Neural Networks - Research Paper Example

?NEURAL NETWORKS Neural Networks Affiliation This paper presents an overview of the Artificial Neural Networks (ANNs). This paper will outline the basic idea, some history and uses of ANN. Basically an artificial neural network is a collection of programs of data structures that is almost capable of the process of the human brain. In this scenario, a neural network encompasses a wide variety of processors that work collaboratively, each with its own small sphere of facts and makes use of the data in its local memory. In addition, an ANN can be considered as a computer program or simulation that is built through a wide variety of clear-cut and extremely unified processing components that are capable of managing information by means of their active state reaction to inputs. In fact, artificial neural networks have the capability to solve those problems that are difficult to resolve using traditional processes and techniques, as well as sometime a lot of them have been tackled through neural networks, for instance forecasting grassland community, analysis of seismic signals or solar radiation, control of chaotic dynamical systems, data and air quality control and categorization of remotely sensed information (Benvenuto & Marani, 2000; Giles, 1998; Laudon & Laudon, 1999). According to (Stergiou & Siganos, 2010), an artificial neural network is a data processing model that is inspired by the method of biological nervous systems information processing. Neural Network is also a renowned and heavily researched area of the AI. Stergiou & Siganos (2010) described that in the neural network field the working of human neuron is copied into the computer based systems. In addition, the neural network is an AI based computer program which imitates the human nervous systems’ working. It is based on neural net which has weights on each node that can be adjusted to make suitable decisions. Such systems are typically seen in automatic driving systems that are under-research and at the early stages of development (Stergiou & Siganos, 2010; Laudon & Laudon, 1999). The history of the artificial neural networks can be traced back to the 1940's with the introduction of the first neural network computing model, which was presented by McCulloch and Pitts. In the 1950's, Rosenblatt's work got a considerable attention of researchers. Additionally, this work was further resulted in a two-layer network, the perceptron, which was able to learn various classifications by making changes to connection weights. Without a doubt, the perceptron attained a considerable success in classifying various patterns, on the other hand it had many flaws. In fact, the perceptron was not capable of solving the typical XOR (exclusive or) scenario. As a result, this kind of flaws caused the decline of the field of neural networks. On the other hand, the perceptron had created the foundations for the future work in neural networks. Later in the early 1980's, researchers in the field of neural networks demonstrated growing concern in neural networks. In addition, latest developments in the field of neural networks comprise Hopfield nets, Boltzmann machines, competitive learning models, multilayer networks, and adaptive resonance theory models (Russell, 1996). According to (Oprea & Matei, 2010), ANN can be helpful in resolving a wide variety of issues. In this scenario, one of the main environmental problems and challenges that require using efficient software tools is the forecast issues such as meteorological forecast, water, soil, air, flood prediction, pollution forecast and many more. In addition, a wide variety of tools and techniques based on the artificial neural networks have been designed and implemented by taking into account that they are able to offer additional informed techniques that make use of domain specific information as well as offer solutions faster than the traditional techniques that are based on mathematical techniques (Oprea & Matei, 2010). Without a doubt, the worldwide incidence of floods cannot be avoided, but they can be minimized, as well as their impact can be minimized by adopting a systematic procedure that guides to a succession of events and proceedings intended to add to the minimization of the risk associated with these phenomena. In addition, artificial neural networks are very helpful for modeling, in a discrete domain, of the hydrological procedures. In this scenario, at the present artificial neural networks are extensively utilized in flood prediction for the avoidance of possible issues and problems regarding the floods and their influences (Oprea and Matei, 2010; Benvenuto & Marani, 2000). According to (Smith, 2003; Orr, 1999), at the present neural networks are used in a number of fields. Some of the uses of neural networks are outlined below: Neural networks are used in investment analysis in an attempt to forecast various aspects of stocks currencies etc., by making use of existing data. In this way they are taking place of previous simpler linear models. Neural networks are extensively used in signature analysis as an effective technique for comparing signatures made (for instance in a financial organization) with those stored in the database. In fact, it is one of the initial across-the-board implementations of neural networks in the USA as well as is also one of the initial to make use of a neural network chip. There are certain applications of the neural networks in process control. In view of the fact that most processes cannot be determined as computable algorithms. Some of the applications in process control include backing up a truck, cab position, rear position, and compare with the dock get converted to steering instructions. Manufacturing plants for controlling automated machines. For instance, Siemens productively makes use of neural networks for automating its business processes in basic industries such as in rolling mill control more than 100 neural networks perform this task, 24 hours a day. In this scenario, Chemical Engineering Department of Newcastle University is working with industrial partners Zeneca and BP in this field. Neural networks can be used to keep an eye on the condition of aircraft engines. In addition, by means of monitoring vibration levels and sound, early warning of engine failure can be given. In this scenario, British Rail has recently implemented a similar application monitoring diesel engines. Neural networks can be used to enhance marketing mailshots. This technique requires execution of a test mailshot, and observes the pattern of results from this test. The basic aim of this test is to discover an analytical modeling of the data collected about the customers in opposition to their responses. This modeling is afterward used to direct further mailshots. In signal processing neural networks are used to suppress line noise, with adaptive echo canceling, blind source separation. Neural networks are used in robotics (which is the branch of artificial intelligence and deals with making intelligent robots). In this scenario, neural networks are used to control navigation, vision recognition. In conclusion, this paper has presented a detailed overview of neural networks. An artificial neural network is a data processing model that is inspired by the method of biological nervous systems information processing. Neural Network is also a renowned and heavily researched area of the artificial intelligence. This paper has discussed the brief history of neural networks. This paper has discussed various applications of neural networks. In future, there are more developments expected in the field of neural networks. References Benvenuto, F., and Marani, A, 2000, Neural networks for environmental problems: data quality control and air pollution nowcasting: Global Nest: the International Journal, Volume 2 Issue 3, 281-292. Giles, L., 1998, Neural Network, http://searchnetworking.techtarget.com/definition/neural-network, Accessed Februray 09, 2011 Laudon, K. C., & Laudon, J. P. (1999). Management Information Systems, Sixth Edition (6th ed.). New Jersey: Prentice Hall. Oprea, M., and Matei, A, (2010). Applying Artificial Neural Networks in Environmental Prediction Systems. Retrieved July 20, 2012, from http://www.wseas.us/e-library/conferences/2010/Iasi/ICAI/ICAI-18.pdf Orr, G. (1999). CS-449: Neural Networks. Retrieved July 20, 2012, from http://www.willamette.edu/~gorr/classes/cs449/intro.html Russell, I. (1996). Neural Networks Module. Retrieved July 21, 2012, from http://uhaweb.hartford.edu/compsci/neural-networks-history.html Smith, L. (2003, April 02). An Introduction to Neural Networks. Retrieved July 22, 2012, from http://www.cs.stir.ac.uk/~lss/NNIntro/InvSlides.html Stergiou, C., & Siganos, D. (2010). NEURAL NETWORKS. Retrieved July 18, 2012, from http://www.doc.ic.ac.uk/~nd/surprise_96/journal/vol4/cs11/report.html Read More