October 13, 2023
5. Activation Functions:
Activation functions introduce non-linearity into the neural
network. Common activation functions include the sigmoid function,
the hyperbolic tangent (tanh) function, and the rectified linear unit
(ReLU) function. The activation function defines the output of a neuron
based on its weighted sum of inputs.
6. Output Layer:
The output layer produces the final results of the neural network's
computation. The number of neurons in the output layer depends on
the specific task the neural network is designed for. For example, in a
binary classification task, there may be one output neuron representing
the probability of belonging to one class, while in multi-class
classification, there would be multiple output neurons, each
representing a different class.
Structure of an artificial neural system
October 13, 2023
7. Bias Neurons:
Some neural networks include bias neurons in each layer. These
neurons have a constant input value of 1 and are associated with their
own weights. Bias neurons allow the network to account for systematic
errors or biases in the data.
8. Training Algorithm:
To make the neural network perform a specific task, it must be
trained on labeled data. Training involves adjusting the weights of the
connections to minimize a cost function or loss function that measures
the difference between the network's predictions and the true values in
the training data. Common training algorithms include backpropagation
and variants like stochastic gradient descent (SGD) and Adam.
Structure of an artificial neural system
October 13, 2023
9. Output Activation Function:
The choice of activation function in the output layer depends on the
nature of the task. For regression tasks, a linear activation function may be
used. For binary classification, a sigmoid or softmax function is common,
and for multi-class classification, softmax is typically used.
The structure and architecture of neural networks can vary significantly
based on the specific problem they are designed to solve. Deep neural
networks (DNNs) have multiple hidden layers and have proven effective in
solving complex tasks, while simpler architectures like feedforward neural
networks have fewer layers and are used for simpler problems.
Convolutional Neural Networks (CNNs) are specialized for image-related
tasks, and Recurrent Neural Networks (RNNs) are designed for sequential
data processing. The choice of architecture and structure depends on the
problem domain and the data characteristics.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Artificial neuron model
October 13, 2023
Artificial neuron model
October 13, 2023
An artificial neuron, also known as a perceptron, is the fundamental
building block of artificial neural networks. It is a simplified computational
model inspired by the functioning of biological neurons in the human
brain.
The artificial neuron processes input data and produces an output
based on a set of weighted connections, an activation function, and a bias
term. Here's an overview of the components and operation of an artificial
neuron:
Artificial neuron model
October 13, 2023
Components of an Artificial Neuron:
1. Inputs (X1, X2, ..., Xn): An artificial neuron receives input signals, often
represented as X1, X2, ..., Xn. Each input represents a feature or
attribute of the data being processed.
2. Weights (W1, W2, ..., Wn): Each input is associated with a weight,
represented as W1, W2, ..., Wn. These weights determine the
importance of each input in the neuron's computation. The weights
are parameters that the neuron learns during training to optimize its
performance.
3. Bias (b): The bias, denoted as "b," is an additional parameter. It allows
the neuron to adjust its output independently of the inputs and is
crucial for modeling translation invariance.
Artificial neuron model
October 13, 2023
Components of an Artificial Neuron:
4. Weighted Sum (Z): The weighted sum (Z) is calculated by multiplying
each input by its corresponding weight and summing these products. It
is expressed as:
Z = (X1 * W1) + (X2 * W2) + ... + (Xn * Wn) + b
5. Activation Function (f): The weighted sum is passed through an
activation function (f). The activation function introduces non-linearity
into the model. Common activation functions include the sigmoid,
hyperbolic tangent (tanh), and rectified linear unit (ReLU). The choice
of activation function depends on the specific problem and the desired
properties of the neuron.
Artificial neuron model
October 13, 2023
Components of an Artificial Neuron:
6. Output (Y): The output (Y) of the neuron is the result of applying the
activation function to the weighted sum:
Y = f(Z)
Artificial neuron model
October 13, 2023
Operation of an Artificial Neuron:
1. The artificial neuron receives input signals (X1, X2, ..., Xn).
2. Each input is multiplied by its corresponding weight (W1, W2, ..., Wn).
3. The weighted products are summed, and the bias term is added to the
sum to calculate the weighted sum (Z).
4. The activation function (f) is applied to the weighted sum (Z) to
produce the neuron's output (Y).
.
Artificial neuron model
October 13, 2023
Operation of an Artificial Neuron:
In essence, the artificial neuron processes inputs by adjusting the
weights and bias during training to find the optimal parameters that
minimize the error between its output and the desired output. Once
trained, the neuron can be used to make predictions or perform various
computations in the context of a neural network.
Artificial neurons are typically organized into layers to form neural
networks. In this network context, they can learn complex patterns and
relationships in data, making them a fundamental component of machine
learning and deep learning models.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Dynamic neuron model
October 13, 2023
Energy-Efficient Tuning of Spintronic Neurons to Imitate the Non-linear
Oscillatory Neural Networks of the Human Brain
Dynamic neuron model
October 13, 2023
A "dynamic neuron model" typically refers to a computational model
of a neuron that incorporates more complex and dynamic behavior than
the simplified artificial neuron model, such as the integrate-and-fire
model. Dynamic neuron models are used in computational neuroscience
and neural network research to better capture the behavior of biological
neurons. One example of a dynamic neuron model is the Hodgkin-Huxley
model. Here's an overview of dynamic neuron models:
Hodgkin-Huxley Model:
The Hodgkin-Huxley model is a well-known dynamic neuron model
that provides a more biologically accurate representation of how
individual neurons operate. This model was developed by Sir Alan Hodgkin
and Sir Andrew Huxley in 1952 and earned them the Nobel Prize in
Physiology or Medicine in 1963. The model is commonly used to describe
the behavior of neurons in the context of action potentials or spikes.
Dynamic neuron model
October 13, 2023
Key features of the Hodgkin-Huxley model:
1. Membrane Potential (Voltage): Unlike the simplified artificial neuron
model, the Hodgkin-Huxley model accounts for the neuron's
membrane potential and how it changes over time.
2. Ion Channels: The model incorporates different types of ion channels,
including sodium (Na+), potassium (K+), and leak channels. These ion
channels allow for the flow of ions in and out of the neuron,
influencing the neuron's membrane potential.
3. Differential Equations: The dynamics of the model are described using
a system of differential equations that govern the behavior of ion
channels and the membrane potential.
Dynamic neuron model
October 13, 2023
Key features of the Hodgkin-Huxley model:
5. Action Potentials: The model accurately represents the generation and
propagation of action potentials, which are the electrical signals that
neurons use to communicate with each other.
6. Nonlinear Behavior: The Hodgkin-Huxley model exhibits nonlinear
behavior, which is essential for modeling the complex dynamics of real
neurons.
The Hodgkin-Huxley model, and similar dynamic neuron models, can be
used to simulate and understand the complex electrical activity in
neurons, including the initiation and propagation of action potentials.
These models are valuable tools in the study of neural circuits,
neurophysiology, and the development of more realistic computational
Dynamic neuron model
October 13, 2023
Key features of the Hodgkin-Huxley model:
The Hodgkin-Huxley model, and similar dynamic neuron models, can
be used to simulate and understand the complex electrical activity in
neurons, including the initiation and propagation of action potentials.
These models are valuable tools in the study of neural circuits,
neurophysiology, and the development of more realistic computational
models of neural networks.
Dynamic neuron models are essential in fields like computational
neuroscience, where the focus is on understanding the intricacies of
biological neurons and their interaction in the brain. While simple artificial
neurons are used in machine learning and deep learning for practical
purposes, dynamic neuron models are crucial for modeling the underlying
biology of the nervous system.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Stochastic neuron model
October 13, 2023
Stochastic neuron model
October 13, 2023
A stochastic neuron model is a type of artificial neuron or
computational model of a biological neuron that incorporates stochastic
(random) elements into its operation. Unlike deterministic neuron models,
which produce fixed outputs given fixed inputs, stochastic neuron models
introduce randomness into the process, making them useful in modeling
certain aspects of neural behavior that involve uncertainty or noise.
There are various ways to incorporate stochasticity into neuron
models, and it depends on the specific context and the level of abstraction
you are interested in. Here are a few common approaches:
Stochastic neuron model
October 13, 2023
1. Stochastic Threshold Model: In this model, the neuron's firing
threshold is not fixed but instead varies randomly. The neuron
integrates incoming signals as usual, but whether it fires or not
depends on a stochastic process. This can help capture the variability
in neural firing observed in real biological neurons.
1. Stochastic Input Model: Instead of a fixed input signal, this model
assumes that the inputs to the neuron have some degree of
randomness. This could be due to noise in the biological system or
modeling uncertainty in the input data.
Stochastic neuron model
October 13, 2023
3. Stochastic Spike Timing Model: In this model, the timing of spikes or
action potentials in the neuron is subject to randomness. This can be
used to mimic the variability in the timing of neural firing, which is
observed in real neural systems.
4. Stochastic Resonance Model: Stochastic resonance is a phenomenon
where the addition of noise to a nonlinear system can actually
enhance the system's response to weak input signals. Stochastic
resonance models can capture this behavior in neurons.
Stochastic neuron model
October 13, 2023
Stochastic neuron models are used in computational neuroscience to
study how neural systems might deal with inherent noise or how they
might perform probabilistic computations. They are also used in machine
learning and artificial neural networks for tasks that require modeling
uncertainty, such as reinforcement learning or certain types of generative
models.
One well-known example of a stochastic neuron model is the
Integrate-and-Fire (I&F) neuron model with added noise. This model
integrates incoming signals and generates spikes when the membrane
potential crosses a randomly fluctuating threshold.
The choice of a stochastic neuron model depends on the specific
research or application context and the level of biological realism or
mathematical abstraction required for the task at hand.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Continuous neuron model
October 13, 2023
Continuous neuron model
October 13, 2023
A continuous neuron model, also known as a continuous-time neuron
model, is a computational representation of a biological neuron that
operates in continuous time, as opposed to discrete time models, which
operate with discrete time steps.
Continuous neuron models are often used in computational
neuroscience and neural network research to provide a more precise and
biologically realistic description of how neurons process information.
Continuous neuron model
October 13, 2023
Key features of continuous neuron models include:
1. Differential Equations: Continuous neuron models are described using
differential equations to model the continuous dynamics of the
neuron's membrane potential and how it changes over time. These
equations can capture the flow of ionic currents across the neuron's
membrane and other biophysical processes.
2. Continuous Signals: In continuous neuron models, the input signals,
synaptic currents, and the neuron's membrane potential are all
represented as continuous functions of time. This allows for a more
detailed and accurate description of the neuron's behavior.
Continuous neuron model
October 13, 2023
Key features of continuous neuron models include:
3. Realistic Time Constants: Continuous neuron models often incorporate
realistic time constants, such as the membrane time constant, synaptic
time constants, and refractory period, which govern the temporal
dynamics of the neuron's response to input.
4. Complex Biophysical Details: Continuous models can capture the
complex biophysical processes occurring in a real neuron, such as the
interplay of ion channels, synaptic transmission, and spike generation.
Continuous neuron model
October 13, 2023
Key features of continuous neuron models include:
One of the most well-known continuous neuron models is the Hodgkin-Huxley
model, which describes the behavior of the squid giant axon neuron. This model
includes a system of differential equations that simulate the flow of sodium and
potassium ions across the neuron's membrane, resulting in action potentials.
Continuous neuron models are computationally more demanding than discrete
time models but are essential for understanding the precise temporal dynamics of
neural processes. They are often used in detailed simulations of single neurons and
networks of neurons to study phenomena like spike timing, synaptic plasticity, and
the propagation of signals in neural circuits.
While continuous neuron models provide a high level of biological realism, they
can be challenging to work with and may require specialized simulation techniques,
such as numerical integration methods, to solve the differential equations and capture
the neuron's behavior accurately.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Analog neuron model
October 13, 2023
Analog neuron model
October 13, 2023
An analog neuron model, also known as an analog neural network or analog
neural model, is a type of computational model inspired by biological neurons that
operates in a continuous, analog fashion.
Unlike digital neural networks, which use discrete numerical values and discrete
operations, analog neural models process information using continuous signals and
continuous operations, often making use of analog electronic components.
Analog neuron model
October 13, 2023
Key characteristics of analog neuron models include:
1. Continuous Signal Processing: In analog neuron models, information is
represented as continuous analog signals, typically in the form of continuous
voltages or currents. These signals can vary over a continuous range.
2. Analog Circuitry: Analog neuron models are often implemented using analog
electronic circuits, such as operational amplifiers (op-amps), resistors, capacitors,
and other analog components. These circuits can mimic the behavior of biological
neurons and synapses.
3. Complex Biophysical Modeling: Analog models may incorporate detailed
biophysical processes, such as the behavior of ion channels and the dynamics of
membrane potentials, to closely emulate the functioning of biological neurons.
Analog neuron model
October 13, 2023
Key characteristics of analog neuron models include:
4. Real-Time Processing: Analog neural networks can perform real-time signal
processing, making them suitable for applications like signal filtering, pattern
recognition, and control systems.
5. Energy Efficiency: Analog circuits are often more energy-efficient than their digital
counterparts, which is a significant advantage in applications where power
consumption is a concern.
Analog neuron model
October 13, 2023
Key characteristics of analog neuron models include:
Analog neural models have been used in various domains, including neurobiology
research, neuromorphic engineering, and specialized analog hardware for AI and
machine learning applications. These models are particularly well-suited for tasks
where continuous and real-time signal processing is essential, such as speech
recognition and sensor data analysis.
One example of analog neural networks is the use of neuromorphic chips, which
are specialized hardware designed to mimic the behavior of biological neurons and
synapses. These chips can be used in applications like robotics and sensor data
processing.
It's important to note that analog neural models have some limitations, such as
the challenges associated with scaling and configurability compared to digital neural
networks. Additionally, these models may be less suitable for complex digital tasks or
tasks that require high precision and reproducibility.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Implementation of neural networks
October 13, 2023
Implementation of neural networks
October 13, 2023
Implementing neural networks can be done using various
programming languages and deep learning libraries.
One of the most popular libraries for building neural networks is
TensorFlow, developed by Google, along with its high-level API, Keras.
Exercise for class
Program a neural network from scratch in a language
preferred by you
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Applications of artificial neural networks
October 13, 2023
Applications of artificial neural networks
October 13, 2023
Artificial Neural Networks (ANNs) have a wide range of applications across
various domains due to their ability to learn complex patterns and make predictions
from data. Here are some of the most common and notable applications of artificial
neural networks:
Image and Video Recognition:
 Object recognition: ANNs are used in computer vision applications to identify and
classify objects within images or videos.
 Facial recognition: ANNs can be trained to recognize and verify faces for security
and authentication purposes.
 Optical character recognition (OCR): ANNs help convert printed or handwritten
text into machine-readable text.
Applications of artificial neural networks
October 13, 2023
Natural Language Processing (NLP):
 Language translation: ANNs, especially recurrent neural networks (RNNs) and
transformers, are used in machine translation applications.
 Sentiment analysis: ANNs are employed to analyze and classify text sentiment,
which is useful in social media monitoring and customer feedback analysis.
 Chatbots and virtual assistants: Many chatbots and virtual assistants utilize ANNs
for natural language understanding and generation.
Speech Recognition:
 Speech-to-text conversion: ANNs, such as deep neural networks (DNNs) and
convolutional neural networks (CNNs), are used in applications like voice assistants
and transcription services.
 Speaker identification: ANNs can recognize and verify individual speakers based on
their vocal characteristics.
Applications of artificial neural networks
October 13, 2023
Medical Diagnosis and Healthcare:
 Disease diagnosis: ANNs assist in diagnosing diseases based on medical images
(e.g., X-rays, MRI scans) and patient data.
 Drug discovery: ANNs are used in drug design and prediction of chemical
properties.
 Predictive analytics: ANNs are employed for forecasting patient outcomes and
hospital resource management.
Autonomous Vehicles:
 Self-driving cars: ANNs are a crucial component in autonomous vehicles, helping
with object detection, path planning, and decision-making.
Applications of artificial neural networks
October 13, 2023
Financial Services:
 Stock market prediction: ANNs are used to analyze financial data and make
predictions about stock prices.
 Credit risk assessment: ANNs help financial institutions assess the creditworthiness
of individuals and businesses.
 Algorithmic trading: ANNs play a role in high-frequency trading and automated
trading strategies.
Manufacturing and Quality Control:
 Predictive maintenance: ANNs are used to predict equipment failures and reduce
downtime in manufacturing facilities.
 Quality control: ANNs help in detecting defects in products on production lines.
Applications of artificial neural networks
October 13, 2023
Gaming and Entertainment:
 Game AI: ANNs are used to create non-player characters (NPCs) with realistic
behaviors in video games.
 Content recommendation: Streaming services use ANNs to recommend content to
users based on their viewing history.
Environmental Monitoring:
 Weather forecasting: ANNs help meteorologists make more accurate weather
predictions.
 Air quality monitoring: ANNs analyze data from sensors to predict air quality and
pollution levels.
Applications of artificial neural networks
October 13, 2023
Cybersecurity:
 Anomaly detection: ANNs can identify unusual patterns in network traffic or
system behavior to detect cyber threats.
 User authentication: ANNs are used in biometric authentication systems, such as
fingerprint or facial recognition.
These are just a few examples of the many applications of artificial neural
networks. ANNs continue to advance and find new applications as the field of deep
learning evolves. Their ability to learn from data and adapt to different tasks makes
them a versatile tool for solving complex problems in various domains.
 • Introduction
 • Structure of an artificial neural system
 • Artificial neuron model
 • Dynamic neuron model
 • Stochastic neuron model
 • Continuous neuron model
 • Analog neuron model
 • Implementation of neural networks
 • Applications of artificial neural networks
 • Development of an application of artificial neural networks
5-9. REACTIVE SYSTEMS – NEURAL NETWORKS
October 13, 2023
Development of an application of artificial neural networks
October 13, 2023
Learning activity for class
Starting from a conceptual map, design the development of an
application of artificial neural networks in a field of knowledge that is
privileged by you.
• • "Artificial Intelligence: A Modern Approach" de Stuart Russell y Peter Norvig
(2021). Este es uno de los libros de referencia sobre inteligencia artificial. Ofrece
una amplia cobertura de los conceptos y técnicas clave.
• • "Sistemas Expertos: Principios y Programación" de Giarratano y Riley. (2021). Un
libro que se enfoca en la programación de sistemas expertos y proporciona
ejemplos prácticos.
• • "Deep Learning" de Ian Goodfellow, Yoshua Bengio y Aaron Courville. (2021).
Este libro es una referencia esencial para el aprendizaje profundo.
• • "Artificial Intelligence: A Guide to Intelligent Systems" de Michael Negnevitsky.
(2021). Ofrece una visión general de la inteligencia artificial y sistemas expertos,
incluyendo conceptos teóricos y aplicaciones prácticas.
• • "AI: A Very Short Introduction" de Margaret A. Boden. (2021). Un libro corto pero
informativo que proporciona una introducción accesible a la inteligencia artificial.
• • "Python Machine Learning" de Sebastian Raschka y Vahid Mirjalili. (2021). Para
aprender a aplicar técnicas de aprendizaje automático en Python.
• • "Building Machine Learning Powered Applications" de Emmanuel Ameisen.
(2021). Se enfoca en cómo construir aplicaciones prácticas basadas en machine
learning e inteligencia artificial.
• • "Sistemas Expertos: Teoría y Práctica" de José A. Olivas Varela. (2021). Un libro
en español que aborda los sistemas expertos desde una perspectiva teórica y
práctica.
• • Fernando Jorge Penousal (Eds.). Inteligencia Artificial y Computación Avanzada.
Fundación Alfredo Brañas Santiago de Compostela. (2021).
Bibliographic references
October 13, 2023
67
Thank you!
Comments & Questions
October 13, 2023
Route of a graph: by width and by depth
October 13, 2023
Route of a graph: by width and by depth

More Related Content

Similar to EXPERT SYSTEMS AND ARTIFICIAL INTELLIGENCE_ Neural Networks.pptx

What are neural networks.pdf
What are neural networks.pdfWhat are neural networks.pdf
What are neural networks.pdfAnastasiaSteele10
 
What are neural networks.pdf
What are neural networks.pdfWhat are neural networks.pdf
What are neural networks.pdfStephenAmell4
 
Nature Inspired Reasoning Applied in Semantic Web
Nature Inspired Reasoning Applied in Semantic WebNature Inspired Reasoning Applied in Semantic Web
Nature Inspired Reasoning Applied in Semantic Webguestecf0af
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its applicationHưng Đặng
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its applicationHưng Đặng
 
Artificial neural network for machine learning
Artificial neural network for machine learningArtificial neural network for machine learning
Artificial neural network for machine learninggrinu
 
CCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdf
CCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdfCCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdf
CCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdfAsst.prof M.Gokilavani
 
Basics of Artificial Neural Network
Basics of Artificial Neural Network Basics of Artificial Neural Network
Basics of Artificial Neural Network Subham Preetam
 
Artificial Neural Network: A brief study
Artificial Neural Network: A brief studyArtificial Neural Network: A brief study
Artificial Neural Network: A brief studyIRJET Journal
 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural NetworkManasa Mona
 
Artificial Neural Networks.pdf
Artificial Neural Networks.pdfArtificial Neural Networks.pdf
Artificial Neural Networks.pdfBria Davis
 
Neural networks of artificial intelligence
Neural networks of artificial  intelligenceNeural networks of artificial  intelligence
Neural networks of artificial intelligencealldesign
 
Neural networks are parallel computing devices.docx.pdf
Neural networks are parallel computing devices.docx.pdfNeural networks are parallel computing devices.docx.pdf
Neural networks are parallel computing devices.docx.pdfneelamsanjeevkumar
 
IRJET-AI Neural Network Disaster Recovery Cloud Operations Systems
IRJET-AI Neural Network Disaster Recovery Cloud Operations SystemsIRJET-AI Neural Network Disaster Recovery Cloud Operations Systems
IRJET-AI Neural Network Disaster Recovery Cloud Operations SystemsIRJET Journal
 
Neural networks in business forecasting
Neural networks in business forecastingNeural networks in business forecasting
Neural networks in business forecastingAmir Shokri
 

Similar to EXPERT SYSTEMS AND ARTIFICIAL INTELLIGENCE_ Neural Networks.pptx (20)

What are neural networks.pdf
What are neural networks.pdfWhat are neural networks.pdf
What are neural networks.pdf
 
What are neural networks.pdf
What are neural networks.pdfWhat are neural networks.pdf
What are neural networks.pdf
 
B42010712
B42010712B42010712
B42010712
 
Nature Inspired Reasoning Applied in Semantic Web
Nature Inspired Reasoning Applied in Semantic WebNature Inspired Reasoning Applied in Semantic Web
Nature Inspired Reasoning Applied in Semantic Web
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its application
 
Artificial neural networks and its application
Artificial neural networks and its applicationArtificial neural networks and its application
Artificial neural networks and its application
 
Artificial neural network for machine learning
Artificial neural network for machine learningArtificial neural network for machine learning
Artificial neural network for machine learning
 
Neural Network
Neural NetworkNeural Network
Neural Network
 
CCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdf
CCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdfCCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdf
CCS355 Neural Networks & Deep Learning Unit 1 PDF notes with Question bank .pdf
 
ANN - UNIT 1.pptx
ANN - UNIT 1.pptxANN - UNIT 1.pptx
ANN - UNIT 1.pptx
 
Basics of Artificial Neural Network
Basics of Artificial Neural Network Basics of Artificial Neural Network
Basics of Artificial Neural Network
 
Artificial Neural Network: A brief study
Artificial Neural Network: A brief studyArtificial Neural Network: A brief study
Artificial Neural Network: A brief study
 
Lesson 37
Lesson 37Lesson 37
Lesson 37
 
AI Lesson 37
AI Lesson 37AI Lesson 37
AI Lesson 37
 
Artificial Neural Network
Artificial Neural NetworkArtificial Neural Network
Artificial Neural Network
 
Artificial Neural Networks.pdf
Artificial Neural Networks.pdfArtificial Neural Networks.pdf
Artificial Neural Networks.pdf
 
Neural networks of artificial intelligence
Neural networks of artificial  intelligenceNeural networks of artificial  intelligence
Neural networks of artificial intelligence
 
Neural networks are parallel computing devices.docx.pdf
Neural networks are parallel computing devices.docx.pdfNeural networks are parallel computing devices.docx.pdf
Neural networks are parallel computing devices.docx.pdf
 
IRJET-AI Neural Network Disaster Recovery Cloud Operations Systems
IRJET-AI Neural Network Disaster Recovery Cloud Operations SystemsIRJET-AI Neural Network Disaster Recovery Cloud Operations Systems
IRJET-AI Neural Network Disaster Recovery Cloud Operations Systems
 
Neural networks in business forecasting
Neural networks in business forecastingNeural networks in business forecasting
Neural networks in business forecasting
 

More from Javier Daza

2. DATABASE MODELING_Database Fundamentals.pptx
2. DATABASE MODELING_Database Fundamentals.pptx2. DATABASE MODELING_Database Fundamentals.pptx
2. DATABASE MODELING_Database Fundamentals.pptxJavier Daza
 
1.BASES DE DATOS- DATABASE MODELING_Overview.pptx
1.BASES DE DATOS- DATABASE MODELING_Overview.pptx1.BASES DE DATOS- DATABASE MODELING_Overview.pptx
1.BASES DE DATOS- DATABASE MODELING_Overview.pptxJavier Daza
 
3. augmented reality uninpahu 2019-1
3. augmented reality uninpahu 2019-13. augmented reality uninpahu 2019-1
3. augmented reality uninpahu 2019-1Javier Daza
 
1. laboratory presentation virtual reality uninpahu_2019-1
1. laboratory presentation virtual reality uninpahu_2019-11. laboratory presentation virtual reality uninpahu_2019-1
1. laboratory presentation virtual reality uninpahu_2019-1Javier Daza
 

More from Javier Daza (6)

2. DATABASE MODELING_Database Fundamentals.pptx
2. DATABASE MODELING_Database Fundamentals.pptx2. DATABASE MODELING_Database Fundamentals.pptx
2. DATABASE MODELING_Database Fundamentals.pptx
 
1.BASES DE DATOS- DATABASE MODELING_Overview.pptx
1.BASES DE DATOS- DATABASE MODELING_Overview.pptx1.BASES DE DATOS- DATABASE MODELING_Overview.pptx
1.BASES DE DATOS- DATABASE MODELING_Overview.pptx
 
3. augmented reality uninpahu 2019-1
3. augmented reality uninpahu 2019-13. augmented reality uninpahu 2019-1
3. augmented reality uninpahu 2019-1
 
1. laboratory presentation virtual reality uninpahu_2019-1
1. laboratory presentation virtual reality uninpahu_2019-11. laboratory presentation virtual reality uninpahu_2019-1
1. laboratory presentation virtual reality uninpahu_2019-1
 
La web 2 top
La web 2 topLa web 2 top
La web 2 top
 
La web 2 top
La web 2 topLa web 2 top
La web 2 top
 

Recently uploaded

"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitecturePixlogix Infotech
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 3652toLead Limited
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsMiki Katsuragi
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDGMarianaLemus7
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLScyllaDB
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Mattias Andersson
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
costume and set research powerpoint presentation
costume and set research powerpoint presentationcostume and set research powerpoint presentation
costume and set research powerpoint presentationphoebematthew05
 
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks..."LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...Fwdays
 
Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clashcharlottematthew16
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...shyamraj55
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024Lorenzo Miniero
 

Recently uploaded (20)

"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC Architecture
 
Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365Ensuring Technical Readiness For Copilot in Microsoft 365
Ensuring Technical Readiness For Copilot in Microsoft 365
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering Tips
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDG
 
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptxE-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQL
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
costume and set research powerpoint presentation
costume and set research powerpoint presentationcostume and set research powerpoint presentation
costume and set research powerpoint presentation
 
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks..."LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
"LLMs for Python Engineers: Advanced Data Analysis and Semantic Kernel",Oleks...
 
Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clash
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024SIP trunking in Janus @ Kamailio World 2024
SIP trunking in Janus @ Kamailio World 2024
 

EXPERT SYSTEMS AND ARTIFICIAL INTELLIGENCE_ Neural Networks.pptx

  • 1. Prepared by: Eng. Javier Daza Piragauta
  • 2. EXPERT SYSTEMS AND ARTIFICIAL INTELLIGENCE REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 3.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 4. REACTIVE SYSTEMS October 13, 2023 Reactive Systems are a class of software systems designed to react to events and changes in their environment in a timely and efficient manner. These systems are typically used in real-time and critical applications where quick and reliable responses are crucial. Key characteristics of reactive systems include: • Responsiveness: Reactive systems respond rapidly to input and events from their environment, minimizing latency. • Scalability: They can scale horizontally or vertically to handle varying workloads. • Resilience: These systems are resilient to failures and can recover from unexpected situations. • Actor Model: Reactive systems often rely on the actor model, where independent components (actors) interact through message passing. Reactive systems find applications in a variety of domains, including online trading systems, real-time gaming systems, flight controllers, and infrastructure monitoring.
  • 5. NEURAL NETWORKS October 13, 2023  Neural Networks: Neural Networks are a type of artificial intelligence model inspired by the functioning of the human brain. They consist of layers of interconnected artificial neurons that process input data, perform computations, and generate outputs. Neural networks are used for machine learning and data processing tasks, including: • Supervised Learning: Neural networks can learn from labeled data, making them suitable for tasks like image classification, text translation, or numeric value prediction. • Unsupervised Learning: They are also used for tasks such as clustering, dimensionality reduction, and pattern detection in unlabeled data. • Deep Learning: Deep Neural Networks (DNNs) are a subtype of neural networks with multiple hidden layers that can handle highly complex tasks like image processing and natural language understanding. Neural networks have revolutionized many fields of computer science and artificial intelligence, including speech recognition, computer vision, autonomous driving, and online content recommendation. In summary, Reactive Systems focus on rapid and efficient reaction to events in real-time applications, while Neural Networks are artificial intelligence models capable of learning and performing complex tasks based on data processing and are widely used in the machine learning domain. Both concepts have significant applications and play crucial roles in different areas of technology and science.
  • 6.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 7. Structure of an artificial neural system October 13, 2023 Neural Networks - I Neural Networks - II
  • 8. Structure of an artificial neural system October 13, 2023
  • 9. Structure of an artificial neural system October 13, 2023
  • 10. Structure of an artificial neural system October 13, 2023
  • 11. Structure of an artificial neural system October 13, 2023
  • 12. Structure of an artificial neural system October 13, 2023
  • 13. Structure of an artificial neural system October 13, 2023
  • 14. Structure of an artificial neural system October 13, 2023
  • 15. Structure of an artificial neural system October 13, 2023 An artificial neural system, also known as an artificial neural network (ANN) or simply a neural network, is a computational model inspired by the structure and function of biological neural networks in the human brain. The structure of an artificial neural system consists of several interconnected layers, each with its unique role in processing information. Here is a breakdown of the typical structure of an artificial neural network:
  • 16. Structure of an artificial neural system October 13, 2023 1. Input Layer: The input layer is the first layer of the neural network. It receives external data or features as input. Each neuron in the input layer represents a feature or attribute of the input data. The number of neurons in this layer is determined by the dimensionality of the input data. 2. Hidden Layers: Between the input and output layers, there may be one or more hidden layers. These layers are called "hidden" because they are not directly connected to the external data or the final output. The hidden layers perform complex computations and feature extraction. The number of hidden layers and neurons in each layer is determined by the architecture of the network, and it can vary widely.
  • 17. Structure of an artificial neural system October 13, 2023 3. Neurons (Nodes): Each layer consists of interconnected neurons, also known as nodes or units. Neurons in one layer are connected to neurons in the subsequent layer through weighted connections. Neurons process information by summing up the weighted inputs, applying an activation function, and passing the result to the next layer. 4. Weights and Connections: Connections between neurons are associated with weights. These weights determine the strength of the connection and play a crucial role in learning. During training, the network adjusts the weights to improve its performance in making predictions or classifications.
  • 18. Structure of an artificial neural system October 13, 2023 5. Activation Functions: Activation functions introduce non-linearity into the neural network. Common activation functions include the sigmoid function, the hyperbolic tangent (tanh) function, and the rectified linear unit (ReLU) function. The activation function defines the output of a neuron based on its weighted sum of inputs. 6. Output Layer: The output layer produces the final results of the neural network's computation. The number of neurons in the output layer depends on the specific task the neural network is designed for. For example, in a binary classification task, there may be one output neuron representing the probability of belonging to one class, while in multi-class classification, there would be multiple output neurons, each representing a different class.
  • 19. Structure of an artificial neural system October 13, 2023 7. Bias Neurons: Some neural networks include bias neurons in each layer. These neurons have a constant input value of 1 and are associated with their own weights. Bias neurons allow the network to account for systematic errors or biases in the data. 8. Training Algorithm: To make the neural network perform a specific task, it must be trained on labeled data. Training involves adjusting the weights of the connections to minimize a cost function or loss function that measures the difference between the network's predictions and the true values in the training data. Common training algorithms include backpropagation and variants like stochastic gradient descent (SGD) and Adam.
  • 20. Structure of an artificial neural system October 13, 2023 9. Output Activation Function: The choice of activation function in the output layer depends on the nature of the task. For regression tasks, a linear activation function may be used. For binary classification, a sigmoid or softmax function is common, and for multi-class classification, softmax is typically used. The structure and architecture of neural networks can vary significantly based on the specific problem they are designed to solve. Deep neural networks (DNNs) have multiple hidden layers and have proven effective in solving complex tasks, while simpler architectures like feedforward neural networks have fewer layers and are used for simpler problems. Convolutional Neural Networks (CNNs) are specialized for image-related tasks, and Recurrent Neural Networks (RNNs) are designed for sequential data processing. The choice of architecture and structure depends on the problem domain and the data characteristics.
  • 21.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 23. Artificial neuron model October 13, 2023 An artificial neuron, also known as a perceptron, is the fundamental building block of artificial neural networks. It is a simplified computational model inspired by the functioning of biological neurons in the human brain. The artificial neuron processes input data and produces an output based on a set of weighted connections, an activation function, and a bias term. Here's an overview of the components and operation of an artificial neuron:
  • 24. Artificial neuron model October 13, 2023 Components of an Artificial Neuron: 1. Inputs (X1, X2, ..., Xn): An artificial neuron receives input signals, often represented as X1, X2, ..., Xn. Each input represents a feature or attribute of the data being processed. 2. Weights (W1, W2, ..., Wn): Each input is associated with a weight, represented as W1, W2, ..., Wn. These weights determine the importance of each input in the neuron's computation. The weights are parameters that the neuron learns during training to optimize its performance. 3. Bias (b): The bias, denoted as "b," is an additional parameter. It allows the neuron to adjust its output independently of the inputs and is crucial for modeling translation invariance.
  • 25. Artificial neuron model October 13, 2023 Components of an Artificial Neuron: 4. Weighted Sum (Z): The weighted sum (Z) is calculated by multiplying each input by its corresponding weight and summing these products. It is expressed as: Z = (X1 * W1) + (X2 * W2) + ... + (Xn * Wn) + b 5. Activation Function (f): The weighted sum is passed through an activation function (f). The activation function introduces non-linearity into the model. Common activation functions include the sigmoid, hyperbolic tangent (tanh), and rectified linear unit (ReLU). The choice of activation function depends on the specific problem and the desired properties of the neuron.
  • 26. Artificial neuron model October 13, 2023 Components of an Artificial Neuron: 6. Output (Y): The output (Y) of the neuron is the result of applying the activation function to the weighted sum: Y = f(Z)
  • 27. Artificial neuron model October 13, 2023 Operation of an Artificial Neuron: 1. The artificial neuron receives input signals (X1, X2, ..., Xn). 2. Each input is multiplied by its corresponding weight (W1, W2, ..., Wn). 3. The weighted products are summed, and the bias term is added to the sum to calculate the weighted sum (Z). 4. The activation function (f) is applied to the weighted sum (Z) to produce the neuron's output (Y). .
  • 28. Artificial neuron model October 13, 2023 Operation of an Artificial Neuron: In essence, the artificial neuron processes inputs by adjusting the weights and bias during training to find the optimal parameters that minimize the error between its output and the desired output. Once trained, the neuron can be used to make predictions or perform various computations in the context of a neural network. Artificial neurons are typically organized into layers to form neural networks. In this network context, they can learn complex patterns and relationships in data, making them a fundamental component of machine learning and deep learning models.
  • 29.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 30. Dynamic neuron model October 13, 2023 Energy-Efficient Tuning of Spintronic Neurons to Imitate the Non-linear Oscillatory Neural Networks of the Human Brain
  • 31. Dynamic neuron model October 13, 2023 A "dynamic neuron model" typically refers to a computational model of a neuron that incorporates more complex and dynamic behavior than the simplified artificial neuron model, such as the integrate-and-fire model. Dynamic neuron models are used in computational neuroscience and neural network research to better capture the behavior of biological neurons. One example of a dynamic neuron model is the Hodgkin-Huxley model. Here's an overview of dynamic neuron models: Hodgkin-Huxley Model: The Hodgkin-Huxley model is a well-known dynamic neuron model that provides a more biologically accurate representation of how individual neurons operate. This model was developed by Sir Alan Hodgkin and Sir Andrew Huxley in 1952 and earned them the Nobel Prize in Physiology or Medicine in 1963. The model is commonly used to describe the behavior of neurons in the context of action potentials or spikes.
  • 32. Dynamic neuron model October 13, 2023 Key features of the Hodgkin-Huxley model: 1. Membrane Potential (Voltage): Unlike the simplified artificial neuron model, the Hodgkin-Huxley model accounts for the neuron's membrane potential and how it changes over time. 2. Ion Channels: The model incorporates different types of ion channels, including sodium (Na+), potassium (K+), and leak channels. These ion channels allow for the flow of ions in and out of the neuron, influencing the neuron's membrane potential. 3. Differential Equations: The dynamics of the model are described using a system of differential equations that govern the behavior of ion channels and the membrane potential.
  • 33. Dynamic neuron model October 13, 2023 Key features of the Hodgkin-Huxley model: 5. Action Potentials: The model accurately represents the generation and propagation of action potentials, which are the electrical signals that neurons use to communicate with each other. 6. Nonlinear Behavior: The Hodgkin-Huxley model exhibits nonlinear behavior, which is essential for modeling the complex dynamics of real neurons. The Hodgkin-Huxley model, and similar dynamic neuron models, can be used to simulate and understand the complex electrical activity in neurons, including the initiation and propagation of action potentials. These models are valuable tools in the study of neural circuits, neurophysiology, and the development of more realistic computational
  • 34. Dynamic neuron model October 13, 2023 Key features of the Hodgkin-Huxley model: The Hodgkin-Huxley model, and similar dynamic neuron models, can be used to simulate and understand the complex electrical activity in neurons, including the initiation and propagation of action potentials. These models are valuable tools in the study of neural circuits, neurophysiology, and the development of more realistic computational models of neural networks. Dynamic neuron models are essential in fields like computational neuroscience, where the focus is on understanding the intricacies of biological neurons and their interaction in the brain. While simple artificial neurons are used in machine learning and deep learning for practical purposes, dynamic neuron models are crucial for modeling the underlying biology of the nervous system.
  • 35.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 37. Stochastic neuron model October 13, 2023 A stochastic neuron model is a type of artificial neuron or computational model of a biological neuron that incorporates stochastic (random) elements into its operation. Unlike deterministic neuron models, which produce fixed outputs given fixed inputs, stochastic neuron models introduce randomness into the process, making them useful in modeling certain aspects of neural behavior that involve uncertainty or noise. There are various ways to incorporate stochasticity into neuron models, and it depends on the specific context and the level of abstraction you are interested in. Here are a few common approaches:
  • 38. Stochastic neuron model October 13, 2023 1. Stochastic Threshold Model: In this model, the neuron's firing threshold is not fixed but instead varies randomly. The neuron integrates incoming signals as usual, but whether it fires or not depends on a stochastic process. This can help capture the variability in neural firing observed in real biological neurons. 1. Stochastic Input Model: Instead of a fixed input signal, this model assumes that the inputs to the neuron have some degree of randomness. This could be due to noise in the biological system or modeling uncertainty in the input data.
  • 39. Stochastic neuron model October 13, 2023 3. Stochastic Spike Timing Model: In this model, the timing of spikes or action potentials in the neuron is subject to randomness. This can be used to mimic the variability in the timing of neural firing, which is observed in real neural systems. 4. Stochastic Resonance Model: Stochastic resonance is a phenomenon where the addition of noise to a nonlinear system can actually enhance the system's response to weak input signals. Stochastic resonance models can capture this behavior in neurons.
  • 40. Stochastic neuron model October 13, 2023 Stochastic neuron models are used in computational neuroscience to study how neural systems might deal with inherent noise or how they might perform probabilistic computations. They are also used in machine learning and artificial neural networks for tasks that require modeling uncertainty, such as reinforcement learning or certain types of generative models. One well-known example of a stochastic neuron model is the Integrate-and-Fire (I&F) neuron model with added noise. This model integrates incoming signals and generates spikes when the membrane potential crosses a randomly fluctuating threshold. The choice of a stochastic neuron model depends on the specific research or application context and the level of biological realism or mathematical abstraction required for the task at hand.
  • 41.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 43. Continuous neuron model October 13, 2023 A continuous neuron model, also known as a continuous-time neuron model, is a computational representation of a biological neuron that operates in continuous time, as opposed to discrete time models, which operate with discrete time steps. Continuous neuron models are often used in computational neuroscience and neural network research to provide a more precise and biologically realistic description of how neurons process information.
  • 44. Continuous neuron model October 13, 2023 Key features of continuous neuron models include: 1. Differential Equations: Continuous neuron models are described using differential equations to model the continuous dynamics of the neuron's membrane potential and how it changes over time. These equations can capture the flow of ionic currents across the neuron's membrane and other biophysical processes. 2. Continuous Signals: In continuous neuron models, the input signals, synaptic currents, and the neuron's membrane potential are all represented as continuous functions of time. This allows for a more detailed and accurate description of the neuron's behavior.
  • 45. Continuous neuron model October 13, 2023 Key features of continuous neuron models include: 3. Realistic Time Constants: Continuous neuron models often incorporate realistic time constants, such as the membrane time constant, synaptic time constants, and refractory period, which govern the temporal dynamics of the neuron's response to input. 4. Complex Biophysical Details: Continuous models can capture the complex biophysical processes occurring in a real neuron, such as the interplay of ion channels, synaptic transmission, and spike generation.
  • 46. Continuous neuron model October 13, 2023 Key features of continuous neuron models include: One of the most well-known continuous neuron models is the Hodgkin-Huxley model, which describes the behavior of the squid giant axon neuron. This model includes a system of differential equations that simulate the flow of sodium and potassium ions across the neuron's membrane, resulting in action potentials. Continuous neuron models are computationally more demanding than discrete time models but are essential for understanding the precise temporal dynamics of neural processes. They are often used in detailed simulations of single neurons and networks of neurons to study phenomena like spike timing, synaptic plasticity, and the propagation of signals in neural circuits. While continuous neuron models provide a high level of biological realism, they can be challenging to work with and may require specialized simulation techniques, such as numerical integration methods, to solve the differential equations and capture the neuron's behavior accurately.
  • 47.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 49. Analog neuron model October 13, 2023 An analog neuron model, also known as an analog neural network or analog neural model, is a type of computational model inspired by biological neurons that operates in a continuous, analog fashion. Unlike digital neural networks, which use discrete numerical values and discrete operations, analog neural models process information using continuous signals and continuous operations, often making use of analog electronic components.
  • 50. Analog neuron model October 13, 2023 Key characteristics of analog neuron models include: 1. Continuous Signal Processing: In analog neuron models, information is represented as continuous analog signals, typically in the form of continuous voltages or currents. These signals can vary over a continuous range. 2. Analog Circuitry: Analog neuron models are often implemented using analog electronic circuits, such as operational amplifiers (op-amps), resistors, capacitors, and other analog components. These circuits can mimic the behavior of biological neurons and synapses. 3. Complex Biophysical Modeling: Analog models may incorporate detailed biophysical processes, such as the behavior of ion channels and the dynamics of membrane potentials, to closely emulate the functioning of biological neurons.
  • 51. Analog neuron model October 13, 2023 Key characteristics of analog neuron models include: 4. Real-Time Processing: Analog neural networks can perform real-time signal processing, making them suitable for applications like signal filtering, pattern recognition, and control systems. 5. Energy Efficiency: Analog circuits are often more energy-efficient than their digital counterparts, which is a significant advantage in applications where power consumption is a concern.
  • 52. Analog neuron model October 13, 2023 Key characteristics of analog neuron models include: Analog neural models have been used in various domains, including neurobiology research, neuromorphic engineering, and specialized analog hardware for AI and machine learning applications. These models are particularly well-suited for tasks where continuous and real-time signal processing is essential, such as speech recognition and sensor data analysis. One example of analog neural networks is the use of neuromorphic chips, which are specialized hardware designed to mimic the behavior of biological neurons and synapses. These chips can be used in applications like robotics and sensor data processing. It's important to note that analog neural models have some limitations, such as the challenges associated with scaling and configurability compared to digital neural networks. Additionally, these models may be less suitable for complex digital tasks or tasks that require high precision and reproducibility.
  • 53.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 54. Implementation of neural networks October 13, 2023
  • 55. Implementation of neural networks October 13, 2023 Implementing neural networks can be done using various programming languages and deep learning libraries. One of the most popular libraries for building neural networks is TensorFlow, developed by Google, along with its high-level API, Keras. Exercise for class Program a neural network from scratch in a language preferred by you
  • 56.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 57. Applications of artificial neural networks October 13, 2023
  • 58. Applications of artificial neural networks October 13, 2023 Artificial Neural Networks (ANNs) have a wide range of applications across various domains due to their ability to learn complex patterns and make predictions from data. Here are some of the most common and notable applications of artificial neural networks: Image and Video Recognition:  Object recognition: ANNs are used in computer vision applications to identify and classify objects within images or videos.  Facial recognition: ANNs can be trained to recognize and verify faces for security and authentication purposes.  Optical character recognition (OCR): ANNs help convert printed or handwritten text into machine-readable text.
  • 59. Applications of artificial neural networks October 13, 2023 Natural Language Processing (NLP):  Language translation: ANNs, especially recurrent neural networks (RNNs) and transformers, are used in machine translation applications.  Sentiment analysis: ANNs are employed to analyze and classify text sentiment, which is useful in social media monitoring and customer feedback analysis.  Chatbots and virtual assistants: Many chatbots and virtual assistants utilize ANNs for natural language understanding and generation. Speech Recognition:  Speech-to-text conversion: ANNs, such as deep neural networks (DNNs) and convolutional neural networks (CNNs), are used in applications like voice assistants and transcription services.  Speaker identification: ANNs can recognize and verify individual speakers based on their vocal characteristics.
  • 60. Applications of artificial neural networks October 13, 2023 Medical Diagnosis and Healthcare:  Disease diagnosis: ANNs assist in diagnosing diseases based on medical images (e.g., X-rays, MRI scans) and patient data.  Drug discovery: ANNs are used in drug design and prediction of chemical properties.  Predictive analytics: ANNs are employed for forecasting patient outcomes and hospital resource management. Autonomous Vehicles:  Self-driving cars: ANNs are a crucial component in autonomous vehicles, helping with object detection, path planning, and decision-making.
  • 61. Applications of artificial neural networks October 13, 2023 Financial Services:  Stock market prediction: ANNs are used to analyze financial data and make predictions about stock prices.  Credit risk assessment: ANNs help financial institutions assess the creditworthiness of individuals and businesses.  Algorithmic trading: ANNs play a role in high-frequency trading and automated trading strategies. Manufacturing and Quality Control:  Predictive maintenance: ANNs are used to predict equipment failures and reduce downtime in manufacturing facilities.  Quality control: ANNs help in detecting defects in products on production lines.
  • 62. Applications of artificial neural networks October 13, 2023 Gaming and Entertainment:  Game AI: ANNs are used to create non-player characters (NPCs) with realistic behaviors in video games.  Content recommendation: Streaming services use ANNs to recommend content to users based on their viewing history. Environmental Monitoring:  Weather forecasting: ANNs help meteorologists make more accurate weather predictions.  Air quality monitoring: ANNs analyze data from sensors to predict air quality and pollution levels.
  • 63. Applications of artificial neural networks October 13, 2023 Cybersecurity:  Anomaly detection: ANNs can identify unusual patterns in network traffic or system behavior to detect cyber threats.  User authentication: ANNs are used in biometric authentication systems, such as fingerprint or facial recognition. These are just a few examples of the many applications of artificial neural networks. ANNs continue to advance and find new applications as the field of deep learning evolves. Their ability to learn from data and adapt to different tasks makes them a versatile tool for solving complex problems in various domains.
  • 64.  • Introduction  • Structure of an artificial neural system  • Artificial neuron model  • Dynamic neuron model  • Stochastic neuron model  • Continuous neuron model  • Analog neuron model  • Implementation of neural networks  • Applications of artificial neural networks  • Development of an application of artificial neural networks 5-9. REACTIVE SYSTEMS – NEURAL NETWORKS October 13, 2023
  • 65. Development of an application of artificial neural networks October 13, 2023 Learning activity for class Starting from a conceptual map, design the development of an application of artificial neural networks in a field of knowledge that is privileged by you.
  • 66. • • "Artificial Intelligence: A Modern Approach" de Stuart Russell y Peter Norvig (2021). Este es uno de los libros de referencia sobre inteligencia artificial. Ofrece una amplia cobertura de los conceptos y técnicas clave. • • "Sistemas Expertos: Principios y Programación" de Giarratano y Riley. (2021). Un libro que se enfoca en la programación de sistemas expertos y proporciona ejemplos prácticos. • • "Deep Learning" de Ian Goodfellow, Yoshua Bengio y Aaron Courville. (2021). Este libro es una referencia esencial para el aprendizaje profundo. • • "Artificial Intelligence: A Guide to Intelligent Systems" de Michael Negnevitsky. (2021). Ofrece una visión general de la inteligencia artificial y sistemas expertos, incluyendo conceptos teóricos y aplicaciones prácticas. • • "AI: A Very Short Introduction" de Margaret A. Boden. (2021). Un libro corto pero informativo que proporciona una introducción accesible a la inteligencia artificial. • • "Python Machine Learning" de Sebastian Raschka y Vahid Mirjalili. (2021). Para aprender a aplicar técnicas de aprendizaje automático en Python. • • "Building Machine Learning Powered Applications" de Emmanuel Ameisen. (2021). Se enfoca en cómo construir aplicaciones prácticas basadas en machine learning e inteligencia artificial. • • "Sistemas Expertos: Teoría y Práctica" de José A. Olivas Varela. (2021). Un libro en español que aborda los sistemas expertos desde una perspectiva teórica y práctica. • • Fernando Jorge Penousal (Eds.). Inteligencia Artificial y Computación Avanzada. Fundación Alfredo Brañas Santiago de Compostela. (2021). Bibliographic references October 13, 2023
  • 67. 67 Thank you! Comments & Questions October 13, 2023
  • 68. Route of a graph: by width and by depth October 13, 2023 Route of a graph: by width and by depth