Volume 2, Issue 3, March 2012 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Convergence of Artificial Intelligence, Emotional Intelligence, Neural Network and Evolutionary Computing Sandeep Kumar Tata Consultancy Services Limited Pune, Maharashtra India 411028 sandeep.k2@tcs.com Medha Sharma Tata Consultancy Services Limited Gurgaon, Delhi India 122001 Abstract— This paper presents a new perspective of Artificial Intelligence (AI). Although, number of attempts has been made to make an artifact intelligent, including evolution theory, neural network etc and a number of problems have been solved using these concepts but each of this theory covers only some aspect of human intelligence. Still there is a large gap between artificial intelligence agent and human being. In this paper I have discuss the extended version of Artificial Intelligence by augmenting it with emotions, and inheritance of neural architecture from parent generation to child generation that can make an artificial agent to match the intelligence and behaviour of a human being. At the same time it adds the power of two well knows Artificial Intelligence techniques viz. Neural Computing and Genetic Computing or Evolutionary Computing. The paper gives an idea of an artefact which is supposed to match the intelligence and behaviour of a human being. Paper also discusses some natural phenomenon and how they can be confirmed by the revised definition of artificial intelligence. The paper does not claim that existing definition of artificial intelligence has some faults. The paper just augments the existing definition by some other features that can make it more close to natural intelligence. The features augmented are naturally inspired similarly as AI, Neural Network and genetics all are naturally inspired. Keywords— Neural Computing, Evolutionary Computing, Intelligence, Emotions, Soft Computing I. INTRODUCTION Artificial Intelligence (AI) Although, artificial intelligence is a very general term but defining it precisely is very difficult. And the design of an artificially intelligent agent totally depends on the fact how we define the term ‗Artificial Intelligence‘. Possibly, the right definition can lead us to develop a successful intelligent artifact. There are a number of definitions to define artificial intelligence. As discussed in [1] the successful definitions are along two dimensions: firstly, whether it is with respect to reasoning (thought) or behavior (action) and secondly, whether it is with respect to human or ideal (i.e. rational) as shown in the fig-1. If we consider the category-1 or category- 3 definition of AI then we try to develop an artifact that can think like and can act like human being respectively. Further, if we consider category-2 or category-4 definition of AI then we try to develop an artifact that thinks or acts optimally respectively. Even development of optimal agents (based on definition from category-2 and category-4) could be really useful for solving problems and Category-4 definition (i.e. Acting Rationally) is the standard and modern definition of artificial agent [1]. Fig.1 Four perspective of Artificial Intelligence Definitions given by Haugeland [2] and Bellman[3] belongs to the first category. Definitions given by Charniak[4] and Winston[5] falls under second category. Definitions given by Kurzweil[6] and Rich and Knight[7] belongs to third category. Finally, definitions given by Poole et al. [8] and Nilsson[9] give support to forth category. Brief discussion of all these definition is covered in [1]. A lot of research work has been done among all the dimensions depending upon the need. As a result we have different model of artificial agent following the definitions from different category. In this paper, we would be considering definitions from category 1 and 3 only, means we will be defining the intelligence in terms of human only. The motivation behind this is that rational agent is supposed to take rational decision but human is not expected to do so. And here the main