AI NOTES

- Developed in 1943

What is Artificial Intelligence?

 Branch of science that deals with making intelligent machines

 machines which can provide solutions to any complex problems
 So that we can have better solutions for complex problems
 They borrow certain charecteristics from human beings
 It is a flexible approach to give solutions depending upon the environment
 Same object, same person, the same object provides different results

-What is AI? -> Creating a machine which has the ability to think

- Automation of various activities

- Art of creeating machines which have decision making capabilty

-> weather forcast

- Study of mental abilities

-> Study of computations that can precieve, reason and act

- Field of study that seeks intelligence in terms of computational process

-> Now we can say that This is a field of CS, Which deals with computations

-> anything that is not possible in / dont know how to do it properly

- AI, developed first by John McCartly

 Goals of AI
Scientific Knowledge

1. Scientific= Develop some kind of knowledge based system

2. Engineering goal- Various types of applications we can build, knowledge based system
3. Game playing

Engineering goal

1. Learning system- What is learning? to enhance our understanding on the base of past experience,

or enhance decision making process

2. NLP based system= Natural language Processing (based application)

3. Real world problem solving

- knowledge based system

- common sense reasoning system

- learning system

- Natural language understanding system

- Intelligent Robots

- Speech and voice Recognition

- game play

 Foundations of AI
 Foundation defined the structure

 Philosophy
 Deserve the nature and environment of human being
 Studying of human nature and how that person is going to react to any situation

 Mathematics :- logical and probability

 Factual based truth
 Certain reasons behind that
 Probability/Possibillity :- best possible outcome to any situation to a problem

 Computer Engineering:- Creation of algorithm on the basis of screnario

 Fully automated or not
 Legally justified or not
 Linguistics:- NIP [Natural language processing]
 Google translator

 Behavioural characteristics: - parameters on that particular situation should be followed that

are the minimum thing we want from model how to react on particular situation
  Application of AI

HISTORY OF AI

1943

- Model for Artificial neurons

1950

- Tuning test is for computers as IQ test for humans

1955

- First artificial intelligence program/ algorithm

1956

- Formly AI is defined by John Mckinly

1966

- AI algorithm that has capability to solve maths equation

- First chatbot was developed

1972

- First Robot was developed in Japan

1980

- Fast expect system was developed

1993

- AI enters a new field namely gaming

2002
 - Home automation: automated vacuum cleaner

2006

- AI enters the business world, Facebook, Netflix, YouTube

2012

- Enters predictions

2015

2018

Problem solving in AI
What is problem? A situation which requires a solution

 States
- Initial state
- Goal state
- Failure state/ impossible state : not reaching desired destination

 Step for problem solving

 Goal formulation
 Problem formulation
- Initial state: beginning state
- Successor function
- Goal state test
- Path cost calculation: calculate the cost

 Choosing best possible solution: suggestions provided to solve a given problem

 Algorithm for problem solving

 Analyze the problems to define initial and goal state
 Find the data related to problem
 Production riles must be defined
  Specify the rule for initial state
 Specify the riles for next succession state
 Information about the goal state

 Procedural Technique
 It implements numeric processing
 Implements algorithm search technique
 Solution steps are explicit
 Knowledge is precise
 Modifications are rare
 Works on implement repetition mechanism
 It has large database
 Results are usually optimal

 AI Techniques
 Implements symbolic reasoning processing
 Implements different search techniques
 Sol states are explicit
 Knowledge is imprecise
 Require frequent modifications

 Components of AI
 Agent
 AI

 Types
i) Simple reflex agent
 Decision according to the environment take decision on basis of current scenario and
ignored the rest & previous data & previous scenarios

 Main component of simple reflex agent sensors, components action rule activation
perfoms action
ii) Limitations of simple reflex
 Works only on observed components
 Intelligent level is less, take decision on the basis of only given condition
 Does not hold or store any info, no part experience, no enhancement in decision
 Always modify the condition action rule on the changing with environment

 Goal based agent

 Obj orientated
 Just to expand the model based agent with goal
 Enhance the decision making capability of model based reflex age
 In this just check objective

 Utility based agent

 Enhancement in goal based agent it checks whether decision takes beneficial or not,
how much effective decision it is

 Learning agent
 Learns from past experience
 Reduce chanvces of mistakes
 Provides data to external environment/source
 Also works on external environment

 Intelligent agent
 Which helps in correct decision making
 Will provide correct decision only when other agents provide correct info, so we can say
that all the agents are intelligent agents

 Min max algo

 Recursive backtracking algo used for decision making games
 Game theory it provides optimal move for a player
 Best move strategy is used
 Max will try to make maximize its utlity( move, which is best)
 Min will minimize its utility (worst move)

 Game Tree

 Difference between AI technique and traditional technique

Date- 17/05/23

STATISTICS

 What is stats?
 Mean, median mode, standard deviation, variation, correlation, frequency distribution
 Dataset
 Prediction of future value
 Description of my dataset, then we use all the given statistics
 Dataset -> p

 Steps for SD
 Calculate the mean value
 Find the difference between mean and each value
 Square of each diff
 Calculate sum of squares
 Calculate the variation ( sun of sequence)/n-1
 Calculate the standard deviation sqrt variation
 Eg (92, 95, 85, 80, 75, 50)

 Correlation

X1 X2 X3 X4 X5 X6 X7 Output
value

 Finding out the relations between the variables

Different libraries

 Numpy - is a library for arrays

 Matplotliib- to generate graphs
 Pandas- whenever we use data sets

 read.csv()

 df.head()
 df.tail()
 data = pd.read.csv(“sommu”)
dataframe
 df.columns
 df.drop() = deletes values
  df= df.drop(‘x3’,’x2’) =

Min, max, mean, SD, variation, datatype

 Df.dtypes() -> this displays the attributes of datatypes

 Df.select[“x3’].datatypes -> this shows attributes of given datatype
 Df.replace() -> to replace
 Df.count() =
 Df.rename() = to rename column
 Df.fillname() =
 Df = df.query(‘X3>65”)
 D

AI is the study of complex info processing problems that often have their roots in some aspect of
biological information processing. The goal of the subject is to identify solvable and interesting
processing problems, and solve them

AI is the design study and construction of computer program that behave intelligently

AI is the enterprise of construction a physical symbol system that can be reliably pass the turing test