Artificial Intelligence

Dr. Meghana Harsh Ghogare

Definition
• Artificial intelligence is the simulation of human intelligence
processes by machines.

• What is AI
(AI) is the ability of a computer or a robot controlled by a
computer to do tasks that are usually done by humans
Diffrence between AI and ML
• AI –Engineering of making Intelligent Machines and Programes
• ML-Ability to learn without being explicitly programmed
In 1997 Deep Mind was the first AI to beat Garry
Kasparov in Chess.
• Most widely telecasted game ever
Lee Sedol 18 times world champion in Go
Chess was defeated by AI AlphaGo
• Complex than Chess, China Japan & Korea
• We use games to train AI
• Games are a confined Environment,
• If we can beat a game than same principals can be applied to Business
They applied artificial intelligence
to a Google warehouse to control the air cooling
• As soon they switch on the AI, it goes down.
• And then they switch it back off, it goes up.
• They were able to save 40%
• on their electricity bill,
Chat GPT
• https://chat.openai.com/
• The full form of GPT is "Generative Pre-trained Transformer." GPT is a type
of deep learning model that uses a transformer architecture and is trained
on large amounts of text data to generate human-like responses
• Best library for web scraping data
• Syntax error
• Error message
• I have a dataset named data.csv with a dependent variable in the last
column and an index in the first column. Can you please write a python
code that builds a logistic regression model and trains it on this dataset.
Also please return accuracy at the end
What is Reinforcement Learning
• maze is our representation of an environment.
• agent is our artificial intelligence, it will be learning from the feedback
• is the agent performs certain actions in this environment.
• And as a result, the state will change
• Based upon the action it gets a reward +/-
• It learns from the feedback
• Actions leading to bad rewards and sates are avoided
• Driving a car and policeman is a feedbback
• Training a Dog
Major Areas of AI
•Computer vision
•Fuzzy logic
•Expert systems
•Robotics
•Machine learning
•Neural networks/deep learning
•Natural language processing
Computer Vision
• techniques used computer to understanding digital images and videos
• Uses ML to identify objects, faces, people, animals etc.
• CNN breaks images down into pixels, giving them tags
• The neural network then uses the labels to conduct convolutions and
make predictions about what it sees.
• Object tracking
• Image classification
• Facial recognition
Fuzzy Logic

• copies human decisions by considering all existing possibilities between

digital values of ‘yes’ and ‘no’.
• Washing Machine
• Expert systems
• is a program specializing in a singular task, just like a human expert.
• Eg: CaDeT, a diagnostic support system help detecting cancer in its early
stages.
• Robotics
• help humans with tedious and repetitive tasks,
• Industry, Space exploration, Sophia
• Machine learning
• ability of machines to automatically learn from data and algorithms
• Supervised learning
• UnSupervised learning
• Reinforcement learning
• Neural networks/deep learning
• consisting of an input layer, one or more hidden layers, and an output
layer.
• Natural language processing
• allows computers to understand both text and spoken words like
humans can
• Virtual chatbots, Spam detection, Sentiment analysis
AI Techniques
AI History
1.1950- Allan Turing –Turing test( a test of intelligence of computer, a human should
not be unable to distinguish between a machine and human. Eg: chatbots
2.1951, Isaac Asimov proposed the Three Laws of Robotics.
3.1955, the first AI-based software was created
4.In 1959, the first self-learning video game
5.1964 The first robot is installed on the GM assembly line
6.1974 the first chatbot Eliza
7.1997, IBM’s Deep Blue chess program defeated Garry Kasparov
8.2004 first emotional AI in 2004
9.2010 Google begins developing a self-driving automobile. IBM Watson
defeats the champions of Jeopardy that year. Siri, Google Now, and Cortana
have gained popularity.
AI Problems
• Safety: The self-improving AI systems can become so mighty than humans
that could be very difficult to stop from achieving their goals, which may
lead to unintended consequences.
• Data Privacy and Security: An AI program that recognizes speech and
understands natural language is theoretically capable of understanding
each conversation on e-mails and telephones.
• Human-level
• Trust Deficit
• Computing Power
• Insufficient Data (sufficient amount of data is needed to train the model)
• Job Loss Concerns
• Threat to Human Dignity: nursing, surgeon, judge, police officer, etc.
Production System
• Production systems are computer programmes, It consists of a set of
rules as the system reacts to external conditions.
• Condition --------→ Action
• Components of a Production System in AI
•
• Global Database
• A database contains all the necessary data and information
required for the successful completion of a task.
• It can be divided into two parts as
• Permanent (consists of fixed actions) and temporary(part alters
according to circumstances).
• Production Rules : rules operate on the global database.
• Each rule consists of a precondition and postcondition that the global
database either meets or not. For example, if a condition is met by
the global database, then the production rule is applied successfully.
• Control System: decides which production rule should be applied. It
stops computation when a termination condition is met
What are the Features of a Production System?
• A production system has the following features:
1.Simplicity: Due to the use of the IF-THEN structure, each sentence is
unique and Simple
2.Modularity: The code is in discrete pieces, which makes it easy to
add, modify, or delete the information without any side effects.
3.Modifiability: Easy to Modify because defined in the skeletal form
and then modified to suit an application.
4.Knowledge-intensive: As the name suggests, the system only stores
knowledge to solve the problem
Problem Characteristics
• To choose an appropriate solution we need to categorize the problem based on
the following characteristics.
1.Is the problem decomposable into small sub-problems which are easy to solve?
The decomposable problem can be solved easily.
1.Can solution steps be ignored or undone?
1. Eg In the 8-Puzzle, Moves can be undone and backtracked Recoverable problems.
2. Irrrecoverable
2.Is the universe of the problem is predictable?
1. Playing Bridge, We cannot know exactly where all the cards are or what the other players will
do on their turns. For uncertain-outcome problems, a sequence of generated operators can
only have a good probability of leading to a solution. Plan revision is made as the plan is
carried out and the necessary feedback is provided.
2. For certain-outcome problems, planning can be used to generate a sequence of operators
that is guaranteed to lead to a solution.
1.Is a good solution to the problem is absolute or relative?
1.Asolute without campareing
2.Relative with camparision
2.Is the solution to the problem a state or a path?
• The Water Jug Problem, the path that leads to the goal must be reported.
•

1.What is the role of knowledge in solving a problem using artificial

intelligence?
2.Does the task of solving a problem require human interaction?
Intelligent Agent

entity which act upon an environment using sensors and actuators for achieving
goals. An intelligent agent may learn from the environment to achieve their goals. A
thermostat(a device that automatically regulates temperature, or
that activates a device when the temperature reaches a certain point.
) is an example of an intelligent agent.
Following are the main four rules for an AI agent:
•Rule 1: An AI agent must have the ability to perceive the environment.
•Rule 2: The observation must be used to make decisions.
•Rule 3: Decision should result in an action.
•Rule 4: The action taken by an AI agent must be a rational action.
AI Applications(Ecommerce )
• Increase in volume of voice search
• chatbots
• Recommendation systems
• Image tagging
• Personalized search
• shopping from Mobile witnessing a rise(73%)
• Handling fake reviews
• Image Searching and Recognition
• Augmented and Virtual Reality (3D presentations)
• Sales Forecasting
AI Applications in Medicine
• Error reduction in diagnosis
• Reducing the costs of care
• Develop drugs faster
• Personalize treatment(design the right treatment plan.)
• Improve gene editing
• But this is just the beginning
AI Representation
• Humans are best at understanding, reasoning, and interpreting knowledge.
• But how machines do all these things comes under knowledge representation and
reasoning
• Following are the kind of knowledge which needs to be represented in AI systems:
• Object: All the facts about objects in our world domain. E.g., Guitars contains strings,
trumpets are brass instruments.
• Events: Events are the actions which occur in our world.
• Performance: It describe behavior which involves knowledge about how to do things.
• Meta-knowledge: It is knowledge about what we know.
• Facts: Facts are the truths about the real world and what we represent.
• Knowledge-Base: The central component of the knowledge-based agents is the
knowledge base. It is represented as KB.
Future Scope of AI
• The scope of Artificial Intelligence is limited to domestic and
commercial purposes as the medical and aviation sectors are also
using AI to improve their services. If AI is outperforming human
efforts, then opting for AI automation will reduce costs in the long run
for a business.
• Banking- customer support, enhanced data quality, fraud
prevention, digital assistants etc
• Health care and medicine
• Jobs in AI will increase by 74
Issues in the design of search algorithm
1.The direction in which to conduct search (forward versus backward
reasoning). If the search proceeds from start state towards a goal
state, it is a forward search or we can also search from the goal.
2.How to select applicable rules (Matching). Production systems
typically spend most of their time looking for rules to apply. So, it is
critical to have efficient procedures for matching rules against states.
3.How to represent each node of the search process (knowledge
representation problem).
Defining Problems as a State Space Search,
• It has a start state, end state and all states in between the two where
the answer to the problem is to be searched.
• Problem: Question to be solved, It defienes the start, end and all valid
states in between
• State Space search: All possible states are considered in order to find
the goal state with desired property
• Eg: 8 Puzzle Problem
END

This Photo by Unknown Author is licensed under CC BY-SA