Welcome to MIND of YOUTH

BE THE CHANGE ! MAKE A DIFFERENCE IN THE LIVES OF YOUTH..........

BRANCHES OF ARTIFICIAL INTELLIGENCE

BRANCHES OF ARTIFICIAL INTELLIGENCE

branches of artificial intelligence

BRANCHES OF ARTIFICIAL INTELLIGENCE

Computational Creativity

Fuzzy Systems

Evolutionary Computation, including evolutionary algorithms and genetic algorithms

Probabilistic methods: Bayesian Network, Hidden Markov Model

Chaos Theory.

 

COMPUTATIONAL CREATIVITY

ARTIFICIAL INTELLIGENCE AND THE ART OF CLEVERNESS IN COMBINING.

Computational creativity is also known as artificial creativity, mechanical creativity, creative computing, and creative computation.

CREATIVITY: The word simulates to our brains. As the human brain is equipped by evolution, to invent, to write, to make. THEREFORE “humans are special” they have art to make music. But wait, what if a robot is asked to do the same?

Over some 40,000 years human creativity had exploded drawing on cave walls, seek to use both sides of your brain, cultivate a thrust for knowledge, explain things back to yourself, take breaks to switch on the creative side.

Let your imagination run wild and moreover centuries to come for advanced technology.

Can a robot truly imagine or it can be programmed to do masterwork??

Can also artificial intelligence imagine, feel emotion, and turn it into art and algorithm with various options like  YES OR NO. And have potential to unlimited creativity and unimagined creation??

 

HENCE, Why computational creativity?

  • To produce artificial systems appearing to be creative, to some degree.
  • For better understanding of human creativity.
  • To design programs that can support and enhance human creativity.
  • Also to explore the potential of cognitive architectures.

 

FUZZY SYSTEMS

AN INTRODUCTION TO THE WORLD OF FUZZY LOGIC

What is fuzzy logic?

To understand fuzzy logic, one needs to understand traditional logic.

TRADITIONAL LOGIC:-

Something which can be represented by either value true which is one OR false which is zero.

Traditional logic                                                                  Fuzzy logic

True:- 1                                                                                  True:-  0.7

False:-0                                                                                   False:- 0.3

Let us take an example:-

Example 1:-Tap Water Temperature

Traditional logic:-

As per traditional logic, cold water will be represented as “zero” and hot water as “one”

Fuzzy logic:- In fuzzy logic system rise in temperature will not be in terms of 0 and 1 it will be in form of fractions.

Here in the following strip BLUE strip represents “COLD” and red represents “HOT”.

Example 2:- AUTOMATIC BRAKING SYSTEM

Traditional Logic

 

Is car close : 0 or 1 (no or yes)

brakes : 0 or 1 (off or on)

In the above situation

A Car is not close

therefore is car close? : 0 (off)

brakes : 0(off)

IN NEXT SITUATION

Is car close: 1 (YES)

Brakes: 1 (ON)

In the above situation, there may be a problem that car will apply brakes with full pressure. And if not then there is a risk of breakages hence from here FUZZY LOGIC SYSTEM came into the picture.

FUZZY LOGIC SYSTEM

Is car close? : 0-1 (Range of no to yes )

Brakes? : 0-1(Range of off to on)

Is car close?  : 0.8 (pretty close)

Brakes? : 0.8(fairly heavy pressure)

 

DIFFERENCE BETWEEN TRADITIONAL CONTROLLERS AND FUZZY LOGIC CONTROLLERS

TRADITIONAL CONTROLLERS:-

  •  In this one need to know detailed physical properties of the system.
  • Most systems are too complex and have to be idealised to develop a traditional controller.
  • This is not a free representation of a system.
  • The condition when the traditional controller will work is usually  fairly limited.

FUZZY LOGIC CONTROLLERS:-

  • Don’t need much detailed knowledge of the system.
  • What the controller should do is determined by linguistic rules.
  • If optimisation tools are used like Genetic Algorithms can get away with not knowing much of anything.
  • The system doesn’t need to be reduced or idealised to develop a      working fuzzy logic controller.
  • The conditions when Fuzzy Logic Controller will work are much more robust because they can account for more variability in the inputs.

WHAT IS THE POTENTIAL OF FUZZY LOGIC CONTROLLERS

  1. No of applications virtually limitless. (control systems, kitchen appliances, car control systems, aerospace vehicle control systems)
  2. Image processing.
  3. Decision making.
  4. Route planning.
  5. Still in development and testing for safety critical system.
  6. Has already been tested in some instances like small satellite attitude control.

 

EVOLUTIONARY COMPUTATION

Being Computer science engineer, When I see stuff like this which has colours and attraction or tingled network. This is exciting as well but the real excitement is the way to solve this puzzle.

Which is not that easy one needs a great motivation for that which comes from the development of fruit flies brain. As the brain develops the chemical signalling which type of signal is going. And what type of brain cell will be generating which type of signal.

 

What concept lies behind the actual evolution?

Now let us talk about nature the main beauty lies on the surface and below the surface there are layers and layers. Biological processes running under very special care processes every min, hour, year and centuries. Various different processes combine to create beautiful patterns.

A lot of different patterns are produced using different processes and there is a common theme. A common concept lying behind this that is THE CONCEPT OF EMERGENCE.

Emergence system

Emergence system is a system in which there is a global pattern which comes about. There is nothing in the system which understands the global pattern all individuals go on the action.

There is a fundamental difference where nature builds things and where the man builds things. We people get inspiration from nature which gives us different biological patterns known as bio-mimicry.

If we look computer and computer applications it seems far from nature. But going back in early days we actually find biology and psychology as really important topics developed for computing which envision life.

So for many computer science engineers going back to the basic computer does not mean going back to nature.

For example:-

AVEN THEVEY, He considered as a gentleman to make the computational theory a legitimate science. But few people realized that he was also involved in biology.

He discovered that by combining chemical reactions in diffusion through loco processes one can generate many different types of biological patterns.

 

WHAT IS ARTIFICIAL LIFE?

In artificial intelligence computers in one room and human in other. They can see each other and can communicate back and forth if the human actually thinks that the human communicating with another human.

Some great personalities who showed their great contribution in artificial intelligence

1.John Von Neumann:(1903-1957)

 

 

 

 

 

 

 

 

He was a great mathematician who made an important contribution in the area of artificial life. His main work on the artificial life is based on self-reproduction of machines.

2. Nils Barricelli(Oslo University)[1968-1993]:-

Nils Barricelli, he traveled in other states like Europe where he worked on simulations, reproduction and evolution on a the computer.

Many different inventions we see like air-planes, air crafts, various display technologies based on butterfly wings or water purification techniques.

Now scientists are looking for inventions at the deeper level not just the product of evolution but also the process of evolution. Various algorithms are developed such that they looks like natural selection.

Let us begin with the initial population to a problem you send it to a performance test each of them goes through a score’s test one with higher score its chances of reproduction is more often, you lose yourself in this way we go round and round and eventually one get the solution to the problem.

EXAMPLES OF EVOLUTIONARY COMPUTATION:-

Evolutionary art(Steven Rooke)

 

 

 

 

 

 

Picbreeder.org (ken Stanley)

Andrew Lincoln Nelson(2006)living robot

Nero evolving Robotic operations(NERO)

Chris robots, Crab lab, NTNU(2013)

 

PROBABILISTIC METHODS

A. PROBABILISTIC REASONING:-

Before this, first, we need to know why probabilistic reasoning?

  • It is because monotonic reasoning and non-monotonic reasoning are not sufficient to handle every situation.
  • In Monotonic Reasoning, we talk more about logic and in logic we talk either logic will be 100% true or 100% false i.e this logic was based on Boolean Algebra i.e true =1 or falsely represented as false=0.
  • But in real life scenario reasoning are not like 100% sure. So for such type of reasoning, probabilistic reasoning is used.

What probabilistic reasoning actually is?

  • Probability provides a way to handle uncertainty that comes from one’s laziness and ignorance.
  • Using logic and probability to handle the uncertain situation.
  • Probability-based reasoning is the same as understanding from the knowledge that, how much uncertainty is present in that.
  • Probabilistic Reasoning is used when outcomes are unpredictable.

Example:-

The Doctor examines a patient, his history, Symptoms based on test result. Example:- 90% chance disease will be cured, but some result or test might be missing.

One way to express such an event is the probability.

Sources that cause Uncertainty  Information:-

1. Information is obtained personally.

2. Experimental errors.

3. A random event occurs in the major event.

4. Equipment fault.

5. Temperature variation or climatic changes occurred.

(a)BAYESIAN NETWORK—–>

Bayesian Belief Network:-

  • It is a type of classifier which does not show any type of dependency between attributes.
  • It is a condition independent attribute, i.e no attribute is related to each other.
  • The namespace is a type of classifier in which one attribute is dependent nature on other for giving its output.
  • But if we talk about Bayesian network this dependency ends, but in real life if we talk about dependent nature,one attribute is related to other in one or other way then QUESTION ARISES HOW BAYESIAN NETWORK WILL WORK.

This is possible through Joint Probability:-

Joint probability is a type of probability which is based on some condition, which is

p(ATTRIBUTE|PARENT)

1. Phase directed to a-cyclic graph.

2. Conditional probability table(CPT)

 

Phase directed a-cyclic graph:-

           In this, we have various nodes and other connectivity in that.

The above figure shows that an that in a family history a member suffering from Lung Cancer and shows positive x-ray , also a smoker has emphysema and dyspnea, a smoker will also suffer from lung cancer and a person suffering from lung cancer will suffer from dyspnea too but at these levels also one level is  independent with another level.

In actual dataset we have two nodes:-

Parent(Immediate predecessor of z)       y->z      (Generate descendent of y)

 
 CPT(Conditional Probability Table)

Take various variables as FH= family history, S=Smoker, LC=Lung Cancer.

               FH,S                           FH,~S                        ~FH~S                        ^FH ~S


LC           0.1                                 0.8                              0.5                               0.7                                   


~LC         0.9                                0.2                                0.5                             0.3                                   


p(Xi|Parent (yi))

P[LC =’Yes’| FH=yes & Smoker =’Yes’ =0.8]

P[LC =’No’| FH=no & Smoker =’No’=0.9]

B. HIDDEN MARKOV MODEL(HMM)

  • It is a statistical model, which is used to generate the model.
  • Model generation based on some set of the input sequence.
  • These input sequences are S1,S2,S3,S4.

A PROPERTY OF HIDDEN MARKOV MODEL

Must required Markov property is:-

  1. It is memory-less,Why it is memory-less? because it considers only 1 state.
  2. Its soul part includes- future’s prediction relies on present state.

HMM is more or less similar to FINITE STATE MACHINE:-

TYPES OF STATE ARE:-

  1. Hidden state(w)
  2. Visible State/Output state(v)

Example model:-

 

   CHAOS THEORY (1903)

The world of classical science is one of the simple systems. For Example, we have two magnets that are aligned to each other, push one and other moves in a predictable way. What if there were five or twenty magnets? HOW would they behave? Were there movements be predictable?

It is a matter of faith, the scientific world of the 90th century, if one of the magnets is moved how the other will behave is needs to predict?

In 1903 french mathematician, Olly Pinckney studied and said about the three body problems. Predicting the movement of our simplified solar system consisting of sun, moon, and earth, governed by the laws of motion.

What he found that even this system behaves in an unpredictable way and the three-body problem was unsolvable. Understanding this and other more complex systems such as earth’s weather, would go in another direction, that direction was the new brand of science known as KR SERIES.

What is KR series?

A series whose main premise was that small change in initial conditions could result in the vast difference in the final outcome. The KR series remains curious into the advent of computer modeling in the 1960s.

What is butterfly effect?

In 1961, Advent servlet was modeling the earth’s weather with the computer and found that the small change, at the one million addressable points, can make his prediction useless, one can call this result a BUTTERFLY EFFECT.

And in the maze, a butterfly flaps his wings, in two years later, an under stone appears over cancers no flaps, no stone, the butterfly effect is everywhere around this.

Changes in KR series:-

Later, in 1970s Bernard Mendel added fructose to KR SERIES. Fructose is bi-product, the pattern’s left behind by dynamically changing systems. Fructose is self simulators, illustrated by these designs.

Where inherent property as with could or could not is that no matter how close or how far away a pattern gets, the basic pattern is always the same. At the end of 20th century, the Bernard had come to a conclusion, the long row to discover the simplicity of everyday life.

Besides using a computer one could start with a simple system such as color squares on the boat are change by the simple known rules. In time life like complexes emerge self-organized clusters come to play and interact with each other, structures are formed that are always on the edge of seven incredible change. This is the way, in fact, the way eco-system, human societies, and the society behave.

At, the beginning of 21st century KR series combined it with the quantum mechanics. The most promising area perhaps in the area of artificial intelligence. Quantum theory is reductively reflecting the unpredictability of human behavior.

Monika Khilrani

Hii, I am Monika Khilrani amazed by Computer Science, I am a technology enthusiast, author and founder of www.mindofyouth.com, writing is not only my profession, it is my passion too because DATA=Knowledge + experience.