COMPUTATIONAL INTELLIGENCE – A DETAILED STUDY OF THE Essay

COMPUTATIONAL INTELLIGENCE – A DETAILED STUDY OF THE PROMINENT PARADIGMS (WITH BIOLOGICAL AND APPLICATION RELEVANCE)

Mr. Shivoham Tiwari [1], Mukul Bhatt [2]

[1] Assistant professor, IT department, IPEM Ghaziabad, U.P, [email protected]

[2]PG Student, IT department, IPEM GHAZIABAD, U.P, [email protected]

ABSTRACT –

“Computational Intelligence (CI), commonly recalled now-a-days as the most powerful advancement in the field of Artificial Intelligence can be denoted as an important subset of AI.”

In simple terms, Computational Intelligence (CI) as a technological expression can be explained as the intelligence or the ability of any computer device for the purpose of learning a pre-determined task with the help of a given observational data or experimental recordings.

With the brief research of CI as an advancement in the field of AI, the paper represents a specific and detailed study of the major and the most prominent paradigms under which Computational Intelligence (CI) works as a successive technology. The paper reflects the basics starting with the Introduction to CI (as many sources till today, have no specified or refined data, even of the accurate definition of CI), continuing with brief description of each major CI paradigm explained along with biological and application relevance of each CI paradigm.

The vision of this paper is to highlight the major CI paradigms in detail with their individual biological relevance they are inspired upon; along with the modern and recent times applications being used through the nature inspired algorithms.

KEYWORDS: Computational Intelligence, Soft Computing, Artificial Neural Networks, Evolutionary Computation, Swarm Intelligence, Artificial Immune Systems, Fuzzy Systems.

________________________________________

1. INTRODUCTION –

According to a certain technological fact, “Technology since the beginning upholds its growth upon discovery, and discovery on the other hand depends upon the advancement of technology”. This has certainly become the case with Computational Intelligence now-a-days. The most common question that particularly arises is “What exactly Computational Intelligence (CI) is?” And, how much difference is there in between Artificial Intelligence (AI) and Computational Intelligence (CI) as both share a similar involving relationship leading towards strong advancements in the field of technology.

Clearing it down through a simple and clear definition, Computational Intelligence (CI) can be explained as the intelligence or the ability of any computer device for the purpose of learning a pre-determined task with the help of a given observational data or experimental recordings. Specifically, CI focuses on some major adaptive mechanisms or techniques with which computers reflect intelligent behaviour in different complex and constantly changing situations. According to [1], Good science produces theories that are explored through experimentation and the experiments depend upon the theories for direction. The discipline of Computational Intelligence (CI) is a new one with some strong and deep ancient roots.

For the world reaching out to a technological milestone with each passing day, CI is relatively a new area of improving technology with the goal of achieving perfection at a glance and bringing change in the current scenario for good. Most relevantly, the sole purpose and the scientific aim of computational intelligence (CI) is and has always been to properly acknowledge all the principles that carry out and succeed in creating intelligence possible in more specifically artificial systems. There are also numerous complexities which can’t be solved or could be procedured with the traditional intelligence approaches like mathematical modelling. The reason for which is, as they might not be able to solve the large complex problems which requires more accurate and advanced approach of processing and handling problems in a precise and exact manner. As, with the reference to [2], many real-life problems cannot be translated into binary language (unique values of 0 and 1) for computers to process it. Computational Intelligence (CI) therefore provides solutions for such problems. The P used are close to the human’s way of reasoning, i.e. it uses inexact and incomplete knowledge, and it is able to produce control actions in an adaptive way. CI therefore uses a combination of ?ve main complementary techniques. These Complementary techniques may be denoted as Computational Intelligence Paradigms.

2. COMPUTATIONAL INTELLIGENCE (CI) PARADIGMS –

Computational Intelligence with the scientific vision of establishing intelligence amongst natural and artificial systems provides accurate and optimum solutions for complex real world problems. All the major paradigms included in accomplishing the purpose of generating Computational intelligence are nearly close to human’s methods of reasoning as it comprises of shuffled bits of informational patterns, inexact knowledge with the capacity of producing actions under control in a procedural and adaptive way.

Therefore, Computational Intelligence uses a combination of 5 major complimentary paradigms which are represented as follows:

1. Artificial Neural Networks (NN)

2. Evolutionary Computation (EC)

3. Swarm Intelligence (SI)

4. Artificial Immune Systems (AIS)

5. Fuzzy Systems (FS)

MAJOR COMPUTATIONAL INTELLIGENCE (CI) PARADIGMS [3]

The figure represented above showcases the main vision of the paper as in addition to the major paradigms of Computational Intelligence (CI), probabilistic techniques are often taken into utilization purposes with Computational Intelligence paradigms. And, with the reference to [3], Computational Intelligence paradigms are based on “Soft computing” methods, a term coined by Lot? Zadeh, is a di?erent grouping of paradigms, which usually refers to the collective set of CI paradigms and probabilistic methods. The arrows indicate that techniques from di?erent paradigms can be combined to form hybrid systems. Moreover, each of the CI paradigms has its origins in biological systems. NNs model biological neural systems, EC models natural evolution (including genetic and behavioural evolution), SI models the social behaviour of organisms living in swarms or colonies, AIS models the human immune system, and FS originated from studies of how organisms interact with their environment.

2.1. ARTIFICIAL NEURAL NETWORKS (NNs) –

Living in the scenario of day-by-day advancements & recognitions in AI, Artificial Neural Networks (NNs) is one certain issue to be strongly identified. Upon its source, NNs are recognised as one of the most prominent techniques used in machine learning. According to [4], as the “neural” part of their name suggests, they are brain-inspired systems which are intended to replicate the way that we humans learn. Neural Networks consists of input and output layers, as well as (in most cases) a hidden layer consisting of units that transform the input into something that the output layer can use. They are excellent tools for finding patterns which are far too complex or numerous for a human programmer to extract and teach the machine to recognize.

2.1.1 BIOLOGICAL BASIS: ARTIFICIAL NEURAL NETWORKS (NNs)

Since the last several decades, Artificial Neural Networks (NNs) prominently established itself as one of the major tools of Artificial Intelligence (AI). And, the only reason of which is the experts that deal with CI research and development work of Artificial Neural Networks (NNs) totally based upon its biological basis and evolution.

With the reference to [3], the basic building blocks of biological neural systems are nerve cells, referred to as neurons which consists of a cell body, dendrites and an axon. A neuron functions either by inhibiting or exciting a signal only when the cell “fires”. An arti?cial neuron (AN) is a model of a biological neuron (BN). Each Artificial Neuron receives signals from the environment, or other ANs, gathers these signals, and when ?red, transmits a signal to all connected ANs. An arti?cial neural network (NN) is a layered network of ANs. An NN may consist of an input layer, hidden layers and an output layer. ANs in one layer are connected, fully or partially, to the ANs in the next layer. Feedback connections to previous layers are also possible.

2.1.2 APPLICATION BASIS: ARTIFICIAL NEURAL NETWORKS (NNs)

On application basis, Artificial Neural Networks (NNs) can be utilised within a wide range of areas. As, according to [2], neural networks can be classi?ed into ?ve groups which are as follows:

1. Data analysis

2. Data classi?cation

3. Associative memory

4. Clustering generation of patterns

5. Control

And, with the proper classification and development; various kinds of neural networks have been arranged for usage purposes in a wide range of applications. As, with the inclusion of the numerous, following are some of the applications that are widely used now-a-days in the field of Information Technology, Medical science and in music and gaming industries:

1. Disease diagnosis

2. Data mining

3. Music composition

4. Speech synthesis and recognition

5. Image processing

6. Pattern recognition

7. Game strategies planning, and many more.

2.2. EVOLUTIONARY COMPUTATION (EC) –

A sub-field of Artificial Intelligence (AI) closely associated with Computational Intelligence (CI), Evolutionary Computation involves numerous of combinational optimization problems, queries and constant optimization engaged in problem-solving natural or artificial systems with the implicational use of computational models and patterns with evolutionary procedures and processes (taken as the major design elements).

The main and foremost objective of this paradigm is to mimic or copy the processes or the various procedures from natural evolution processes (which reflects the key idea “Survival” or the concept of showcasing “only the fittest survives, the rest [mainly weak] die”). Also, completely based upon the techniques and process of Natural selection (which was first and foremost introduced by Charles Robert Darwin), evolutionary computational methodologies comprises the concept of capitalizing (on the certain strength of natural evolution); for the purpose of introducing new artificial evolutionary computational methodologies to solve a large and wide variety of complex problems.

2.2.1 BIOLOGICAL BASIS: EVOLUTIONARY COMPUTATION (EC)

Evolutionary Computation is generally based upon the theory of biological evolution (Natural selection, inheritance basis on the genes to name a few). It mimics the concept of natural or biological processes and then intakes a certain group of problem-solving strategies to be applied to numerous problems or complexities. Moreover, the complexities or problems to be solved, belong to a wide range of practical industrial applications.

With reference to [3], in natural evolution, survival is achieved through reproduction. O?spring, reproduced from two parents (sometimes more than two), contain genetic material of both (or all) parents – hopefully the best characteristics of each parent. Those individuals that inherit bad characteristics are weak and lose the battle to survive. This is nicely illustrated in some bird species where one hatchling manages to get more food, gets stronger, and at the end kicks out all its siblings from the nest to die. Most relevantly, evolutionary algorithms on the same side, use a population of individuals, where an individual is referred to as a chromosome and the survival strength of an individual is measured using a fitness function which re?ects the objectives and constraints of the problem to be solved.

2.2.2 APPLICATION BASIS: EVOLUTIONARY COMPUTATION (EC)

Evolutionary computational techniques (based upon the biological evolution processes) in the modern era establishes its place in numerous range of applications ranging from industrial applications like analytics, algorithms (based over predictions), etc. to scientific research applications like the one named as “Protein folding”. Specifically, on its implementation over the real-world problems; evolutionary computation techniques have numerous applications that can be understood by the following data:

1. Combinational optimization

2. Data mining

3. Time series approximation

4. Fault diagnosis

5. Data clustering

6. Data Classification

7. Scheduling of data, and many more.

2.3. SWARM INTELLIGENCE (SI) –

“The reality of nature reflects the importance of even simplest of creatures, of how through simple rules and norms; the small creatures display an astonishing amount of creativity and efficiency by solving complex problems.” Swarm Intelligence, relatively recognised as Collective Intelligence is an important branch of Computational Intelligence (CI) which deals with the discussion, research, and developments made through by the complete study and observation of collective behaviour and responses emerging out within or from self-organising societies and groups of natural agents.

Swarm intelligence collectively represents a trait that can be observed throughout the nature and its constituents. But in the recent times, numerous of biologists, researchers or natural scientists have begun its utilisation for the purpose of transforming numerous of fields (including data mining, robotics, medicine, etc.). As, reference to [3], Swarm intelligence (SI) originated from the study of colonies, or swarms of social organisms. Studies of the social behaviour of organisms (individuals) in swarms prompted the design of very e?cient optimization and clustering algorithms. For example, simulation studies of the graceful, but unpredictable, choreography of bird ?ocks led to the design of the particle swarm optimization algorithm, and studies of the foraging behaviour of ants resulted in ant colony optimization algorithms.