What is Soft Computing? Soft computing is an approach whe e compute solutions to the existing complex problems, where output results are imprecise or fuzzy in nature, one of the most important features of soft computing is it should be adaptive so that any change in environment does not affect the present process. The following are the characteristics of soft computing. elt does not require any mathematical modeling for solving any given problem e It gives different solutions when we solve a problem of one input from time to time e Uses some biologically inspired methodologies such as genetics, evolution, particles swarming, the human nervous system, etc. ¢ Adaptive in nature.or fuzzy in nature, one of the most importan features of soft computing is it should be adaptive so that any change in environmen does not affect the present process. The following are the characteristics of soft computing. elt does not require any mathematical modeling for solving any given problem e It gives different solutions when we solve a problem of one input from time to time e Uses some biologically inspired methodologies such as genetics, evolution, particles swarming, the human nervous system, etc. e Adaptive in nature. There are three types of soft computing techniques which include the following. e Artificial Neural Network e Fuzzy Logic e Genetic algorithm aDifference Between Hard Computing and Soft Computing The difference between hard computing and soft computing are as follows Hard Computing Soft Computing ¢ The analytical e Itis based on model required by —_ uncertainty, partial hard computing truth tolerant of must be precisely imprecision and represented approximation. Computation time » Computation time is is more less « It depends on ¢ Based on binary logic, approximation and numerical systems, dispositional. crisp software. Sequential ¢ Parallel computation computation ¢ Gives exact output « Gives appropriate output nl OB¢ Examples: ¢ Example: Neural Traditional networks like methods of Adaline, Madaline, computing using ART networks, etc. Our personal computer. Advantages The benefits of soft computing are ¢ The simple mathematical calculation is performed ¢ Good efficiency ¢ Applicable in real-time e Based on human reasoning. Disadvantages The disadvantages of soft computing are e It gives an approximate output value If a small error occurs the entire system stops working, to overcome its entire system must be corrected from the beginning, which is time taking process.T—S'XCOC”~C~C#é€C*st PR ® chanakyaresearch.net Visit Site Need of soft computing Sometimes, conventional computing or analytical models does not provide a solution to some real- world problems. In that case, we require other technique like soft computing to obtain an approximate solution. ° Hard computing is used for solving mathematical problems that need a precise answer. It fails to provide solutions for some reallife problems. Thereby for real-life problems whose precise solution does not exist, soft computing helps. 0 When conventional mathematical and analytical models fail, soft computing helps, e.g., You can map even the human mind using soft computing. ° Analytical models can be used for solving mathematical problems and valid for ideal cases. But the real-world problems do not have an ideal case; these exist in a non-ideal4:33 AM | 0.6KB/s © errr Hard computing is used for solving mathematical problems that need a precise answer. It fails to provide solutions for some real-life problems. Thereby for real-life problems whose precise solution does not exist, soft computing helps. When conventional mathematical and analytical models fail, soft computing helps, e.g., You can map even the human mind using soft computing. Analytical models can be used for solving mathematical problems and valid for ideal cases. But the real-world problems do not have an ideal case; these exist in a non-ideal environment. Soft computing is not only limited to theory; it also gives insights into real-life problems. Like all the above reasons, Soft computing helps to map the human mind, which cannot be possible with conventional mathematical and analytical models. Daet alnarithaiminr AARC+. there is no general convergence theory applicable to practical, multidime problems ‘mt 1.6 Classification of . © Classification of Artificial Neural Network are FeedForward and Feedback 11. Feedforward ANN output layer and one or more layers of neurons. « The information flow is unidirectional. A unit sends information to other unit from which it does not receive any information. There are no feedback loops. * Fig. 1.6.1 shows feedforward ANN Input Layer Hidden Layer Output Layer Fig. 1.6.1 : Feedforward ANN + Through evaluation of its output by reviewing its input, the power of the network can be noticed base on group behavior of the connected neurons and the output is decided. + The main advantage of this network is that it leams to evaluate and recognize input patterns. TECHNICAL PUBLICATIONS ® « An up thrust for knowledge Soft Computing (1-14) Introduction to Soft Computing 0 2. Feedback ANN « In these types of ANN, the output goes back into the network to achieve the best- evolved results internally. «Ina basic structure, a feedback ANN consists of a set of processing units, whose outputs are fed back as inputs to all other units of the same layer including the same unit. «Fig. 1.6.2 shows Feedback ANN. Fig. 1.6.2 : Feedback ANN In its basic form, a feedback ANN doesn’t have any structure and hence it is not a useful at pattern recognition. + The feedback network feeds information back into itself and is well suited to solve optimization problems. Feedback ANNs are used by the Internal system error corrections,ee ee ar ee ee eT Ne ee Le Se Sag po 3. Fill in skeletal or incomplete specifications about a domain ‘=> 1.8.1 Supervised Learning ‘+ Supervised learning is the machine learning task of inferring a function from supervised training data The training data consist of a set of training examples. The task of the supervised leamer is to predict the output behavior of a system for any set of input values, afier an initial training phase. ‘* Supervised learning in which the network is trained by providing it with input and matching output pattems. These input-output pairs are usually provided by an external teacher. ‘+ Human leaming is based on the past experiences. A computer does not have experiences, + A computer system leams from data, which represent some “past experiences” of an application domain * To learn a target function that can be used to predict the values of diserete class attribute, €.8., approve or not-approved, and high-risk or low risk. The task is commonly called : Supervised learning, Classification or inductive learning. + Training data includes both the input and the desired results. For some examples the correct results (largets) are known and are given in input to the model during the learning process. The construction of a proper training, validation and test set is crucial. These methods are usually fast and accurate TECHNICAL PUBLICATIONS © Anup wus for krowoage ‘Soft Computing (ee) Introduction to Soft Computing ‘* Have to be able to generalize : give the correct results when new data are given in input without knowing a priori the target. ‘Supervised learning is the machine leaming task of inferring a function from supervised training data. The training data consist of a set of training examples. In supervised learning, each example is a pair consisting of an input object and a desired output value. ‘A supervised learning algorithm analyzes the training data and produces an inferred function, which is called a classifier or a regression function. Fig. 1.8.1 shows supervised leaming process. Training Testing Fig, 1.8.1 : Supervised learning process ‘* The learned model helps the system to perform task better as compared to no learning. ‘* Each input vector requires a corresponding target vector. Training Pair = (Input Vector, Target Vector) Fig. 1.8.2 © Supervised leaming denotes a method in which some input vectors are collected and presented to the network. The output computed by the net-work is observed and the deviation from the expected answer is measured. The weights are corrected according to the magnitude of the error in the way defined by the learning algorithm. © Supervised leaming is further divided into methods which use reinforcement or error correction. The perceptron leaming algorithm is an example of supervised learning with reinforcement. * Inorder to solve a given problem of supervised learning, following steps are performed 1. Find out the type of training examples. 2. Collect a training set. TEGHNIGAL PUBLICATIONS up thst for knomldge SS ‘Soft Computing (1-18) Introduction to Soft ComputingSupervised learning, Classification or inductive leaning. © Training data includes both the input and the desired results. For some examples the correct results (targets) are known and are given in input to the model during the learning process. The construction of a proper training, validation and test set is crucial. These methods are usually fast and accurate TECHNICAL PUBLICATIONS ® - An up thrust for knowledge Soft Computing (1-17) Introduction to Soft Computing © Have to be able to generalize : give the correct results when new data are given in input without knowing a priori the target. © Supervised learning is the machine learning task of inferring a function from supervised training data. The training data consist of a set of training examples. In supervised learning, cach example is a pair consisting of an input object and a desired output value. © A supervised learning algorithm analyzes the training data and produces an inferred function, which is called a classifier or a regression function. Fig. 1.8.1 shows supervised learning process. Training Testing Fig. 1.8.1 : Supervised learning process ‘© The learned model helps the system to perform task better as compared to no learning. © Each input vector requires a corresponding target vector. Training Pair = (Input Vector, Target Vector) Fig. 1.8.2 © Supervised leaming denotes a method in which some input vectors are collected and presented to the network, The output computed by the net-work is observed and the deviation from the expected answer is measured. The weights are corrected according to the magnitude of the error in the way defined by the learning algorithm. © Supervised learning is further divided into methods which use reinforcement or error correction. The perceptron leaning algorithm is an example of supervised learning with reinforcement. ‘© In order to solve a given problem of supervised learning, following steps are performed 1. Find out the type of training examples. 2. Collect a training set. TECHNICAL PUBLICATIONS ® - An up thrust fr knowledge Soft Computing (1-18) Introduction to Soft Computing 3. Determine the input feature representation of the learned function 4, Determine the structure of the learned function and corresponding learning algorithm. 5. Complete the design and then run the learning algorithm on the collected training set. 6. Evaluate the accuracy of the learned function. After parameter adjustment and learning, the performance of the resulting function should be measured on a test set that is, separate from the training set.1.8 Review Questions 1.9 Bibliography, References and Further Reading 1.0 Objectives In this chapter, we will try to learn what is soft computing, difference betwe¢n hard computing and soft computing and reason for why soft computing evolvgd. end, some application of soft computing will be discussed. 1.1 Computational Paradigm { ietpeens ea Conpaing = aaa [Taal [igre | etter Sesh a Search) | Remserine Figure 1.1: Computational Paradigms SOFT COMPUTING TECHNIQUES Computational paradigm is classified into two viz: Hard computing and soft computing. Hard computing is the conventional computing. It is based on the principles of precision, certainty, and inflexibility. It requires mathematical model to solve problems. It deals withs the precise models. This model is further classified into symbolic logic and reasoning, and traditional numerical modelling and search methods. The basic of traditional artificial intelligence is utilised by these methods. It consumes a lot of time to deal with real life problem which contains imprecise and uncertain information. The following problems cannot accommodate hard computing techniques: 1. Recognition problems 2. Mobile robot co-ordination, forecasting 3. Combinatorial problems Soft computing deals with approximate models. This model is into two approximate reasoning, and functional optimization & methods. It handles imprecise and uncertain information of thg be used in all industries and business sectors to solve probl can be designed with soft computing to deal with the incomple where the system behaviour is not completely known or the existence of measures of variable is noisy.(hap nts To So Computing 1.2 Introduction to Soft Computing ‘The real-world problems require systems that combines knowledge, techniques, ‘and methodologies from various source. These systems should possess humanlike ‘expertise within specific domain, adapt themselves and leam to do better in the changing environments and explain how they make decisions or take actions, Natural language is used by human for reasoning and drawing conclusion. In conventional Al, the human intelligent behaviour is expressed in the language form ‘or symbolic rules. It manipulates the symbols on the assumption that such behaviour can be stored in symbolically structured knowledge base known as physical symbol system hypothesis. ‘Basically, Soft Computing is not a homogenous body of concepts & techniques. Rather. it is partnership of distinet methods that in one way or another conform to its guiding principle. At this juncture, the dominant aim of soft computing is to exploit the tolerance for imprecision and uncertainty to achieve tractabitits, robustness and low solutions cost. The principal constituents of soft computing are fuzzy logic, neurocomputing and probabilistic reasoning, withthe later subsuming ‘genetic algorithms, belief networks, chaotic systems, and parts of learning theory. In partnership of fuzzy logic, neurocomputing, and probabilistic reasoning, logic is mainly concerned with imprecision and approximate reasoning: neurocomputing with learning and curve-fiuing: and probabilistic reasoning with uncertainty and belief propagation.” ~Zadeh (1994) Soft computing combines different techniques and concepts. It can handle imprecision and uncertainty. Fuzzy logic, neurocomputing, evolutionary and genetic programming, and probabilistic computing are fields of soft computing, Soft computing is designed to model and enable solutions to real world problems, ‘which cannot be modelled mathematically. It docs not perform much symbolic manipulation, ‘The main computing paradigm of soft computing are: Fuzzy systems, Neural Networks and Genetic Algorithms. # Fuzzy set are for knowledge representation via fuzzy If— Then rules. + Neural network for leaming and adaptivity and Genetic algorithm for evolutionary computation. (Ques To achieve close resemblance with human like decision making, soft computing aims to exploit the tolerance for approximation, uncertainty, imprecision, and partial truth ‘© Approximation: the mode! has similar features but not same. © Uncertainty: the features of the model may not be same as that of the entity/belief. ‘© Imprecision: the model features (quantities) are not same as that the real ones but are close to them. 1.3 Premises of Soft Computing ‘© The real-world problems are imprecise and uncertain, ‘© Precision and certainty carry a cost. © There may not be precise solutions for some problems. 1.4 Guidelines of Soft Computing The guiding principle of soft computing is to exploit the tolerance for approximation, uncertainty, imprecision and partial truth to achive tractability, robustness and low solution cost. Human mind is the role model for soft computing.approximation, uncertainty, imprecision and partial truth to achieve tractability, robustness and low solution cost. Human mind is the role model for soft computing. 1.5 Uncertainty of AI . Objective (features of whole environment) ° There are lot of uncertainty in the world. We have limited capabilities to sense these uncertainties. . Subjective (features of interaction with concrete environment o For the same/similar situation people may have different experiences. This experience maps on the features of semantics of different languages.3. Distributed representation of knowledge over the connections 4, Knowledge is acquired by network through a leaming process ‘™ 1.4 Constituents of Soft Computing ‘* The principal constituents, ic., tools, techniques, of Soft Computing (SC) are Fuzzy Logic (FL), Neural Networks (NN), Support Vector Machines (SVM), Evolutionary Computation (EC), and Machine Learning (ML) and Probabilistic Reasoning (PR), ‘+ Fuzzy theory plays a leading role in soft computing and this stems from the fact that human reasoning is not crisp and admits degrees. ‘= 1.4.1 Fuzzy Logic ‘+ Fuzzy Logic (FL) is a multivalued logic, that allows intermediate values to be defined between conventional evaluations like true/false, yes/no, high/low, etc. ‘+ Fuzzy Logic is based on fuzzy set theory and provides methods for modeling and reasoning under uncertainty, a characteristic present in many problems, which makes FL a valuable approach. ‘+ It allows data to be represented in intuitive linguistic categories instead of using precise (crisp) numbers which might not be known, necessary or in general may be too restrictive TECHNICAL PUBLICATIONS ® - An uphrst or krowiedge Soft Computing (1-8) Introduction to Soft Computing ‘© Fuzzy logic offers a practical way for designing nonlinear control systems. It achieves nonlinearly through piece-wise linear approximation. The basic building blocks of a fuzzy logical control system are set of fuzzy if-then (i.e, fuzzy rule based models) that approximate a functional mapping. ‘* Fuzzy logic provides an inference morphology that enables approximate human reasoning capabilities to be applied to knowledge-based systems. The theory of fuzzy logic provides a ‘mathematical strength to capture the uncertainties associated with human cognitive processes, such as thinking and reasoning, ‘+ Fuzzy systems are suitable for uncertain or approximate reasoning, especially for the system with a mathematical model that is difficult to derive. ‘+ Fuzzy logic allows decision making with estimated values under incomplete or uncertain information. ‘* Fuzzy logic is viewed as a formal mathematical theory for the representation of uncertainty. Uncertainty is crucial for the management of real systems : if you had to park your car precisely in one place, it would not be possible. Instead, you work within, say, 10 em tolerances. ‘+ The presence of uncertainty is the price you pay for handling a complex system. Nevertheless, fuzzy logic is a mathematical formalism, and a membership grade is a precise ‘number. What's crucial to realize is that fuzzy logic is a logic of fuzziness, not a logic which is itself fuzzy ‘=> 1.4.2 Genetic Algorithms ‘© Genetic algorithms are inspired by Darwin's theory of natural evolution. In the natural world, organisms that are poorly suited for an environment die off, while those well-suited, prosper. ‘© Genetic algorithms search the space of individuals for good candidates. The chance of an individual's being selected is proportional to the amount by which its fitness is greater or less than its competitors’ fitness. ‘© It is basically a method for moving from one population of chromosomes to a new generated population using selection together with the genetic inspired operators of crossover, mutation and inversion. ‘= It is basically a search technique that will map data for the problems where no particular formula can be implemented. © Genetic algorithms are broadly applicable and have the advantage that they require little knowledge encoded in the system, ‘=> 1.4.3 Neural Networks ‘* Neural networks consists of many number of simple elements (neurons) connected between them in system. Whole system is able to solve of complex tasks and to learn for it like a natural brain. TECHNICAL PUBLIGATIONS © An upihnst for krowedpe Soft Computing (1-7) Introduction to Soft Computing ee ee eeknowledge encoded in the system. ‘=> 1.4.3 Neural Networks ‘+ Neural networks consists of many number of simple elements (neurons) connected between them in system, Whole system is able to solve of complex tasks and to learn for it like natural brain. TECHNICAL PUBLICATIONS © - Anup or krowedge Soft Computing (7 Introduction to Soft Computing ‘+ For user NN is black box with Input vector (source data) and Output vector (result) ‘* A Neural network is usually structured into an input layer of neurons, one or more hidden layers and one output layer. ‘* Neurons belonging to adjacent layers are usually fully connected and the various types and architectures are identified both by the different topologies adopted for the connections as well by the choice of the activation function ‘+ The values of the functions associated with the connections are called “weights”. ‘+ The whole game of using NNs is in the fact that, in order for the network to yield appropriate outputs for given input, the weight must be set to suitable values. The way this is obtained allows a further distinction among modes of operations. ‘+ A neural network is a processing device, either an algorithm or actual hardware, whose design was motivated by the design and functioning of human brains and components thereof. ‘+ Most neural networks have some sort of “training” rule whereby the weights of connections are adjusted on the basis of presented patterns. ‘+ Inother words, neural networks “learn” from examples, just like children learn to recognize dogs from examples of dogs, and exhibit some structural capability for generalization. ‘+ Neural networks normally have great potential for parallelism, since the computations of the components are independent of each other. ‘+ Neural networks are a different paradigm for computing 1, Vone neumann machines are based on the processing/memory abstraction of human information processing. 2, Neural networks are based on the parallel architecture of animal brains. ‘+ Neural networks are a form of multiprocessor computer system, with a. Simple processing elements _b. A high degree of interconnection ¢. Simple scalar messages . Adaptive interaction between elements ‘=> 1.4.4 Probabilistic Reasoning ‘+ Uncertainty is described by probabilities. Probability may be use for simulation of fuzziness ‘+ Relations between events are described as conditional probabilities (Bayesian nets) or probabilities of transition probabilities (Markovian process). ‘+ For example, action of system may be described as graph of states ~ ‘* Hard computing deals with precise models where accurate solutions are achieved quickly. On the other hand, soft computing deals with approximate models and gives solution to complex problems TECHNICAL PUBLICATIONS ©- Anup ihn or knonieage Soft Computing (ee) Introduction to Soh Computing — Faye se ‘Neuro fuzzy computing ‘© Soft computing methods have been applied to many real-world problems. Applications can be found in signal processing, pattern recognition, quality assurance and industrial inspection, business forecasting, speech processing, eredit rating, adaptive process control, robotics control, natural language understanding, etc.* Neurons belonging to adjacent layers are usually fully connected and the various types and architectures are identified both by the different topologies adopted for the connections as well by the choice of the activation function. * The values of the functions associated with the connections are called “weights”. * The whole game of using NNs is in the fact that, in order for the network to yield appropriate outputs for given input, the weight must be set to suitable values. The way this is obtained allows a further distinction among modes of operations * A neural network is a processing device, either an algorithm or actual hardware, whose design was motivated by the design and functioning of human brains and components thereof. * Most neural networks have some sort of “training” rule whereby the weights of connections are adjusted on the basis of presented patterns. * In other words, neural networks “learn” from examples, just like children learn to recognize dogs from examples of dogs, and exhibit some structural capability for generalization. # Neural networks normally have great potential for parallelism, since the computations of the components are independent of each other. * Neural networks are a different paradigm for computing 1, Vone neumann machines are based on the processing/memory abstraction of human information processing. 2. Neural networks are based on the parallel architecture of animal brains. * Neural networks are a form of multiprocessor computer system, with a. Simple processing elements __b. A high degree of interconnection ¢. Simple scalar messages 4. Adaptive interaction between elements ‘=> 1.4.4 Probabilistic Reasoning * Uncertainty is described by probabilities. Probability may be use for simulation of fuzziness * Relations between events are described as conditional probabilities (Bayesian nets) or probabilities of transition probabilities (Markovian process) * For example, action of system may be described as graph of states - Sy Py 8, P, Pu Sw + Hard computing deals with precise models where accurate solutions are achieved quickly. On the other hand, soft computing deals with approximate models and gives solution to complex problems. TECHNICAL PUBLICATIONS ® - An up thrust for knowiedge Soft Computing (1-8) Introduction to Soft Computing ‘eat [Fayre ane Neuro fuzzy computing Derivative free optimization ‘Soft computing * Soft computing methods have been applied to many real-world problems. Applications can be found in signal processing, pattem recognition, quality assurance and industrial inspection, business forecasting, speech processing, credit rating, adaptive process control, robotics control, natural language understanding, etc. © Possible new application arcas are programming languages, user friendly application interfaces, computer networks, database management, fault diagnostics and information security etc, * For learning and adaptations, soft computing requires extensive computation. In this sense, soft computing shares the same characteristics as computational intelligence. * Soft computing does not perform much symbolic manipulation, so user can view it as a new discipline that complements conventional artificial intelligence approaches and vice versa. Lo — Co