Medsker L.R., Hybrid Intelligent Systems, Kluwer Academic Publishers, 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....[6] Our current research relies heavily on the concept of agency across a number of different domains, and categories of processes, agencies and agents. For example, we can consider our ongoing research into decision support systems [7] as that of an investigation into tightly and loosely coupled [8] agent communities making use of modern but relatively orthodox AI techniques. Again we can consider the investigation of (co operative and competitive) emergence in our simulation work [9] as complimentary to that discussed in research related to artificial societies [2] This current paper ....

L. R. Medsker, Hybrid Intelligent Systems, ISBN 0-7923-9588-3, Kluwer Academic Publishers, 1995.

....knowledge extraction, SRN networks, analysis of connectionist learning 1. Introduction There has been a lot of interest lately in knowledge structures and their representation in artificial neural networks [HSlldobler, 1990, Kurfefi, 1991, Sperduti et at. 1995, Wermter, 1995, Hallam, 1995, Medsker, 1995, Sun, 1995, Wermter et at. 1996, Elman et at. 1996, Craven, 1996, Wermter, 1999] Artificial neural networks (or connectionist networks) have already demonstrated interesting learning results for various classification tasks. However, it continues to be very difficult to understand the ....

Medsker, L. R. (1995). Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston.

....available software packages. On the other hand, standalone techniques are not transferable; neither can support the weakness of the other technique. 2. 2 Transformational Hybrid Architecture In a transformational hybrid model the system begins as one type of system and end up as the other [57]. Determining which technique is used for development and which is used for delivery is based on the desirable features that the technique offers. Expert Systems and ANNs have proven to be useful transformational models. Variously, either the expert system is incapable of adequately solving the ....

Medsker LR, Hybrid Intelligent Systems, Kluwer Academic Publishers, 1995.

....1997, Papka et al. 1997] For these more sophisticated retrieval tasks it may be an option to consider the integra tion of hybrid neural techniques for improving information retrieval in the future. Pre viously there has been some work on hybrid neural integration [Reilly and Sharkey, 1992, Medsker, 1995, Wermter et al. 1996, Elman et al. 1996] which potentially could be useful to information retrieval. In general, robust and learning architectures have been identi fied as important current areas for natural language processing and information retrieval [Lewis, 1991, Briscoe, 1997, Cunningham ....

Medsker, L. R. (1995). Hybrid Intelligent Systems. Kluwer Academic Pub- lishers, Boston.

....the interpretation and combination of various neural preference Moore machines with additional real world examples. i Introduction Previously, there has been a lot of work on hybrid neural integration and combination [Reilly and Sharkey, 1992, Miikkulainen, 1993, Yager, 1994, Wermter, 1995, Medsker, 1995, Wermter et al. 1996, Dorffner, 1997] Currently, it is an open question whether neural or symbolic approaches alone will be sufticient to provide a general framework for intelli gent performance, e.g. for processing natural language [Dyer, 1991, Honavar and Uhr, 1994, Sun and Bookman, 1995, ....

....and neural networks with symbolic interpretations are currently examined and this paper is a contribution to this hybrid neural integration. Hybrid neural symbolic representations have been found advantageous in some con texts since different mutually complementary properties can be combined [Medsker, 1995, Wermter et al. 1996, Dorffner, 1997] Symbolic representations have advantages with re spect to easy interpretation, explicit control, fast initial coding, dynamic variable binding and knowledge abstraction. On the other hand, neural representations show advantages for gradual plausibility, ....

Medsker, L. R. (1995). Hybrid Intelligent Systems. Kluwer Academic Pub- lishers, Boston.

....of experimental research) becomes an impossible task. The underlying cause behind this is the lack of theories (conceptual models of reference) methodologies and tools for research and development purposes. Fortunately, some recent works represent considerable contributions in these areas (e.g. [17] or, in this volume, 12, 15, 16] 1 To address these aforementioned problems, a framework devised for the interoperation of heterogeneous systems is proposed as a well suited approach. The main characteristics of this approach are identified in section 2. These objectives have nurtured the ....

Larry R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, 1995.

....them have had deep impact on development of intelligent systems. Together, they form a context within which modern knowledge modeling techniques should be discussed. Such streams include object oriented software design [9] 32] layered software architectures [56] development of hybrid systems [41], multimedia systems [23] and, of course, distributed systems and the Internet [36] Along with such a context, any discussion of knowledge modeling should also include a reference to the kinds of knowledge that can be represented in the knowledge base of an intelligent system, as well as to the ....

L.R. Medsker, Hybrid Intelligent Systems, Kluwer Academic Publishers, Amsterdam, 1994.

....CI methods in order to construct an overall system, but also in the sense of the theoretical unification of AI and CI ideas (Figure 1) In the following, we will adopt a taxonomy of hybrid intelligent systems, as shown in Figure 2. This taxonomy is done on the basis of previous work by (see [16] [25]) The basic innovation with respect to the latter lies in the use of a different terminology and in the distinction of hybrid intelligent systems into three categories. The first category, Combined Intelligent Systems, comprises systems that use NNs as tools for symbolic processing (top down ....

Medsker L.R., Hybrid Intelligent Systems, Kluwer Academic Publishers, 1995

....of front line operators and experts, in order to achieve continuous improvements in productivity. The use of many concepts from discipline areas such as information theory, neural networks and fuzzy logic has been proposed to model and control systems that would create hybrid intelligent systems [13]. During the last decades manufacturing systems have utilised the advantages of computer and automation technology and have made many advances. The requirements for more advanced manufacturing systems, which are characterised by high autonomy and intelligence, have led engineers to investigate ....

Medsker, R.: Hybrid Intelligent Systems, Kluwer Academic, Dordrecht, 1995.

.... approaches in the learning community from the theory of neural network ensembles and modular networks [31] to multistrategy learning [26] Hybridization in a broader sense is seen in e orts to combine two or more of neural network, Bayesian, GA, fuzzy logic and knowledge based systems [25, 1, 35, 4]. The goal is again to incorporate diverse sources and forms of information and to exploit the somewhat complementary nature of di erent methodologies. The main form of hybridization of interest in this chapter involves the integration of symbolic and connectionist approaches [24, 35] 15] 6, ....

Medsker, L. R. (1995). Hybrid Intelligent Systems. Kluwer Academic, Norwell, MA.

....of the main methods used, outline the work that is presented here, and provide additional references. We will also highlight some important general issues and trends. 1 Introduction In recent years, the research area of hybrid and neural processing has seen a remarkably active development [62, 50, 21, 4, 48, 87, 75, 76, 25, 49, 94, 13, 74, 91]. Furthermore, there has been an enormous increase in the successful use of hybrid intelligent systems in many diverse areas such as speech natural language understanding, robotics, medical diagnosis, fault diagnosis of industrial equipment and financial applications. Looking at this research ....

.... of Hybrid Neural Architectures Various classification schemes of hybrid systems have been proposed [77, 76, 89, 47] Other characterizations of architectures covered specific neural architectures, for instance recurrent networks [38, 52] or they covered expert systems knowledge based systems [49, 29, 75]. Essentially, a continuum of hybrid neural architectures emerges which contains neural and symbolic knowledge to various degrees. However, as a first introduction to the field, we present a simplified taxonomy here: unified neural architectures, transformation architectures, and hybrid modular ....

L. R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston, 1995.

....interpretation and simultaneous processing of mean ring rate and pulse coding schemes in a preferences framework. 1 Introduction The development of hybrid models, integrating neural and symbolic approaches, has received a fair amount of interest. Some signi cant work has been done in this area [Medsker, 1995, Dor ner, 1997] but much less has been done on fundamental principles of neural symbolic hybrid systems [Smolensky, 1988, Sharkey and Jackson, 1995, Wermter, 1995] Furthermore, most neural network models considered in hybrid approaches use the mean ring rate as a concept of encoding the ....

Medsker, L. R. (1995). Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston.

....of the main methods used, outline the work that is presented here, and provide additional references. We will also highlight some important general issues and trends. 1 Introduction In recent years, the research area of hybrid and neural processing has seen a remarkably active development [62, 50, 21, 4, 48, 87, 75, 76, 25, 49, 94, 13, 74, 91]. Furthermore, there has been an enormous increase in the successful use of hybrid intelligent systems in many diverse areas such as speech natural language understanding, robotics, medical diagnosis, fault diagnosis of industrial equipment and nancial applications. Looking at this research area, ....

.... Forms of Hybrid Neural Architectures Various classi cation schemes of hybrid systems have been proposed [77, 76, 89, 47] Other characterizations of architectures covered speci c neural architectures, for instance recurrent networks [38, 52] or they covered expert systems knowledge based systems [49, 29, 75]. Essentially, a continuum of hybrid neural architectures emerges which contains neural and symbolic knowledge to various degrees. However, as a rst introduction to the eld, we present a simpli ed taxonomy here: uni ed neural architectures, transformation architectures, and hybrid modular ....

L. R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston, 1995.

....Symbolic Agent Architectures based on Neuroscience Constraints University of Sunderland, Sunderland, UK, stefan.wermter sunderland.ac. uk Adaptive symbolic and neural agents have received a lot of interest for different tasks, for instance speech language integration and image text integration [159, 155, 132, 216, 44, 220]. Hybrid neural symbolic methods have been shown to be able to reach a level where they can actually be further developed in real world scenarios. A combination of symbolic and neural agents is possible in various hybrid processing architectures, which contain both symbolic and neural agents ....

L.R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston, 1995.

....complete for computing the solution. Mainly we put in this theoretical container four goods that, without any claim to exhaustiveness, cover the principal ingredients of many instances of knowledge integration: i. we deal with symbolical functions as results of available though incomplete theories [4]; ii. we focus on dynamical systems since they are able to host the recursive kernel of any non trivial computation [5,6] iii. we manage decision trees as top down counterparts of both mixtures of experts a recent paradigm much used in both neural network [7,8] and computational learning [9] ....

L.R. Medsker, Hybrid Intelligent Systems. Kluwer Academic Publishers, 1995.

....language processing has a lot of potential to overcome the gap between a neural level and a symbolic conceptual level. ii 1 Motivation for hybrid symbolic connectionist processing In recent years, the field of hybrid symbolic connectionist processing has seen a remarkable development [48, 38, 18, 2, 36, 59, 55, 20, 37, 61, 9]. Currently it is still an open issue whether connectionist or symbolic approaches alone will be sufficient to provide a general framework for processing natural language [51, 11, 27, 56] However, since human language capabilities are based on real neural networks in the brain, artificial neural ....

....systems, their interpretation, and communication principles within various architectures. Previous characterizations of architectures have covered certain specific connectionist architectures, for instance recurrent networks [31, 39] or they have covered expert systems knowledge based systems [37, 24, 55]. In contrast, here we will concentrate on various types of hybrid connectionist natural language processing. In figure 1 there is an overview of different possibilities for integration in natural language processing. Continuous connectionist representations are represented by a circle, discrete ....

L. R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston, 1995.

.... Workpackage: 2 Document Type: Deliverable XX Document ID: MIX WP2 UPM DX Status: Final Date of Issue: May 15, 1997 Distribution: Consortium, reviewers 1 1 Introduction The MIX approach to hybrid systems has taken advantage of the intelligent agent paradigm, as remarked by Medsker [ Medsker, 1995, page 238 ] Intelligent agents have been defined for encapsulating the different symbolic and connectionist techniques. For this purpose, the MIX multiagent platform [ Gonz alez et al. 1995b; Iglesias et al. 1994 ] was developed during the first year of the project. This platform was ....

Larry R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston, MA, 1995.

No context found.

Medsker L.R., Hybrid Intelligent Systems, Kluwer Academic Publishers, 1995.

No context found.

L.R. Medsker, Hybrid Intelligent Systems, Kluwer Academic Publishers, 1995.

No context found.

Larry R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, 1995.

No context found.

Larry R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, 1995.

No context found.

Larry R. Medsker. Hybrid Intelligent Systems. Kluwer Academic Publishers, Boston, 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC