Towell, G., Shavlik, J.: Knowledge-based Artificial Neural Networks. Artificial Intelligence, Vol. 70. (1994)  Home/Search   Document Details and Download   Summary   Related Articles   Check

....learning algorithm, that approximates Plotkin s least general generalization (lgg) Results show that our initial goal, learning logic programs using neural networks, has been achieved. 1 Introduction The Cascade ARTMAP [21] system is a knowledge based neural network (KBNN) 20] like KBANN [22], RAPTURE [11] and C IL2p [6] that has been shown to outperform other purely analytical or inductive systems in the task of propositional theory refinement: a prior incomplete and or partially correct propositional symbolic knowledge about a problem domain is given to a theory refinement system ....

G. Towell and J. Shavlik. Knowledge-Based Artificial Neural Networks. Artificial Intelligence, pp.119-165, vol. 70, 1994.

....survey of the latter. An important point of the present work is the systematic way the di#erent interpretations has been presented in order to permit their comparison. RBFNs are particularly suitable for integrating the symbolic and connectionist paradigms in the line draw by Towell and Shavlik [78] whose recent developments has been surveyed by Cloete and Zurada [22] This symbolic interpretation permits to consider RBFNs as intrinsically Knowledge Based Networks. Moreover, RBFN have also very di#erent interpretations. They are Regularization Networks so there is the possibility of tuning ....

G. Towell and J.W. Shavlik. Knowledge based artificial neural networks. Artficial Intelligence, 70(4):119--166, 1994.

.... been extensive research activity at combining (or integrating) the symbolic and the connectionist approaches for knowledge representation in expert systems [3, 15, 16, 19] Especially, there are a number of efforts combining symbolic rules and neural networks that map rules into neural networks [4, 9, 18]. In addition, connectionist expert systems [6, 7, 17] are a type of integrated systems that represent relationships between concepts, considered as nodes of a neural network. The above approaches give pre eminence to connectionism and use a neural network as a knowledge base. The strong point of ....

Towell, G., Shavlik, J., "Knowledge-Based Artificial Neural Networks", Artificial Intelligence 70, 1994, 119-165.

.... Logic Programming (ILP) techniques, ii) Hybrid (neural and symbolic) Systems; and (iii) Explanation Based Learning (EBL) algorithms (see [31] Among these, hybrid systems seem to be more appropriate as far as dealing with incorrect background knowledge and theory refinement is concerned [10, 46, 45]. Hybrid systems are not normally push button techniques though, as they typically use traditional neural learning algorithms (such as Backpropagation) which require the adaptation of a learning rate via trial and error. On the other hand, explanation based learning algorithms seem to require ....

....learning could be one of these actions [11] An extension of this work would be to investigate the use of other techniques of machine learning for revising requirements specifications. These include extensions of Inductive Logic Programming, Knowledge based Artificial Neural Networks (KBANN) [46] and Explanation based Neural Networks EBNN [44] and their hybrids, e.g. 33] Experiments on a number of real world problems would allow us to perform more detailed technical evaluation of these techniques, and draw general conclusions on when, why, and for which type of requirements ....

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1):119-165, 1994.

....refined. This has the following advantages: 1) the agent is able to perform reasonably well initially because it can utilize the users prior knowledge and 2) users prior knowledge does not have to be correct as it is refined through learning. Information is derived by extracting rules from KBNNs [62]. In order to map large sized web pages into fixed sized NNs, a concept of sliding window is used. This parses each page considering three words at a time, and the html tags like act as window breakers. Using self generated training examples it can act also as a self tuning agent. Rules of the ....

J. Shavlik and G. G. Towell, "Knowledge-based artificial neural networks, " Artificial Intell., vol. 70, no. 1/2, pp. 119--165, 1994.

....series. We now present a method of obtaining rules directly from the time series. We encode these rules into a neural network using the KBANN encoding method. 3.9. 1 Knowledge Based Artificial Neural Networks Methods for encoding a Boolean rule set into a feedforward network have been proposed [66]. Other methods differ only in the way that they combine the input neurons. The initial network is constructed based on the relationship between rules in the rule set. Rule inputs become input neurons, intermediate results become hidden neurons, final results become output neurons, and ....

G. Towell and J. Shavlik, "Knowledge-based artificial neural networks," Artificial Intelligence, vol. 70, no. 1,2, pp. 119--160, 1994.

....it as possible. 1 INTRODUCTION There has been extensive research activity at combining (or integrating) the symbolic and the connectionist approaches for knowledge representation in expert systems [7, 8, 10] Especially, there are a number of efforts combining symbolic rules and neural networks [2, 3, 6, 9]. They give pre eminence to connectionism and use a neural network as a knowledge base. The main objective is to reduce knowledge elicitation from experts to a minimum. In such approaches, connectionism is mainly used as a means for refining an initial background rule base. Integration with ....

G. Towell and J. Shavlik, `Knowledge-Based Artificial Neural Networks', Artificial Intelligence, 70(1-2), 119-165, (1994).

....programmers to arrive at more efficient solutions, and they also facilitate debugging during software development. The issue of integrating learning into inference systems has been studied intensely before. For example, recent work on explanation based learning [69, 42, 20] theory refinement [94, 108, 80, 78], and inductive logic programming [73, 85] has led to a variety of learning algorithms that modify programs written in first order logic based on examples. Several research teams have integrated such learning algorithms into problem solving architectures, such as SOAR [89, 26, 66, 57] PRODIGY [67, ....

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1/2):119--165, 1994.

.... the symbolic and the neurocomputing approaches (see e.g. 6, 9] To that end, there are a number of efforts at combining production rules and neural networks for knowledge representation [5] Some of them follow the unified approach [3, 4, 8] whereas others follow a pseudo hybrid approach [1, 2, 7], called the translational approach in [5] A weak point of both approaches is that the resulted system lacks the naturalness and modularity of symbolic rules. In this paper, we introduce a KR formalism which attempts to incorporate aspects of neurocomputing within the symbolic framework of ....

Towell G. G. and Shalvik J. W., Knowledge-based artificial neural networks, Artificial Intelligence 70 (1994) pp. 119-165.

....proposed approach achieves good performance on different datasets. 2 1 Introduction Promoters are transcription signals, which regulate gene expressions. Characterization and recognition of such signals is an important research topic and has been studied by many researchers. 15] 18] 20] [27], for example, analyzed E. Coli promoters. 19] compiled and clustered a set of promoters recognized by E. Coli RNA polymerase. More recently, 6] 12] 23] considered eukaryotic promoters and presented techniques for detecting these signals. In this paper we focus on the recognition of E. Coli ....

....to assign unlabeled test data to either the target class or the non target class. The importance of the binary classification problem has been addressed in the DM literature [9] 30] In the past, several researchers have considered the binary classification problem for E. Coli promoters. In [27], Towell and Shavlik proposed to initialize the topology and weights of a neural network according to the characteristics of E. Coli promoters. They built a system, called KBANN, for recognizing the promoters. Later, Opitz [18] employed a genetic algorithm to search through the topology space of ....

[Article contains additional citation context not shown here]

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence 70, 119--165, 1994.

.... purpose, many methods have already been proposed in the literature (see [1] 2] and references therein) Talking about integration of DTs and NNs alone, for example, we can design a DT first, and ENN ENN ENN ENN ENN Figure 1: A neural network tree (NNTree) then derive an NN from the DT [3] [5]. This method is good for fast design of NNs. Inversely, we can design an NN first, and then extract a DT from it [6] 8] This approach is helpful for opening the NN black boxes. We can also combine the above two to refine domain knowledge. Using the above transformational approaches, however, ....

G. G. Towell and J. W. Shavlik, "Knowledgebased artificial neural networks," Artificial intelligence, 70(1-2), pp. 119-165, 1994.

....have been presented. An important point of the work is the unifying view, that have been proposed in order to compare and to integrate all the different approaches. RBFNs are particularly suitable for integrating the symbolic and connectionist paradigms in the line draw by Towell and Shavlik [Towell and Shavlik, 1994] for the usual ANN s architec 90 ture. With respect to this work the symbolic interpretation emerges from RBFNs in a more natural way and so knowledge extraction requires simpler algorithms. More specifically, this thesis has presented a technique for using classical symbolic learning algorithms ....

Towell, G. and Shavlik, J. (1994). Knowledge based artificial neural networks. Artficial Intelligence, 70(4):119--166.

....serve as a blueprint for the synthesis of proteins. Interspersed among the genes are segments, called non coding regions, that do not encode proteins. 196 A.S. d Avila Garcez et al. Artificial Intelligence 125 (2001) 155 207 Fig. 22. Part of the network for Promoter Recognition. Following [36], we use a special notation to identify the location of nucleotides in a DNA sequence. Each nucleotide is numbered with respect to a fixed, biologically meaningful, reference point. For example, 3 atcg states the location relative to the reference point in the DNA, followed by the sequence of ....

G.G. Towell, J.W. Shavlik, Knowledge-based artificial neural networks, Artificial Intelligence 70 (1994) 119--165.

....of a DT, and make the global decision based on all terminal nodes [9] 12] 5. Embed NNs directly into DTs [13] 14] The first approach was proposed originally for fast design of neural networks. This approach can be extended to a more general framework named as knowledge based neural networks [3]. NNs so obtained are expected to cover domain knowledge better with fewer connections. However, once they are exposed to changing environments, they become black boxes again. The second approach enable us to open the black boxes, and see what are in them. The main problem is that the ....

G. G. Towell and J. W. Shavlik, "Knowledge-based artificial neural networks," Artificial intelligence, 70(1-2), pp. 119-165, 1994.

....This has the following advantages: a) the agent is able to perform reasonably well initially because it can utilize the users prior knowledge, and (b) users prior knowledge does not have to be correct as it is refined through learning. Information is derived by extracting rules from KBNN s [62]. In order to IEEE TRANS. NEURAL NETWORKS 11 map large sized web pages into fixed sized neural networks, a concept of sliding window is used. This parses each page considering three words at a time, and the html tags like p ; p ; br act as window breakers. Using self generated training ....

J. Shavlik and G. G. Towell, "Knowledge-based artificial neural networks," Artificial Intelligence, vol. 70, no. 1/2, pp. 119--165, 1994.

....knowledge based classifier, linear programming, quadratic programming 1 Introduction Support vector machine classifiers [15, 1, 12] generate separating planes or surfaces by training on labeled data, that is data for which the class of each point is given. Knowledge based classifiers [13, 14] on the other hand utilize prior knowledge, e.g. an expert s experience such as a doctor s knowledge in diagnosing a certain disease. Recently [7] a precise incorporation of prior knowledge into a linear support vector machine classifier was achieved by placing nonempty polyhedral sets ....

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70:119--165, 1994.

....concurrency and real time control. Consequently, the issues addressed here are therefore entirely orthogonal. To our knowledge robotics language design has not addressed these issues before. The idea of learning with prior knowledge has been studied extensively in the machine learning community [8, 12]. Prior knowledge is usually represented in declarative form (e.g. Horn clauses) but to our knowledge the idea of integrating learning into a procedural programming language is novel. The most related approach are evolutionary algorithms (EAs) 5] which modify program code directly instead of ....

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1/2):119--165, 1994. 11

....they have the capacity to model highly non linear data distributions [8] ANNs generally consider a fixed topology of neurons connected by links in a pre defined manner. Recently, there have been some attempts in improving the efficiency of neural computation by using knowledge based nets. These [9] constitute a special class of ANNs that consider crude domain knowledge to generate the initial network architecture, which is later refined in the presence of training data. Recently, the theory of rough sets has been used to generate knowledge based networks. The theory of rough sets [10] has ....

G. G. Towell and J. W. Shavlik, "Knowledge-based artificial neural networks," Artificial Intelligence, vol. 70, pp. 119--165, 1994.

....set of operators than we do. As suggested in the previous section, however, we believe that our diagnostic approach can be generalized to work with other operators. A different view of our approach is as an instance of a theory refinement algorithm [Pazzani Kibler, 1992; Ourston Mooney, 1994; Towell Shavlik, 1994] In theory refinement, the goal is to improve the accuracy of an incomplete or incorrect domain theory, from a set of labeled training examples. The primary difference between our work and previous work in theory refinement is the representation used by our learned models. Whereas previous ....

Towell, G., and Shavlik, J. 1994. Knowledge-based artificial neural networks. Artificial Intelligence 70(1,2):119--165.

....are usually initialized by small random values. Recently, there have been some attempts in improving the efficiency of neural computation by using knowledge based nets. This helps in reducing the searching space and time while the network traces the optimal solution. Knowledge based networks [5,15] constitute a special class of ANNs that consider crude domain knowledge to generate the initial network architecture, which is later refined in the presence of training data. Such a model has the capability of outperforming a standard MLP as well as other related algorithms including symbolic and ....

....class of ANNs that consider crude domain knowledge to generate the initial network architecture, which is later refined in the presence of training data. Such a model has the capability of outperforming a standard MLP as well as other related algorithms including symbolic and numerical ones [5,15]. Recently, the theory of rough sets has been used to generate knowledge based networks. A recent trend in neural network design for large scale problems is to split the original task into simpler subtasks, and use a subnetwork module for each of the subtasks [8] The popular methods available ....

[Article contains additional citation context not shown here]

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70:119--165, 1994.

.... have been achieved in mining M of N rules that model patterns of the form If three of these five features are present, the patient has coronary heart disease [Fu, 1999] Neural networks also su#er from other drawbacks, such as the capacity to incorporate prior knowledge in only a limited form [Towell and Shavlik, 1994] and excessive dependence on the original network topology [Opitz and Shavlik, 1997] For an excellent introduction, we refer the reader to [Jordan and Bishop, 1997] Logical Representations While neural networks are attribute value based techniques, more expressive schemes can be obtained by ....

Towell, G. and Shavlik, J. (1994). Knowledge-Based Artificial Neural Networks. Artificial Intelligence, Vol. 70:pp. 119--165.

....related to the learning task exists, pure inductive systems have easily an inherent handicap, especially if the search space is large and the number of examples is small. Integrating background knowledge and inductive learning has been widely studied within the field of theory revision, e.g. [29, 92, 116, 90, 12, 103], where the purpose is to use a coarse domain theory 14 CHAPTER 2. BACKGROUND together with training examples to produce a more accurate theory of the domain. Most theory revision systems are theory centered and concentrate on refining the initial theory through a chain of small local changes, ....

....in general, be guaranteed in real world domains. Our purpose is to use an initial domain theory together with training ex 7.2. BACKGROUND KNOWLEDGE IN CONCEPT LEARNING 75 amples to produce a more accurate theory of the domain. Thus our approach falls within the field of theory revision, see e.g. [29, 92, 116, 90, 12, 103]. Most theory revision systems are theory centered and concentrate on refining the initial theory through a chain of small local changes, guided by the classified examples. Examples of such local changes include adding and removing rules, as well as changing the individual rules by adding or ....

[Article contains additional citation context not shown here]

G. Towell and J. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70:119--165, 1994.

....reoptimized for frozen weights. Such a simplified network has effectively lower number of independent inputs, therefore it is easier to analyze. If symbolic knowledge is used to specify initial weights, as it is done in the Knowledge Based Artificial Neural Networks (KBANN) of Towell and Shavlik [24], weights cluster before and after training. The search process is further simplified if the prototype weight templates (corresponding to symbolic rules) are used for comparison with the weight vectors [25] weights are adjusted during training to make them more similar to templates) The RuleNet ....

G. Towell, J. Shavlik, "Knowledge-based artificial neural networks." Arti- ficial Intelligence 70 (1994) 119-165

....experimentation on a challenging artificial case study shows that the network converges quite fastly and generalizes much better than propositional learners on an equivalent task definition. 1 INTRODUCTION Several papers appeared in the last decade, both in the field of machine learning [16, 15, 10, 1] and in the field of connectionism [17, 8] have shown that 1 combining knowledge based methods with connectionist learning, produced algorithms exhibiting excellent performances on non trivial cases studies. Even if the methods proposed in the literature differ for many substantial aspects, all ....

.... Even if the methods proposed in the literature differ for many substantial aspects, all of them share the basic approach of converting a propositional theory into a neural network which can then be refined using numeric algorithms such as error backpropagation [14] The method developed by [16, 15] is based on multi layer perceptron, whereas the one described in [17, 1] is based on Radial Basis Function Networks (RBFNs) 12] An alternative line is followed by [10] and [8] which adapt the backpropagation to knowledge bases with certainty factors in the style of MYCIN [6] This paper ....

[Article contains additional citation context not shown here]

G. Towell and J.W. Shavlik. Knowledge based artificial neural networks. Artficial Intelligence, 70(4):119--166, 1994.

....require the domain theory to be complete and correct. This has led to attempts to produce hybrid systems that aim at integrating the two approaches, by improving existing domain theories using further examples. EITHER [6] used three different types of inferences to refine incorrect theories. KBANN [15] is a hybrid system that takes nearly complete rulebases, expressed as Horn clauses, and translates them into initial neural networks. Subsequently, these networks are further trained using a separate training set. Although clearly capable of impressive learning performance, the concepts developed ....

G. G. Towell and J. W.Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70:119--165, 1994.

....The output of each unit of the trained ANN must be mapped into boolean values, i.e. each unit represents a boolean rule, which leads to a transparent vision of the ANN. One of the first projects following this approach was proposed by Fu [23] who developed the KT algorithm. Towell Shavlik [24] developed the SUBSet al..gorithm, based on the decompositional technique, which was very computationally expensive and produced rules with a large number of conditions in their antecedents. In order to improve their method, they developed a new algorithm called MofN [9] Two other systems using the ....

G.G. Towell, and J.W. Shavlik. Knowledge-based Artificial Neural Networks. Artificiail Intelligence, v. 69, n. 1, 1994.

....produce specialized and personalized IR agents. W W IE is a general extractor system, which creates specialized agents that extract pieces of information from documents in the domain of interest. W w builds its agents based on ideas from the theory refinement community within machine learning [28,29,45]. First, the user provided domain knowledge is compiled into knowledge based neural networks [45] Then, this prior knowledge is refined whenever mining examples become available. By using theory refinement, we are able to find an appealing middle ground between nonadaptive agent programming ....

....agents that extract pieces of information from documents in the domain of interest. W w builds its agents based on ideas from the theory refinement community within machine learning [28,29,45] First, the user provided domain knowledge is compiled into knowledge based neural networks [45]. Then, this prior knowledge is refined whenever mining examples become available. By using theory refinement, we are able to find an appealing middle ground between nonadaptive agent programming languages and systems that solely learn user preferences from training examples. On one hand, ....

[Article contains additional citation context not shown here]

Towell G.G., Shavlik J.W. (1994). Knowledge-Based Artificial Neural Networks. Artificial Intelligence, 70, 119-165.

....sets, support vector machine classifier, knowledge based classifier, linear programming 1. INTRODUCTION Support vector machines (SVMs) have played a major role in classification problems [24, 3, 12, 1, 13, 5, 6] However un like other classification tools such as knowledge based neural networks [21, 22, 17, 7], little work [20] has gone into incor porating prior knowledge into support vector machines. In this work we present a novel approach to incorporating prior knowledge in the form of polyhedral knowledge sets in the input space of the given data. These knowledge sets, which can be as simple as ....

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70:119 165, 1994.

....trees, iii) neural network ensembles perform better than standard neural networks, iv) knowledge based neural networks perform better than standard neural networks, and (v) an ensemble of knowledge based neural networks performs the best. 1 Introduction Knowledge based neural networks [8] (KNNs) are networks that are derived from a set of rules describing what is currently known about a task. KNNs have been shown to classify novel instances better than a wide variety of machine learning algorithms, including ordinary artificial neural networks (ANNs) In this paper, we describe ....

....The first neural network learning algorithm we used was backpropagation [7] applied to feed forward networks with sigmoidal hidden units. The networks we used have a single layer of 25 hidden units, and are fully connected between layers. We also used two knowledge based network algorithms: KBANN [8] and ADDEMUP [3] In a knowledgebased network, the topology and initial weights of the network are specified by a domain theory consisting of symbolic inference rules. The domain theory used in our experiments was adapted by us from MAX s knowledge base of 75 ART rules. This adapted domain theory ....

G. Towell and J. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1,2):119--165, 1994.

....make it time consuming and di#cult to port an IE system from 1 By annotated examples, we mean the result of the tedious process of reading the training documents and tagging each extraction by hand. one domain to another. In this paper, we demonstrate how the theory refinement approach (e.g. Towell Shavlik, 1994) can be used to build an IE system. By using theory refinement, we are able to strike a balance between needing a large number of labeled examples and having a complete (and correct) set of domain knowledge. Our system takes advantage of the intuition that information retrieval (IR) and IE are ....

....instructions describe how the system should score possible bindings to the slots being filled during the the IE process. We will call the names of the slots to be filled variables, and use binding a variable as a synonym for filling a slot. These initial instructions are then compiled (Towell Shavlik, 1994) into a neural network (called ScoreThisPage) which rates the goodness of a document in the context of the given variable bindings. We refer to the user provided instructions as advice to emphasize that our system does not blindly follow the user s instructions, but instead refines them based on ....

Towell, G. G., & Shavlik, J. W. (1994). Knowledge-based artificial neural networks. Artif. Intell., 70, 119--165.

....Wawa s fundamental operations. These operations are used in both the IR and the IE subsystems of Wawa (see Sections 3 and 5 for further details) Our approach is based on ideas from the theory refinement community within machine learning (Pazzani and Kibler, 1992; Ourston and Mooney, 1994; Towell and Shavlik, 1994). Users input properties of pages and links they like and dislike using a high level advice language. 1 Following Maclin and Shavlik (1996) we call our programming language an advice language, since this name emphasizes that the underlying system does not blindly follow the user provided ....

....have one of the following strength levels: a) weak, b) moderate, c) strong, and (d) definite. These levels represent the degree to which the user wants to increase or decrease the score of a page or a link. The user provided instructions are compiled into knowledge based neural networks (Towell and Shavlik, 1994), thereby allowing subsequent refinement via neural network learning whenever training examples are available. There are two knowledge based neural networks at 1 We envision that there are two types of potential users of our system: a) developers who build an intelligent agent on top of Wawa and ....

[Article contains additional citation context not shown here]

Towell, G. G. and J. W. Shavlik: 1994, `Knowledge-based artificial neural networks'. Artificial Intelligence 70(1/2), 119--165.

No context found.

Towell, G., Shavlik, J.: Knowledge-based Artificial Neural Networks. Artificial Intelligence, Vol. 70. (1994)

No context found.

G. Towell and J. Shavlik, "Knowledge-based artificial neural networks, " Artif. Intell., vol. 70, pp. 119--165, 1994.

No context found.

Towell, G., Shavlik, J.: Knowledge-based artificial neural networks. Artificial Intelligence 70 (1994)

No context found.

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119--165, 1994.

No context found.

Geoffrey G. Towell and Jude W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119-- 165, 1994.

No context found.

G. G. Towell and J. W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119--165, 1994.

No context found.

G. G. Towell and J. W. Shavlik. Knowledge-Based Artificial Neural Networks. Artificial Intelligence, 70(1-2):119--165, 1994.

No context found.

G. G. Towell and J. W. Shavlik; " Knowledge-Based Artificial Neural Networks" ; Artificial Intelligence, Vol. 70; 1994.

No context found.

G. G. Towell and J. W. Shavlik; " Knowledge-Based Artificial Neural Networks" ; Artificial Intelligence, Vol. 70, pp.119-165; 1994.

No context found.

G. Towell and J. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119--165, 1994.

No context found.

Towell G, Shavlik J. Knowledge-based artificial neural networks. Artif Intell 1994;70:119--65.

No context found.

Geo#rey G. Towell and Jude W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119--165, 1994.

No context found.

G. G. Towell and J. W. Shavlik. Knowledge-Based Artificial Neural Networks. Artificial Intelligence, 70(1-2):119--165, 1994.

No context found.

G. G. Towell and J. W. Shavlik, "Knowledge-based artificial neural networks, " Artif. Intell., vol. 70, pp. 119--165, 1994.

No context found.

Geo#rey G. Towell and Jude W. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119--165, 1994.

No context found.

G. Towell and J. Shavlik, "Knowledge-based artificial neural networks, " Artif. Intell., vol. 70, pp. 119--165, 1994.

No context found.

G. Towell and J. Shavlik. Knowledge-based artificial neural networks. Artificial Intelligence, 70(1--2):119--165, 1994.

No context found.

Towell, G., & Shavlik, J. W. (1994). Knowledge-Based Artificial Neural Networks. Artificial Intelligence, 70(1-2), 119-165.

No context found.

GG. Towell, J.W. Shavlik. Knowledge-Based Artificial Neural Networks. Artificial Intelligence, 70:119-165, 1994.

First 50 documents  Next 50

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