Artificial intelligence Training course

file:///D:/Artificial%20Intelligence%20Training%20Course.pdf

Artificial intelligence a system approach-TIM JHON

file:///C:/Users/mtariq/Downloads/Artificial%20Intelligence%20A%20System%20Approach.pdf

History of AI
I would like to personally recognize John McCarthy for introducing AI in 1955 (at
the Dartmouth Summer Project) and for having created the wonderful Lisp
programming language.
It was August of 1956 when researchers met at the Dartmouth Summer Research
Project on Artificial Intelligence with the agenda of creating intelligent machines.
In the 50 years that followed, AI has become a genuine field of study, but the road
has not been without its bumps.

TABLE OF CONTENTS:
Chapter 1 The History of AI

Chapter 2 Uninformed Search

Chapter 3 Informed Search

Chapter 4 AI and Games

Chapter 5 Knowledge Representation

Chapter 6 Machine Learning

Chapter 7 Evolutionary Computation

Chapter 8 Neural Networks I

Chapter 9 Neural Networks II

Chapter 10 Robotics and AI

Chapter 11 Intelligent Agents

Chapter 12 Biologically Inspired and Hybrid Models

Chapter 13 The Languages of AI

Chapter 1 The History of AI
CHAPTER SUMMARY: The history of AI is a modern-day drama. It’s filled with
interesting characters, cooperation, competition, and even deception. But outside
of the drama, there has been exceptional research and in recent history an
application of AI’s ideas in a number of different settings. AI has finally left the
perception of fringe research and entered the realm of accepted research and
practical development.

Chapter 2 Uninformed Search

CHAPTER SUMMARY: Uninformed search algorithms are a class of graph search
algorithms that exhaustively search for a node without the use of a heuristic to
guide the search. Search algorithms are of interest in AI because many problems
can be reduced to simple search problems in a state space. The state space
consists of states (nodes) and operators (edges), allowing the state space to be
represented as a graph. Examples range from graphs of physical spaces to
massive game trees such as are possible with the game of Chess. The depth-first
search algorithm operates by evaluating branches to their maximum depth, and
then backtracking to follow unvisited branches. Depth-limited search (DLS) is
based on DFS, but restricts the depth of the search. Iterative-deepening search
(IDS) uses DLS, but continually increases the search depth until the solution is
found. The breadth-first search (BFS) algorithm searches with increasing depth
from the root (searches all nodes with depth one, then all nodes with depth two,
etc.). A special derivative algorithm of BFS, bidirectional search (BIDI), performs
two simultaneous searches. Starting at the root node and the goal node, BIDI
performs two BFS searches in search of the middle. Once a common node is
found in the middle, a path exists between the root and goal nodes. The uniform-
cost search (UCS) algorithm is ideal for weight graphs (graphs whose edges have
costs associated with them). UCS evaluates a graph using a priority queue that is
ordered in path cost to the particular node. It’s based on the BFS algorithm and is
both complete and optimal.
Chapter 3 Informed Search
CHAPTER SUMMARY: Informed search methods, unlike the uninformed (or blind)
methods, use a heuristic to determine the quality of any state in the search space.
This allows the algorithm to guide the search and choose the next node to
expand. In this chapter, a number of informed search methods were explored,
including best-first search, A* search, hill-climbing search, simulated annealing,
Tabu search, and finally algorithms for constraint satisfaction. Understanding the
informed methods of search is important because much of problem solving in AI
can be refined to search methods. The key is choosing the algorithm that is best
suited to the particular problem at hand.

Chapter 4 AI and Games

CHAPTER SUMMARY: This chapter provided a broad spectrum of techniques used
in game AI, from classical games, to modern video games. In classical games such
as Chess and Checkers, and even information imperfect games such as Poker, the
minimax game-tree search algorithm with alpha-beta pruning can be found. In
these games, the goal of developers is to build an opponent that can defeat
worldclass players. In contrast, the goal in video games is to build opponents that
are challenging to play, but not perfect. Even with this limitation in mind, various
AI technologies can be found in modern games, from neural networks, to rule-
based systems and decision trees, to embed game opponents that are both
challenging and adaptive to the human’s level of play.

Chapter 5 Knowledge Representation

CHAPTER SUMMARY: Knowledge representation focuses on the mechanisms by
which information can be stored and processed in a way to make the information
useful as knowledge. This chapter presented some of the more important
mechanisms for knowledge representation, such as semantic networks and
frames, and also methods that enable not only storage of knowledge, but also
processing, such as propositional logic and first-order (or predicate) logic. Finally,
applications of knowledge representation were explored including computational
scientific discovery, and common-sense reasoning with the Cyc project.
Chapter 6 Machine Learning
CHAPTER SUMMARY: While machine learning is one of the older techniques of
AI, it remains useful and effective for general learning. In this chapter, the ideas
behind supervised and unsupervised learning were explored and a collection of
algorithms that demonstrate these approaches were given. Decision trees were
introduced from the perspective of supervised learning, and also Markov chains
and nearest neighbor algorithms from the unsupervised learning perspective. All
are useful for learning and classification in a wide variety of problem domains.

Chapter 7 Evolutionary Computation

CHAPTER SUMMARY: While evolutionary algorithms are not new, you’ll find
continued research and even new algorithms being developed today (for
example, differential evolution and swarm intelligence). Evolutionary algorithms
borrow concepts from Darwinian natural selection as a means to evolve solutions
to problems, choosing from more fit individuals to propagate to future
generations. In this chapter we explored a number of evolutionary and
biologically-inspired algorithms. After an introduction to evolutionary algorithms,
we presented the genetic algorithm which is at the core of most evolutionary
algorithms. Next, we explored genetic programming, an evolutionary means to
generate code sequences. We then reviewed one of the original evolutionary
methods, evolutionary strategies. Next, we reviewed the new method of
differential evolution, and finished with a review of particle swarm optimization (a
biologically plausible optimization method).

Chapter 8 Neural Networks I

CHAPTER SUMMARY: Neural networks are an ideal solution to problems that
can’t be formulated easily to an algorithm, and for which lots of examples exist
from which the network can be trained. Additionally, neural networks are highly
parallel and distributed structures with the desirable property that they
generalize well to unseen data. This chapter presented an introduction to neural
networks using supervised learning algorithms, including perceptron learning, the
delta-rule, backpropagation, and probabilistic neural networks.
Chapter 9 Neural Networks II
CHAPTER SUMMARY: This chapter explored a variety of unsupervised neural
network architectures and learning algorithms. Unsupervised learning algorithms
are useful to discover the underlying structure of a data set. We began with a
biologically plausible learning method for patterns called Hebbian Learning and
demonstrated its pattern storage and recall capabilities. Next, we explored a
number of clustering algorithms that were capable of segregating data based on
their similarities and differences. Algorithms explored included vector
quantization, k-Means clustering, and Adaptive Resonance Theory (ART-1). Finally,
we ended the FIGURE 9.14: Sample activations for the Hopfield model
implementation. 328 Artificial Intelligence chapter with a return to pattern
storage and recall with a discussion of the Hopfield auto-associator (another
biologically plausible model).

Chapter 10 Robotics and AI

CHAPTER SUMMARY: The field of robotics and AI are inexorably tied because
robotics permit the visualization of AI algorithms in the physical world. Robotics
exhibit the true systems level of AI, as there are distinct inputs (sensors), outputs
(effectors), and the AI algorithms that provide rational decision-making. Robotics
is an interesting area of study because they encompass many elements of AI. The
algorithms demonstrated throughout this book are applicable to AI systems, from
search for use in planning, to neural networks and evolutionary systems for
learning. Robotics are the extension of AI algorithms into the real world.

Chapter 11 Intelligent Agents

CHAPTER SUMMARY: Intelligent agents are an interesting exploration in artificial
intelligence as they are key users of AI methods and techniques. While agents can
be viewed as an end application of AI, their study introduces concepts of
intelligent systems and fundamental architectures for their implementation. This
chapter provided an introduction to agent systems, characteristics, and
architectures. You’ll find agent systems in games, web applications (such as user
assistance) as well as internally in complex and distributed applications.
Chapter 12 Biologically Inspired and Hybrid Models
AFFECTIVE COMPUTING: Affective computing is all about emotion, but from two
different perspectives. First, it’s about the detection and classification of emotions
of a user. Second, it’s about the synthesis of emotions from a computer to a user.
Let’s explore what this means and why it’s important. Emotions are a critical
aspect of human communication (a side-band, if you will, of information).
Consider the emotion of confusion. If a computer could detect this emotion from
a user, it could alter the way in which it’s presenting information. In order to
present a computer as a believable entity, emotion must be part of its
communication (either through speech, or acts of speech). Emotions may also be
more fundamental in nature than intelligence. For example, animals are capable
of understanding emotion in humans (consider a dog’s response to a human’s
gestures or speech), and while they’re intelligent, it’s certainly at a different level.
Characterizing Human Emotion: Characterizing human emotion can occur in
numerous ways, but provide different types of information. For this reason, a
hybrid approach is necessary Biologically Inspired and Hybrid Models 431 for a
complete and unambiguous view of the emotion of a human. Speech is a multi-
channel interface, and in addition to the information being conveyed, information
is also available in the patterns and pitch of the speaker. For example, the pitch of
the speech, the speed of the speech, and the inflections in the speech that can be
tuned by emotion, are all detectable through a microphone and software for
analysis. Gestures are also an important channel of communication for emotion.
Some of this information is context-dependent. For example, inattention is an
important gesture to understand for learning, but is even more important when
used to understand a driver’s awareness. Hand gestures are also important, as are
posture and facial expressions. Other important indicators of emotion include
physiological parameters, most of which can require more than just visually
monitoring a subject. Gauging a person’s temperature can be done through
infrared, but heart rate or blood pressure requires more invasive monitoring.
Giving a computer the ability to characterize and understand a human’s emotive
state is very important in future human-computer interfaces. With this we can
build systems that not only communicate with a user (from the perspective of
Strong AI), but can use subtle queues to build more efficient interfaces.
Synthesizing Emotion: Synthesizing human emotion in machines is an important
aspect of Strong AI, particularly for the development of believable characters
(computer agents) or in robotic systems. But human emotions are very complex
and multi-dimensional. For example, happiness and sadness can be considered
polar opposites, but a human can be both happy and sad at the same time, so
they’re not opposite ends of a spectrum. Considering the variety and depth of
human emotion, synthesizing an agent with believable emotions will be difficult
indeed. Research exists to pursue the simulation of emotions given its importance
to a number of applications, from Strong AI to computer game agents. The
emotions to be synthesized are based on what’s available to the simulated agent.
Computer agents (those in computer games or other applications) typically have a
full range of channels, as characters are typically fully modeled with synthesized
speech. But visualization of emotions is the simpler aspect of this problem. The
more difficult problem of synthesizing emotion is generating emotion from a
variety of internal agent states. How do these internal states map to the
dimensionality of emotion? First steps would involve minimizing 432 Artificial
Intelligence the emotive states to a more manageable set, and then evolving this
set with experience. Emotions may also play an important role in the internal
function of simulated agents. Consider the survival response of animals. Animals
can intellectually understand a situation and determine a response, but emotions
and other physiological factors certainly apply in determining a rationale response
to a situation.

Chapter 13 The Languages of AI

CHAPTER SUMMARY: Languages are the vehicles by which developers create
solutions to problems. But unfortunately, languages are also religious artifacts
that become the centerpieces of heated debates. But ultimately, languages are
tools, and each has their strengths and weaknesses. This chapter introduced the
major classes of languages for AI, and also discussed four particular languages
that are used in AI application development. The chapter concluded with a review
of some of the features that started in the older functional languages, but have
now been applied to modern domainspecific languages.
KEY TAKEAWAYS

 Artificial intelligence (AI) refers to the simulation or approximation of

human intelligence in machines.
 The goals of artificial intelligence include computer-enhanced learning,
reasoning, and perception.
 AI is being used today across different industries from finance to
healthcare.
 Weak AI tends to be simple and single-task oriented, while strong AI
carries on tasks that are more complex and human-like.
 Some critics fear that the extensive use of advanced AI can have a
negative effect on society.

"Scruffies" use any number of different algorithms and methods to

achieve intelligent behavior.

Algorithms often play a very important part in the structure of artificial

intelligence, where simple algorithms are used in simple applications, while
more complex ones help frame strong artificial intelligence.

Artificial intelligence is based on the principle that human intelligence can be

defined in a way that a machine can easily mimic it and execute tasks, from
the most simple to those that are even more complex. The goals of artificial
intelligence include mimicking human cognitive activity.

Applications of Artificial Intelligence

The applications for artificial intelligence are endless. The technology can be
applied to many different sectors and industries. AI is being tested and used
in the healthcare industry for suggesting drug dosages, identifying
treatments, and for aiding in surgical procedures in the operating room.

Other examples of machines with artificial intelligence include computers that

play chess and self-driving cars. Each of these machines must weigh the
consequences of any action they take, as each action will impact the end
result. In chess, the end result is winning the game. For self-driving cars, the
computer system must account for all external data and compute it to act in a
way that prevents a collision.

Artificial intelligence also has applications in the financial industry, where it is

used to detect and flag activity in banking and finance such as unusual debit
card usage and large account deposits—all of which help a bank's fraud
department. Applications for AI are also being used to help streamline and
make trading easier. This is done by making supply, demand, and pricing of
securities easier to estimate.

Types of Artificial Intelligence

Artificial intelligence can be divided into two different categories: weak and
strong. Weak artificial intelligence embodies a system designed to carry out
one particular job. Weak AI systems include video games such as the chess
example from above and personal assistants such as Amazon's Alexa and
Apple's Siri. You ask the assistant a question, and it answers it for you.

Strong artificial intelligence systems are systems that carry on the tasks

considered to be human-like. These tend to be more complex and
complicated systems. They are programmed to handle situations in which
they may be required to problem solve without having a person intervene.
These kinds of systems can be found in applications like self-driving cars or
in hospital operating rooms.

Content Creation
Artificial intelligence is good for content creators to make unique
content and also they are free of grammatical errors. With the help of
artificial intelligence, the content creators can create good pictures and
good quality content such as educational, news, technical, sports,
celebrity, finance, comedy etc. Everyone is using the internet to get
some knowledge related to their field.
Smart Mentoring Systems
A mentoring system is helpful for teachers because in the classroom
teachers find it difficult to pay attention to every student to teach some
enlightening things. Smart mentoring systems play a very significant
role. Based on the capacity of the student and their ability to grasp
things, every class can be handled accordingly. Some students are not
bright in their studies, but with the help of a smart monitoring system,
it will be possible for the student to get some knowledge based on their
studies. A mentoring system will work on the student's abilities and
create a syllabus according to the student's ability. Although work is still
ongoing, many online learning training institutes have opted for this
type of teaching. If this system is implemented, it will change the
learning process.

Augmented Reality
Augmented reality is a digital medium that allows you to overlay the
virtual world in a physical term in such a way that it appears that the
material is physical. Applying it to the education system can change the
actual learning experience. Students will be able to experience the real
world instead of learning only theoretically. Augmented reality will not
only improve academic standards but will also increase the quality of
students and make them better humans.

Personalized Learning
Personalized learning refers to a variety of techniques that support
each student's unique learning style. Some students are not keen on
their studies. To eradicate this, a new method called personal
education will be introduced very soon. The development of virtual
assistants helps to handle classrooms and will change the entire
experience of education.

AI-Based Virtual Assistants

We know about Alexa and Google Assistant and we also know how
virtual assistants help us in our daily life. Ask Alexa anything and we'll
get an answer. And it also helps the students to get answers with the
help of virtual aid, there will be advancement in the context of research
and a new era will come, where every student can realize their full
potential and serve the society in the best way.

. What is AGI ?
Artificial general intelligence is the software that represents generalized
human cognitive abilities that would allow an Artificial general
intelligence system to solve a problem when given a difficult task. An
Artificial general intelligence system's goal is to carry out all tasks that
human beings could be capable of.
. Why Is AGI Important?

Artificial general intelligence aims to build machines with human-like reasoning, learning, and problem-
solving abilities. Artificial general intelligence systems have the potential to revolutionize numerous
industries and address challenging problems in science, medicine, and environmental issues.

How Does AGI Work?

Different names for artificial general intelligence are strong AI and deep artificial intelligence. This is a
concept for a computer with general intelligence that replicates human intelligence and can reason,
understand, pick up knowledge, and use that ability to resolve any issue in any situation, much like
people do.
Applications Of AGI

1. Spam filters on email

2. Social media monitoring tools

3. Intelligent assistants (like Siri and Alexa)

4. Condition mapping and prediction tools

5. Manufacturing and drone robots

6. Personalized healthcare treatment recommendations

8. Robo-advisors for stock trading

9. TV show recommendations from Spotify and Netflix

What will yoy learn

Python

Machine Learning

Deep Learning

Neural Networks

Data Visualization

Recommendation Systems