UNIT-1

AI
Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are
programmed to mimic human actions and cognitive processes such as learning, problem-
solving, perception, and decision-making.
AI technologies enable computers to perform tasks that typically require human intelligence,
including natural language processing, visual perception, speech recognition, and complex
decision-making.
Artificial intelligence (AI) refers to computer systems capable of performing complex tasks
that historically only a human could do, such as reasoning, making decisions, or solving
problems.
Artificial intelligence is the intelligence of machines or software, as opposed to the
intelligence of living beings, primarily of humans. It is a field of study in computer science
that develops and studies intelligent machines. Such machines may be called AIs.

Problems in AI
1. Representation: Determining how to represent knowledge in a way that is suitable
for processing by AI systems.
2. Reasoning and Inference: Developing algorithms for logical reasoning, deduction,
and inference.
3. Learning: Designing systems that can learn from data and improve performance over
time.
4. Perception: Creating systems capable of interpreting and understanding sensory
input, such as images, sounds, and text.
5. Natural Language Processing: Developing algorithms for understanding and
generating human language.
6. Robotics: Integrating AI with robotics to enable machines to perform physical tasks
in real-world environments.
7. Ethical and Societal Impact: Addressing ethical concerns related to AI, such as bias in
algorithms, job displacement, and privacy issues.

Techniques in AI
1. Machine Learning: Algorithms that enable computers to learn from data and make
predictions or decisions without being explicitly programmed.
2. Natural Language Processing (NLP): Techniques for enabling computers to
understand, interpret, and generate human language.
3. Computer Vision: Algorithms for processing and interpreting visual information from
images or videos.
4. Expert Systems: AI systems that emulate the decision-making ability of a human
expert in a specific domain.
5. Neural Networks: Computational models inspired by the structure and function of
the human brain, used for various AI tasks including pattern recognition and
classification.
 6. Genetic Algorithms: Optimization algorithms inspired by the process of natural
selection and genetics.
7. Reinforcement Learning: Learning from feedback in the form of rewards or
punishments to improve performance in a given task.

Models in AI
1. Symbolic AI
Based on the manipulation of symbols according to predefined rules. Examples include
expert systems and logic-based reasoning systems.
Symbolic AI, also known as good old-fashioned AI (GOFAI), refers to the use of symbols and
abstract reasoning in artificial intelligence.
It involves the manipulation of symbols, often in the form of linguistic or logical expressions,
to represent knowledge and facilitate problem-solving within intelligent systems.
Symbolic Artificial Intelligence, also known as Good Old-Fashioned AI (GOFAI), uses human-
readable symbols that represent real-world entities or concepts as well as logic (the
mathematically provable logical methods) in order to create 'rules' for the concrete
manipulation of those symbols, leading to a rule-based system.

2. Connectionist Models
Inspired by the structure and function of neural networks in the brain. Examples include
artificial neural networks and deep learning models.
A connectionist model consists of a collection of units, often represented as points or
“nodes,” joined by links.
All the information stored in the network is embodied in its interconnections.
In general, the only kind of activity in a connectionist network is activation or inhibition.
 3. Bayesian Models
Based on probabilistic reasoning and inference, often used in decision-making and
uncertainty modeling.
Describes the conditional probability of an event based on data as well as prior information
or beliefs about the event or conditions related to the event.
Bayesian methods can also be used for new product development as a whole. Mainly, one
would look at project risk by weighing uncertainties and determining if the project is worth
it.

Defining Problem as a State Space Search

A process known as state space search is used to explore all potential configurations or
states of an instance until one with the necessary feature is found. A state is a time snapshot
representing some aspect of the problem.

In AI, many problems can be formulated as state space search problems, where:

• State: Represents a configuration or situation in the problem domain.

• Action: Represents a transition from one state to another.
• Goal: The desired state or states that the problem-solving agent aims to reach.
Production System
A production system is a type of knowledge representation used in AI for modeling problem-
solving procedures. It consists of a set of rules (productions) that describe actions to be
taken based on the current state of the system and the available information.
The production system in AI is the collection of those rules which work on different
behaviors and settings.
These production systems help set up the AI algorithms and look into the proper functioning
of the AI software.

Characteristics of AI
1. Adaptability: AI systems can adapt and improve their performance over time
through learning mechanisms.
2. Autonomy: AI systems can operate independently to perform tasks without human
intervention.
3. Perception: AI systems can interpret and understand sensory input from the
environment.
4. Reasoning: AI systems can make decisions and draw conclusions based on available
information and logical rules.
5. Learning: AI systems can acquire knowledge and skills from experience or data.

Search Methods
Search methods are algorithms used to traverse and explore the search space in order to
find a solution to a given problem. Here are some common search methods used in AI:

1. Depth-First Search (DFS):

• DFS explores as far as possible along each branch before backtracking.
• It uses a stack data structure for implementation.
• DFS may encounter problems with infinite paths or deep paths, and it may
not find the shortest path.
2. Breadth-First Search (BFS):
• BFS explores all neighbor nodes at the present depth level before moving on
to nodes at the next depth level.
• It uses a queue data structure for implementation.
• BFS guarantees finding the shortest path to the solution, but it may consume
more memory compared to DFS.
3. Uniform Cost Search:
• Uniform Cost Search expands the node with the lowest path cost.
• It uses a priority queue or heap data structure based on the path cost.
• Uniform Cost Search guarantees finding the optimal solution in terms of path
cost.
4. A*-Search:
• A*-Search evaluates nodes by combining the cost to reach the node (g(n))
and an estimate of the cost to reach the goal from the node (h(n)).
• It uses a priority queue or heap based on the sum of g(n) and h(n) (f(n) = g(n)
+ h(n)).
 • A*-Search is both complete and optimal if the heuristic function h(n) is
admissible and consistent.
5. Iterative Deepening Depth-First Search (IDDFS):
• IDDFS repeatedly applies DFS with increasing depth limits until the goal is
found.
• It combines the advantages of both BFS and DFS.
• IDDFS is complete and optimal for problems with finite depths.
6. Bidirectional Search:
• Bidirectional Search simultaneously explores the search space from the initial
state and the goal state.
• It terminates when the two searches meet in the middle.
• Bidirectional Search is more efficient than unidirectional searches for certain
problems, especially when the branching factor is high.
7. Greedy Best-First Search:
• Greedy Best-First Search expands the node that is closest to the goal
according to a heuristic function.
• It uses a priority queue based solely on the heuristic value.
• Greedy Best-First Search is not guaranteed to find the optimal solution, but it
can be efficient in certain cases.
8. Depth-Limited Search:
• Depth-Limited Search is similar to DFS but limits the depth of exploration to a
predefined level.
• It avoids infinite paths and can find solutions within a specified depth limit.

-> UNINFORMED SEARCH

BFS (Breath First Search)

Algorithm:
1. Place the starting node S on the queue.
2. If the queue is empty return failure and return.
3. If the first element in the queue is _ then return S and _
4. Otherwise remove and expand the first element from the queue and place all the children
at the end of the queue in any order.
5. Return to step 2.

Space Complexity-> O (b^d)

Time Complexity-> O (b^d)
DFS (Depth First Search)
Algorithm:
1. Place the starting node S on the queue.
2. If the queue is empty return failure and return.
3. If the first element in the queue is _ then return S and _
4. Remove and expand the first element and place the children at the first of the queue. 5.
Return to step 2.

Space Complexity-> O (b^d)

Time Complexity-> O (d)

Iterative Deepening DFS

Iterative Deepening Depth-First Search (IDDFS) is a combination of Depth-First Search (DFS)
and Breadth-First Search (BFS) algorithms. It is a variant of DFS that aims to combine the
advantages of both DFS and BFS, overcoming some of the limitations of traditional DFS.
Algorithm: Doesn’t exist

Space Complexity-> O (b^d/2)

Time Complexity-> O (b^d/2)
Worst Case Scenario-> O(b^d)

-> INFORMED SEARCH

Hill Climbed
Algorithm:
1. Put the initial node on the list and start list.
2. If start list is empty or start list is goal then terminate search.
3. Remove the first list from the list, call this node N.
4. If N=Goal, then terminate search with success.
5. Else if node N has successors then generate all of them, find how far they are from the
goal node, sort them by the remaining distance from the goal and add them to the
beginning of the start list.
6. Go to step 2.
Hill Climbing Problem:

1. Plateau:
• A plateau refers to a flat region in the problem space where the objective
function values remain constant over a range of neighboring solutions.
• In hill climbing, encountering a plateau means that the algorithm struggles to
make progress because all neighboring solutions have similar or identical
objective function values.
• Plateaus can pose challenges for hill climbing algorithms as they may cause
the algorithm to get stuck without finding the optimal solution.
2. Foothill:
• A foothill represents an area of the problem space where the objective
function values gradually increase or decrease as the algorithm moves
towards the optimal solution.
• Foothills are regions where the objective function provides guidance to the
algorithm, making it easier to progress towards the optimal solution.
• Hill climbing algorithms typically perform well in foothill regions, as they can
follow the gradient of the objective function towards higher (or lower)
values.
3. Ridge:
• A ridge refers to a narrow and elevated region in the problem space where
the objective function values increase in one direction and decrease in the
perpendicular direction.
• When encountering a ridge, hill climbing algorithms may struggle to traverse
across it due to the conflicting gradients in different directions.
• Ridges can present challenges similar to plateaus, as they may hinder the
progress of the algorithm towards the optimal solution.

Best-First Search/Greedy Search

Algorithm:
1. Place the starting node S on the queue.
2. If the queue is empty return failure and stop.
3. If the first element on the queue is the goal node, return success and stop otherwise
4. Remove the first element from the queue, expand it and compute the estimated goal
distance for each child. Place the children on the queue and arrange all queue elements in
ascending order corresponding to goal distance from the first of the queue.
5. Return to step 2.
Issues in the Design of Search Problems:

1. Completeness: Whether the algorithm guarantees finding a solution if one exists.

2. Optimality: Whether the algorithm finds the best (optimal) solution among all
possible solutions.
3. Time Complexity: The computational resources required to find a solution.
4. Space Complexity: The amount of memory required by the algorithm.
5. Heuristics: Designing effective heuristic functions to guide search algorithms
towards the goal efficiently.
6. Admissibility: Ensuring that heuristic estimates never overestimate the true cost to
reach the goal.
7. Consistency: Ensuring that the heuristic is consistent, meaning it satisfies the
triangle inequality property.