Artificial intelligence
what is AI?
History
foundations of AI
Types of AI
Applications of AI
machine learning and applications
AI Vs Machine learning
Deep learning- advantages and disadvantages
Applications of Deep learning
Why is deep learning better than machine learning
Deep learning vs machine learning
Artificial Neural Network (ANN)
Architecture of ANN
Types of ANN
Applications of ANN
Softwares of ANN and their applications
2. CONTENTS
• ARTIFICIAL INTELLIGENCE
• History
• Foundation
• Types
• Applications
• MACHINE LEARNING
• Types and Applications
• AI vs Machine learning
• DEEP LEARNING
• Advantages & Disadvantages
• Applications
• Machine learning vs Deep learning
• Neural Network (ANN)
• Architecture
• Working
• Types
• Applications
• References
3. ARTIFICAL INTELLIGENCE
Artificial intelligence (AI) refers to the simulation of human
intelligence in machines that are programmed to think like
humans and mimic their actions.The term may also be applied
to any machine that exhibits traits associated with a human
mind such as learning and problem-solving.
A subset of artificial intelligence is Machine learning, which
refers to the concept that computer programs can
automatically learn from and adapt to new data without
being assisted by humans.
Deep learning enable this automatic learning through the
absorption of huge amounts of unstructured data such as
text, images, or video with the help of Neural networks.
4. HISTORY
• 1943-BinaryANN model,W. McCulloch andW. Pitts – the first
nonlinear mathematical model of the neuron (a formal neuron).
• 1950-AlanTuring, Published a landmark paper "Computing Machinery
and Intelligence” in which he speculated about the possibility of
creating machines that think humanly.
• 1951-Game AI, Christopher Strachey wrote a checkers program and
Dietrich Prinz wrote for chess.
• 1956- Birth of AI, John McCarthy first coined the termArtificial
Intelligence at Dartmouth Conference.
• 1959- first AI Laboratory, MIT AI Lab, (Massachusetts Institute of
Technology, Cambridge, MA, United States) where the research on AI
began.
AlanTuring
John McCarthy
5. 1960- General Motors Robot , First robot was introduced to
General Motors assembly line.
1961- First AI Chatbot (ELIZA) was introduced.
1997- IBM Deep Blue beats the world champion Garry Kasparov in
the game of chess.
2005- DARPA Grand Challenge
Stanford Racing team’s autonomous robotic car named Stanley
wins the 2005 DARPA Grand Challenge.
2011- IBMWatson
IBM’s Question answering system namedWatson defeated the
two greatest Jeopardy champions, Brad Rutter and Ken Jennings.
8. • Narrow Artificial Intelligence: Sometimes referred to as “Weak AI” , this kind of AI
operates within a limited context and is a simulation of human intelligence. It is
often focused on performing a single task extremely well and while these machines
may seem intelligent, they are operating under far more constraints and limitations
than even the most basic human intelligence.
Examples: Google search, Image recognition software, Siri, Alexa and other
personal assistants, Self-driving cars.
• Artificial General Intelligence (AGI): Sometimes referred to as “Strong AI” is the
kind ofAI that can understand and learn any intellectual task that a human being
can perform. It allows a machine to apply knowledge and skills in different contexts.
• Super Artificial Intelligence: It surpasses human intelligence and can perform any
task better than a human.The concept sees AI evolved to be so analogous to human
sentiments and experiences that it doesn't merely understand them.It also evokes
emotions, needs, beliefs, and desires of its own. Its existence is still hypothetical.
9. Robotic vehicles: A driverless robotic car named STANLEY
sped through the rough terrain of the Mojave dessert at 22
mph, finishing the 132-mile course first to win the 2005
DARPA Grand Challenge. STANLEY is aVolkswagen
outfitted with cameras, radar, and laser rangefinders to
sense the environment and onboard software to command
the steering, braking and acceleration.
APPLICATIONS
Autonomous planning and scheduling: A hundred million miles from Earth, NASA’s Remote
Agent program became the first on-board autonomous planning program to control the
scheduling of operations for a spacecraft. Remote agent generated plans from high-level goals
specified from the ground and monitored the execution of those plans—detecting, diagnosing,
and recovering from problems as they occurred. MEXAR2 did mission planning—both logistics
and science planning—for the European Space Agency’s Mars Express mission in 2008.
10. MachineTranslation: A computer program automatically translates from Arabic to English,
allowing an English speaker to see the headline “Ardogan ConfirmsThatTurkey Would Not
Accept Any Pressure, UrgingThem to Recognize Cyprus.”The program uses a statistical model
built from examples of Arabic-to-English translations and from examples of English text totaling
two trillion words.
Spam fighting: Each day, learning algorithms classify over a billion messages as spam, saving the
recipient from wasting time deleting , many users may comprise 80% or 90% of all messages, if
not classified away by algorithms. Because the spammers are continually updating their tactics, it
is difficult for a static programmed approach to keep up, and learning algorithms work best.
Robotics: The iRobot Corporation has sold over two million Roomba
robotic vacuum cleaners for home use.The company also deploys the
rugged PackBot to Iraq and Afghanistan, where it is used to handle
hazardous materials, clear explosives, and identify the snipers
location.
11. MACHINE LEARNING
• Machine learning is a branch of artificial intelligence (AI) that allows machines to learn
through algorithms.These algorithms learn from real data with which a model is generated.
This model allows predicting what class or what type is a new data.
• it utilizes automated algorithms to predict the decisions for the future and modeling of
functions based on the data fed to it. Ex-multiple linear regression, random forest, and
K-NN (k-nearest neighbor algorithm).
• It is of three types:
Supervised Learning –We have the labeled/classified data to train the machines.
Unsupervised Learning –We do not have labeled/classified data to train the machines.
Reinforcement Learning –We train the machines through rewards on the right decisions.
12.
13. Applications of Machine Learning
Pinterest – Improved Content Discovery
• Pinterest occupies a curious place in the social media
ecosystem. As it’s primary function is to curate existing
content, it is beneficial to invest in technologies that can make
this process more effective.
• Thus In 2015, Pinterest acquired Kosei, a machine learning
company that specialized in the commercial applications of
machine learning (content discovery and recommendation
algorithms).
• By today, machine learning touches every aspect of Pinterest’s
business operations, from spam moderation and content
discovery to advertising monetization and reducing churn of
email subscribers.
14. Facebook – Chatbot
• Facebook’s Messenger service is one of the most exciting
aspects of the world’s largest social media platform.That’s
because Messenger has become something of an
experimental testing laboratory for chatbots.
• Some chatbots are virtually indistinguishable from humans
when conversing via text.
• Any developer can create and submit a chatbot for
inclusion in Facebook Messenger.This means that
companies with a strong emphasis on customer service
and retention can leverage chatbots, even if they’re a tiny
startup with limited engineering resources.
• It not the only application of machine learning but also to
filter out spam and poor-quality content, and the company
is also researching computer vision algorithms that can
“read” images to visually impaired people.
15. ARTIFICIAL INTELLIGENCE (AI) MACHINE LEARNING (ML)
AI enables machines to simulate human
intelligence/ behavior.
ML enables machines to automatically learn from
humongous past data without being explicitly
programmed.
AI aims to make smart and intelligent systems
that function just like humans so that they can
solve complex problems.
ML aims to enable machines to learn from the
available data so that they can give accurate output.
Subsets of AI – machine learning and deep
learning.
Subset of machine learning – deep learning.
AI comprises the processes of learning,
reasoning, and self-correction.
ML involves learning and self-correction only when it
is introduced to new data.
AI can deal with structured, unstructured, and
semi-structured data.
Machine learning can only deal with structured and
semi-structured data.
AI is working towards the creation of systems
that will be able to perform various complex
tasks.
ML is working towards creating machines that will be
able to perform specific tasks – for which they are
trained.
Applications of AI – Siri, online games, etc. Applications of ML – Google search algorithms,
Recommended products list on Amazon, etc.
16. DEEP LEARNING
• Deep learning is the subset of machine learning that imitates the functionality of a human
brain for managing the data and forming the patterns for referring it in decision making.
• In deep learning, a computer algorithm learns to perform classification tasks directly on
complex data in the form of images, text, or sound.
• These algorithms can accomplish state-of-the-art (SOTA) accuracy, and even sometimes
surpassing human-level performance.They are trained with the large set of labeled data
with neural network architectures, involving many layers.
• The trained dataset can be interconnected, diverse and complex in nature.
• Neural networks help in interpreting the features of data and their relationships in which
important information is processed through multiple stages.
17. ADVANTAGES
• It has ability to generate new features
from the limited available training data
sets.
• It can work on unsupervised learning
techniques helps in generating actionable
and reliable task outcomes.
• It reduces the time required for feature
engineering, one of the tasks that
requires major time in practicing machine
learning.
• With continuous training, its architecture
has become adaptive to change and is
able to work on diverse problems.
DISADVANTAGES
• The complete training process relies on
the continuous flow of the data, which
decreases the scope for improvement in
the training process.
• The cost of computational training
significantly increases with an increase in
the number of datasets.
• Lack of transparency in fault revision. No
intermediate steps to provide the
arguments for a certain fault. In order to
resolve the issue, a complete algorithm
gets revised.
19. WHY IS DEEP LEARNING BETTERTHAN MACHINE LEARNING?
• The main advantage of deep learning over machine learning is the redundancy of
feature extraction.
• Long before deep learning, traditional machine learning methods were most popular.
which are otherwise known as flat algorithms. It means these algorithms cannot
typically be applied directly to raw data (such as .csv, images, text, etc.). Instead we
require a preprocessing step called feature extraction.
• Feature extraction is complicated and requires detailed knowledge of the problem
domain.This step must be adapted, tested and refined over several iterations for
optimal results.
• When it comes to deep learning models, we have artificial neural networks, which don’t
require feature extraction.The layers are able to learn an implicit representation of the
raw data on their own.
20. Machine Learning Deep Learning
Machine Learning is a superset of Deep
Learning
Deep Learning is a subset of Machine
Learning
it uses structured data it uses neural networks(ANN).
Outputs: Numerical Value, like
classification of score
Anything from numerical values to free-
form elements, such as free text and sound.
Uses various types of automated
algorithms that turn to model functions
and predict future action from data
Uses neural network that passes data
through processing layers to, interpret data
features and relations.
Humans explicitly do feature engineering. Feature engineering is not needed because
important features are automatically
detected by neural networks
Machine Learning is highly used to stay in
the competition and learn new things.
Deep Learning solves complex machine
learning issues.
21. Artificial Neural Networks (ANN)
• Neural networks, also known as artificial neural networks (ANNs) or simulated
neural networks (SNNs), are a subset of machine learning and are at the heart
of deep learning algorithms.Their name and structure are inspired by the human
brain, mimicking the way that biological neurons signal to one another.
• The uniqueness of ANN is its ability to deliver desirable results even with the help of
incomplete or historical data results without a need for structured experimental
design by modeling and pattern recognition.
22. ADVANTAGES
1) It has a parallel processing ability. It has the numerical strength that performs more than one
task at the same time.
2) Failure of one element of the network does not affect the working of the whole system.This
characteristic makes it fault-tolerant.
3) A neural network learns from the experience and does not need reprogramming.
DISADVANTAGES
1. Its black-box nature is the most prominent disadvantage.The neural network does not give the
proper explanation of determining the output. It reduces trust in the network.
2.The duration of the development of the network is unknown.
3.There is no assurance of proper network structure.There is no proper rule to determine the
structure.
24. • Similarly to the brain, ANN is composed of many single units, artificial neurons,
associated with coefficients (weights) which constitute the structural neuron ,
otherwise called processing elements (PE) as they process data.
• A neural network consists of three layers.
• The first layer is the input layer. It contains the input neurons that send information
to the hidden layer.
• The hidden layer performs the computations on input data and transfers the
output to the output layer. It includes weight, activation function, cost function.
• To standardize the output from the neuron, the “activation function” is used.
Activation functions are the mathematical equations that calculate the output of
the neural network. Each neuron has its own activation function.
25. A schematic of four-layered artificial network. Input layer units (in blue) receive input signals (x1, x2,
x3) and transfer the signal to the hidden layers via weighted connections. Output layer receives the
signals and provides the representative output signal.
26. In each hidden layer and output layer, the processing unit sums up its input from previous layer by the
sigmoidal function to compute the output to the following layer according to the following equations
yq = ∑ wpq xp
f(yq) =
• Where, wpq is the strength of the connections between unit q in the current layer to unit p in the
previous layer, xp is the output value from the previous layer,
• f(yq) is conducted to the following layer as an output value, and α is a parameter relating to the shape of
the sigmoidal function.
• Most commonly used methods is back-propagation method, which requires two simple steps,
1.Network Design 2.LearningOrTraining
1
1+ exp(- αyq )
WORKING
27. A common design of a node in an artificial neural network. The input signal travels to the
unit though weighted connection. Multiple signals are summated, processed, and
transformed based on the specific function. An output signal is relayed to the following
nodes in the network.
1. Network Design:
28. 2.Training:
• The process of training involves a search for the most optimal network state by
adjustments of the weights of the connections between PEs.The weights are
adjusted, based on training the set of data, until the error is minimized.
• The amount of training is important because undertrained ANNs yields larger
errors in the output signal and overtrained ANNs lose the ability to generalize
and recognize patterns.
• The learning through weight adjustment can be supervised or unsupervised.
• The network is repeatedly presented with an input pattern and a desired output
response in supervised learning.
• The training process terminates when error goal is near zero and neural
network produces correct response for given input patterns.
29. • Akaike’s information criterion (AIC) can be applied to evaluate the optimality ofANN
Equation:
Nhidden =
Ntrn
[R+(Ninp +NOUT)]
• Where, Nhidden is the number of hidden nodes; Ntrn is the number of training sample; R
is a constant with values ranging from 5 to 10, Ninp is the number of inputs and NOUT is
the number of outputs.
• The network uses the cost function to compare the output and expected output. Cost
function refers to the difference between the actual value and the predicted value.
Lower the cost function, closer it is to the desired output.
30. There are two processes for minimizing the cost function.
1) BACK PROPAGATION
Back propagation is the core of neural network training. Data enters the input layer and
propagates in the network to give the output.After that, the cost function will equate the
output and desired output. If the value of the cost function is high then the information goes
back, and the neural network starts learning to reduce the cost function by adjusting the
weights. Proper adjustment of weights lowers the error rate and makes the model definitive.
2) FORWARD PROPAGATION
The information enters into the input layer and forwards in the network to get the output value.
The user compares the value to the expected results.The next step is calculating errors and
propagating the information backward.This permits the user to train the neural network and
modernize the weights. Due to the structured algorithm, the user can adjust weights
simultaneously. It will help the user to see which weight of the neural network is responsible for
error.
32. • This network contains an input, hidden, and output
layer. Signals can move in only one direction.
• Input data passes to the hidden layer to perform the
mathematical calculations.
• Processing element computes according to the
weighted sum of its inputs.
• The output of the previous layer becomes the input of
the following layer.
• This continues through all the layers and determines
the output.
• This network has feedback paths. It means signals can
travel in both the direction using loops.
• Neurons can have all the possible connections.
• Due to loops, it becomes a dynamic system that
changes continuously to reach in the equilibrium
state.
33. SINGLE LAYER NETWORK
• A single-layered neural network often called perceptrons.
• The input layer transmits the signals to the output layer.The
output layer performs computations.
• Perceptron can learn only a linear function and requires less
training output.
• The output can be represented in one or two values(0 or
1).
• Multilayer networks solve the classification problem for non
linear sets by employing hidden layers, whose neurons are
not directly connected to the output.
• An MLP consists of three layers of nodes: an input layer, a
hidden layer and an output layer. Its multiple layers and non-
linear activation distinguish MLP from a linear perceptron.
MULTILAYER NETWORKS
34. RADIAL BASIS FUNCTION NETWORK
• RBNN is composed of input, hidden, and output
layer. RBNN is strictly limited to have exactly one
hidden layer. We call this hidden layer as feature
vector.
• The hidden layer contain Gaussian transfer
functions whose outputs are inversely
proportional to the distance from the center of the
neuron.
• RBF networks have many applications like function
approximation, interpolation, classification and time
series prediction.
35. COMPETITIVE NETWORK
KOHONEN’S NEURAL NETWORK
Competitive learning is a form of unsupervised learning, in which
nodes compete for the right to respond to a subset of the input
data. It works by increasing the specialization of each node in the
network. It is well suited to finding clusters within data.
The Self-Organizing Map (SOM), commonly also known as
Kohonen network is a computational method for the visualization
and analysis of high-dimensional data, especially experimentally
acquired information.
A Kohonen network is composed of a grid of output units and N
input units.The input pattern is fed to each output unit.The input
lines to each output unit are weighted.These weights are
initialized to small random numbers.
36. Hopfield neural network was invented by Dr. John J.
Hopfield in 1982. It consists of a single layer which
contains one or more fully connected recurrent
neurons.The Hopfield network is commonly used for
auto-association and optimization tasks.
Discrete Hopfield Network
A Hopfield network operates in a discrete line fashion
i.e, it can be said the input and output patterns are
discrete vector, which can be either binary 0,1 or
bipolar +1,−1 in nature.The network has symmetrical
weights with no self-connections.
HOPFIELD NETWORK
37. ART (ADAPTIVE RESONANCETHEORY) MODEL
Adaptive resonance theory is developed by Stephen Grossberg and
Gail Carpenter in 1987.
The term “adaptive” and “resonance” used in this suggests that
they are open to new learning(i.e. adaptive) without discarding the
previous or the old information(i.e. resonance).The ART networks
are known to solve the stability-plasticity dilemma.
ART networks implement a clustering algorithm.
It consists of :
• F1 layer or the comparison field(where the inputs are processed)
• F2 layer or the recognition field (which consists of the clustering
units)
• The Reset Module (that acts as a control mechanism)
The reset unit makes the decision whether or not the cluster unit is
allowed to learn the input pattern depending on how similar its top-
down weight vector is to the input vector
40. HUMAN FACE RECOGNITION
• It is one of the biometric methods to identify a given face. It is a typical task because of the
characterization of “non-face” images. However, if a neural network is well trained, then it can
be divided into two classes namely images having faces and images that do not have faces.
• First, all the input images must be preprocessed.Then, the dimensionality of that image must
be reduced. And, at last it must be classified using neural network training algorithm.
-For dimensionality reduction, Principal Component Analysis PCA is used.
SIGNATUREVERIFICATION
• Signatures are one of the most useful ways to authorize and authenticate a person in legal
transactions.
• For this application, the first approach is to extract the feature or rather the geometrical
feature set representing the signature.With these feature sets, we have to train the neural
networks using an efficient neural network algorithm.This trained neural network will classify
the signature as being genuine or forged under the verification stage.
41. FINGER PRINT RECOGNITION
• Finger prints are unique pattern of ridges and valleys different foreach individual.
• AAN uses the following steps to recognize finger print
• Image recognition-a given image is digitalized into 512ₓ512 image with each pixel
assigned a particular gray scale value.
• Edge detection and thinning- these are pre-processing of the image, remove noise and
enhance the image.
• Feature extraction- features like ridge bifurcation and ridge endings of the finger print
were pointed out with the help of neural network
• Classification- class label is assigned to the image depending on the extracted features.
42. STRUCTURAL SCREENING IN DRUG DISCOVERY
PROCESS
• The standard approach to predicting how active a chemical compound will be
against a given target (usually a protein that needs to be inhibited) in the
development of new medicines is to use machine learning models .
• ANN-based QSAR models are broadly picked as the forecast strategies in the
virtual screening.
• For example, if any basic structure of the compound is the input for a neural
network, it displays various structure similar to those compounds screens over
1000 compounds, among them three compounds with high biologic activity can
be identified.
43. DISEASE DIAGNOSIS
• ANN has been applied in the diagnosis of various diseases by the input of clinical
data .
• It helps in the early prediction of cardiovascular risk or myocardial infarction,
diagnosis ofAlzheimer's disease, prediction of outcome in epilepsy surgery,
prediction of development of pregnancy-induced hypertensive disorders, etc.
Lung cancer detection using ANN-
The early detection of the lung cancer is a challenging problem, due to the
structure of the cancer cells.
The manual analysis of the sputum sample is time consuming, inaccurate and
requires highly trained person to avoid diagnostic errors.
44. There are many techniques to diagnosis lung cancer, such as chest radiograph(x-
ray), computed tomography(CT), magnetic resonance imaging(MRI scan) and
sputum cytology. However, most of these techniques are expensive and time
consuming.
Most of these techniques are detecting the lung cancer in its advanced stages,
where the patient’s chance of survival is very low.
Hence, Hopfield neural network segmentation method is used for segmenting
sputum colour images to detect the lung cancer in early stages.
The segmentation results will be used as a base for a ComputerAided Diagnosis (
CAD)system for early detection of lung cancer.
This method is designed to classify the image of N pixels.
45. Software Applications
Statistica Neural network Signal processing, Pattern recognition, Medical diagnostics and
monitoring, Image and speech analysis, Stock market and
forecasting.
Tns2server AppliedAnalytic System, Inc. Genetic Algorithm solutions, StatisticalAnalysis, Linear and
Nonlinear Optimization, Mathematical modeling.
Data Engine Functionality for all the phases of data analysis.
Environment for Computer-Aided Neural
Software Engineering (eCANSE)
Data Analysis, Experiment design and actual experimentation,
Analyzing and packaging results.
Fast Artificial Neural Network (FANN) Backpropagation training, Evolving topology training, Graphical
interface.
Brian Neural Network simulator Plotting and analysis of auditory system modelling,
Electrophysiological data
Neural Simulation Language Supporting neural network layer, Programming of neural elements
Neuroshell Large library of technical analysis, Prediction analysis, Genetic
Optimization and Multicore Distributed Optimization, Multiple
Time Frame analysis, Batch Processing.
46. REFERENCES
• https://www.sas.com/en_us/insights/analytics/what-is-artificial-intelligence.html
• https://builtin.com/artificial-intelligence
• Ankith M., SuryaTeja S. P., Damodharan N.,Artificial Neural Network in Drug Delivery
and Pharmaceutical Research, International Journal of Applied Sciences, 2018;Vol
10(Issue 5), 29-30.
• Vijaykumar Sutariyaa, Anastasia Grosheva, Prabodh Sadanab, Deepak Bhatiab and
Yashwant Pathak, Artificial Neural Network in Drug Delivery and Pharmaceutical
Research, The Open BioinformaticsJournal, 2013, 7, (Suppl-1, M5) 49-62.
• https://techvidvan.com/tutorials/artificial-neural-network/
• https://www.tutorialspoint.com/artificial_neural_network/artificial_neural_network_a
pplications.htm