Introduction to Artificial Intelligence.

Artificial Intelligence
and Machine Learning

Intelligence :
 Human beings are Homo sapiens (man the wise) ----- INTELLIGENCE

The Foundations Of Artificial Intelligence :
 Philosophy
 Can formal rules be used to draw valid conclusions?
 How does the mind arise from a physical brain?
 Where does knowledge come from?
 How does knowledge lead to action?
 Mathematics
 What are the formal rules to draw valid conclusions?
 What can be computed?
 How do we reason with uncertain information?
 Economics
 How should we make decisions so as to maximize payoff?
 How should we do this when others may not go along?
 How should we do this when the payoff may be far in the future?
 Neuroscience
 How do brains process information?

The History Of Artificial Intelligence :
 The gestation of artificial intelligence (1943–1955)
 Warren McCulloch and Walter Pitts (1943)
 Donald Hebb (1949)
 Alan Turing (1950)
 The birth of artificial intelligence (1956)
 John McCarthy
 Early enthusiasm, great expectations (1952–1969)
 Newell and Simon (1976) physical symbol system
 Marvin Minsky (1958)
 Bernie Widrow (Widrow and Hoff, 1960; Widrow, 1962), ADALINES
 Frank Rosenblatt (1962), Perceptrons
 A dose of reality (1966–1973)
 Friedberg (1959), genetic algorithms

The History Of Artificial Intelligence :
 Knowledge-based systems: The key to power? (1969–1979)
 Expert systems
 Certainty factors
 AI becomes an industry (1980–present)
 Digital Equipment Corporation (McDermott, 1982)
 The return of neural networks (1986–present)
 Back propagation learning algorithm first found in 1969 by Bryson and Ho
 AI adopts the scientific method (1987–present)
 DATA MINING
 Bayesian network
 The emergence of intelligent agents (1995–present)
 The availability of very large data sets (2001–present)

Goals of AI :
 Scientiﬁc Goal : Scientific goal is to determine which ideas about knowledge
representation, learning, rule systems, search, and so on, explain various sorts of
real intelligence (e.g., implementation of Expert Systems which exhibit intelligent
behaviour, learn, demonstrate, explain, and advice its users).
 Engineering Goal : Engineering goal is to solve real-world problems by using AI
techniques, such as knowledge representation, learning, rule systems, search, and
so on.
 For example, implementation of Human Intelligence in Machines, which means creating
systems that understand, think, learn, and behave like humans.
 In a traditional manner, computer scientists and engineers are more concerned in
the engineering goal, whereas psychologists, philosophers, and cognitive scientists
have been more absorbed in the scientific goal.
 It makes good sense to be concerned in both, as there are common techniques and
the two approaches can feed off each other.

Categorization of AI :
 Sensing - Through the sensor taking in data about the world which includes:
 1. In image processing field its recognizing important objects, paths, faces, cars, or kittens
and all other part present in the image.
 2. In speech recognition filtering out the noise and then recognizing specific words from the
input speech.
 3. Some examples of other sensors are robotics, sonar, accelerometers, balance detection,
etc.
 Reasoning - Reasoning is thinking or process the data sensed by the sensor about
how things relate to what is known as follows:
 1. In planning/problem solving, reasoning is figuring out what to do to achieve a goal.
 2. In learning, reasoning is building new knowledge based on examples or examination of a
data set.
 3. In natural language generation, reasoning is given a communication goal, generating the
language to satisfy it.

Categorization of AI :
 Reasoning –
 4. In situation assessment, reasoning is figuring out what is going on in the world at a broader
level than the ideas alone.
 5. In logic-based inference, reasoning is deciding that something is true because, logically, it
must be true.
 6. In language processing, reasoning is turning words into ideas and their relationships.
 7. In evidence-based inference, reasoning is deciding that something is true based on the
weight of evidence at hand.
 Acting - On the basis of input and reasoning, acting is generating and controlling
actions in the environment as follows:
 1. Like in speech generation, the action can be given a piece of text or generating the audio that
expresses that text.
 2. In robotic control, action is moving and managing the different effectors that move you about
the world.

Components of AI :
 In AI, the intelligence is intangible which is composed of mainly five techniques as follows :
 1. Reasoning
 2. Learning
 3. Problem solving
 4. Perception
 5. Linguistic intelligence

 Reasoning : Reasoning is the set of processes that enables an intelligent system to
help or to provide basis for actions, making decisions, and prediction. Reasoning is
of two types: Inductive Reasoning and Deductive Reasoning.
 1. Inductive reasoning conducts specific observations to make broad, general
statements.
 2. Deductive reasoning which starts with a general statement and checks the
possibilities to reach a specific or a logical conclusion.
 Learning : Learning is the process of gaining knowledge by understanding,
practicing, being taught, or experiencing one thing. Learning enhances the
awareness of any topic. The flexibility of learning is possessed by humans, some
animals, and AI-enabled systems.
Components of AI :

 Problem Solving : Problem solving is the method during which one perceives and
tries to make a desired answer from a present state of affairs by taking some path,
that is blocked by known or unknown hurdles. Drawback solving also includes
deciding that is the method of choosing the most effective appropriate alternative
out of multiple alternatives to succeed in the specified goal are available.
 Perception : Perception is the method of acquiring, decoding, selecting, and
organizing sensory data. Perception presumes sensing. In humans, perception is
aided by sensory organs. Within the domain of AI, perception mechanism puts the
info acquired by the sensors along in a very meaningful manner.
 Linguistic Intelligence : Linguistic intelligence is one’s ability to use, comprehend,
speak, and write the verbal and written language. It is important in interpersonal
communication.
Components of AI :

Applications of AI :
 What can AI do today?
 Robotic vehicles
 Speech recognition
 Autonomous planning and scheduling
 Game playing
 Spam fighting
 Logistics planning
 Robotics
 Machine Translation

Agents and Environments :
 An agent is anything that can be viewed as perceiving its environment through
sensors and acting upon that environment through actuators.
 The term percept is used to refer to the agent’s perceptual inputs at any given
instant. an agent’s choice of action at any given instant can depend on the entire
percept sequence observed to date, but not on anything it hasn’t perceived.

Example :
 A vacuum-cleaner world with just two
locations.
 Percept sequence Action
[A, Clean] Right
[A, Dirty] Suck
[B, Clean] Left
[B, Dirty] Suck
[A, Clean], [A, Clean] Right
[A, Clean], [A, Dirty] Suck
...
...
[A, Clean], [A, Clean], [A, Clean] Right
[A, Clean], [A, Clean], [A, Dirty] Suck
...
...

The Concept of Rationality :
 A rational agent is one that does the right thing.
 What does it mean to do the right thing?
 When an agent is plunked down in an environment, it generates a sequence of
actions according to the percepts it receives. This sequence of actions causes the
environment to go through a sequence of states.
 If the sequence is desirable, then the agent has performed well. This notion of
desirability is captured by a performance measure that evaluates any given
sequence of environment states.
 Note : it is said environment states, not agent states.

Rationality :
 What is rational at any given time depends on four things:
 The performance measure that defines the criterion of success.
 The agent’s prior knowledge of the environment.
 The actions that the agent can perform.
 The agent’s percept sequence to date.
 This leads to a definition of a rational agent:
 For each possible percept sequence, a rational agent should select an action that is
expected to maximize its performance measure, given the evidence provided by the
percept sequence and whatever built-in knowledge the agent has.
 Example of vacuum cleaner : performance measure
geography of the environment
actions
repeating the sequence

Omniscience, learning, and autonomy :
 An omniscient agent knows the actual outcome of its actions and can act
accordingly; but omniscience is impossible in reality.
 Rationality maximizes expected performance, while perfection maximizes actual
performance. Retreating from a requirement of perfection is not just a question of
being fair to agents.
 Doing actions in order to modify future percepts is sometimes called information
gathering.
 The definition requires a rational agent not only to gather information but also to
learn as much as possible from what it perceives. The agent’s initial configuration
could reflect some prior knowledge of the environment, but as the agent gains
experience this may be modified and augmented. There are extreme cases in which
the environment is completely known a priori.

Omniscience, learning, and autonomy :
 To the extent that an agent relies on the prior knowledge of its designer rather than
on its own percepts, we say that the agent lacks autonomy.
 A rational agent should be autonomous - it should learn what it can to
compensate for partial or incorrect prior knowledge.
 After sufficient experience of its environment, the behavior of a rational agent can
become effectively independent of its prior knowledge.
 Hence, the incorporation of learning allows one to design a single rational agent
that will succeed in a vast variety of environments.

The Nature of Environments :
 To build rational agents, it is primary requirement to think about task
environments, which are essentially the “problems” to which rational agents are the
“solutions.”
 Step 1 : Specifying the task environment
 In designing an agent, the first step must always be to specify the task environment
as fully as possible.
 Under the heading of task environment, ideally we have to consider PEAS
(Performance measure, Environment, Actuators, Sensors)

Step 2 : Properties of task environments
 Fully observable vs. partially observable
 Single agent vs. multiagent
 Deterministic vs. stochastic
 Episodic vs. sequential
 Static vs. dynamic
 Discrete vs. continuous
 Known vs. unknown

The Structure of Agents :
 The job of AI is to design an agent program that implements the agent function the
mapping from percepts to actions.
 The programs needs to run on some computing device with sensors and actuators.
This involves architecture to interface all the required elements.
 Hence, agent = architecture + program
 Agent programs : The agent programs that are designed take the current percept
as input from the sensors and return an action to the actuators.
 Notice the difference between the agent program, which takes the current percept as
input, and the agent function, which takes the entire percept history.
 The agent program takes just the current percept as input because nothing more is
available from the environment; if the agent’s actions need to depend on the entire
percept sequence, the agent will have to remember the percepts.

Agent programs :
 Example : The agent program for a simple reflex agent in the two-state vacuum
environment.
 There are four basic kinds of agent programs that embody the principles underlying
almost all intelligent systems:
 Simple reflex agents;
 Model-based reflex agents;
 Goal-based agents; and
 Utility-based agents.

Simple reflex agents :
 The simplest kind of agent is the simple reflex agent. These agents select actions on
the basis of the current percept, ignoring the rest of the percept history
 A simple reflex agent acts according to a rule whose condition matches the current
state, as defined by the percept. rectangle shape- current internal state Oval-
background information used in the process.
 Only applicable in fully observable situation

Model-based reflex agents :
 The most effective way to handle partial observability is for the agent to keep track
of the part of the world it can’t see now. That is, the agent should maintain some
sort of internal state that depends on the percept history and thereby reflects at
least some of the unobserved aspects of the current state.
Fig : A model-based reflex agent keeps track of the current
state of the world, using an internal model. It then chooses an
action in the same way as the reflex agent.

Goal-based agents :
 The agent needs some sort of goal information that describes situations that are
desirable.
 For example, at a road junction, the taxi can turn left, turn right, or go straight on. The
correct decision depends on where the taxi is trying to get to.
Fig : A model-based, goal-based agent.
It keeps track of the world state as well
as a set of goals it is trying to achieve,
and chooses an action that will
(eventually) lead to the achievement of
its goals.

Utility-based agents :
 An agent’s utility function is essentially an internalization of the performance
measure. If the internal utility function and the external performance measure are
in agreement, then an agent that chooses actions to maximize its utility will be
rational according to the external performance measure.
 Utility based learning gives degree of satisfaction of goal among multiple goals
(prioritise goal) or conflicting goals(trade off between goals)
Fig : A model-based, utility-based agent. It uses
a model of the world, along with a utility
function that measures its preferences among
states of the world. Then it chooses the action
that leads to the best expected utility, where
expected utility is computed by averaging over
all possible outcome states, weighted by the
probability of the outcome.

Simple reflex agents Model-based reflex agents
Goal-based agents Utility-based agents
Agents and Environments

Learning agents :  Learning agent is creating state of art
model and teach them. It is divided
into 4 parts.
 Learning element is responsible for
making improvements from inputs from
critic based on fixed targets
 Performance element is responsible for
selecting external actions based on
percept history.
 The learning element uses feedback from
the critic on how the agent is doing and
determines how the performance element
should be modified to do better in the
future.
 Problem generator is responsible for
suggesting actions that will lead to new
and informative experiences.
Fig : A general learning agent.

How the components of agent programs work?
“What is
the world
like now?”
“What
action
should I do
now?”
“What do
my actions
do?”
“How on
earth do
these
components
work?”
The agent programs consist of various components, whose function it is to answer
questions such as,

Atomic representation :
 In an atomic representation each state of the world is indivisible - it has no internal
structure.
 Consider the problem of finding a driving route from one end of a country to the
other via some sequence of cities.
 The algorithms underlying search and game-playing, Hidden Markov models
(HMM) and Markov decision processes all work with atomic representations

Factored representation :
 A factored representation splits up each state into a fixed set of variables or
attributes, each of which can have a value.
 While two different atomic states have nothing in common - they are just different
black boxes - two different factored states can share some attributes (such as being
at some particular GPS location) and not others (such as having lots of gas or
having no gas); this makes it much easier to work out how to turn one state into
another.
 Many important areas of AI are based on
factored representations, including constraint
satisfaction algorithms, propositional logic,
planning, Bayesian networks and the machine
learning algorithms

Structured representation :
 It is required to understand the world has things
in it that are related to each other, not just
variables with values.
 Structured representations underlie relational
databases and first-order logic, first-order
probability models, knowledge-based learning
and much of natural language understanding.
 The axis along which atomic, factored, and structured representations lie is the axis
of increasing expressiveness.
 To gain the benefits of expressive representations while avoiding their drawbacks,
intelligent systems for the real world may need to operate at all points along the
axis simultaneously.

Introduction to Artificial Intelligence.

More Related Content

Similar to Introduction to Artificial Intelligence. (20)

Recently uploaded (20)

Introduction to Artificial Intelligence.