Helmut Schnelle
Ruhr University Bochum
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Helmut Schnelle.
Archive | 2004
Helmut Schnelle
Turing’s precise notion of computation implies three types of constraints: (1) sub-computational constraints requiring error-free components of the machines, (2) con-computational constraints according to which the machines are not influenced by situational distraction, and (3) trans-computational limits (in Godel’s sense). In contrast, human thought is not marked by these constraints and limitations. This is discussed with reference to thoughts of von Neumann and Weyl, Carnap and Bar-Hillel and, finally, Godel and Wang.
Konnektionismus in Artificial Intelligence und Kognitionsforschung. Proceedings 6. Österreichische Artificial Intelligence-Tagung (KONNAI) | 1990
Rolf Wilkens; Helmut Schnelle
Most connectionist networks in present use employ built-in learning algorithms. Their major disadvantage however is their inability to handle complex knowledge structures. In this paper a connectionist network is presented which can handle highly developed cognitive structures not by making use of learning processes but by the integration of a powerful initial structure. Parsing of context-free phrase structure grammars shall exemplify the workings of the network.
wissensbasierte systeme, . internationaler gi-kongress | 1989
Helmut Schnelle
“Positive results in the field of parallel computation will be based on existing knowledge of computational structures, not on mystical emergent properties of unstructured networks.”(FELDMAN,1988). There are no convincing arguments against this statement, at least not in the field of complicated perceptual behaviour, such as high-level vision and high-level speech (syntax, semantics). As a consequence, we should study ways of translating existing knowledge of computational structures into massively parallel systems, in particular into connectionist systems.
Archive | 1994
Rolf Wilkens; Helmut Schnelle
Die kognitive Linguistik hat die Aufgabe, die Sprache als ein kognitives Phanomen zu beschreiben. Es herrscht keine Einigkeit daruber, in welcher Weise diese Aufgabe zu losen ist. Nach klassischer Auffassung genugt es, kognitive Leistungen, speziell Prozesse der Sprachverarbeitung, als symbolverarbeitende Systeme zu beschreiben. Es ist uberflussig, strukturelle Eigenschaften der Implementierung anzugeben, ihre Erklarung wurde nichts zur Charakterisierung der Kognition beitragen. Zur Erlauterung wird haufig gesagt, das Verhaltnis der kognitiven Beschreibung zur physischen Beschreibung entspreche dem Verhaltnis von Software und Hardware. Zur Bestimmung der Bedingungen korrekten Rechnens reichten ja auch Regeln der Algebra und ihnen zugeordnete Algorithmen. Die genaue Implementierung konne den Konstrukteuren von Computern uberlassen bleiben. Die Plausibilitat der Argumentation setzt die Konzeption des sequentiellen von Neumann-Computers als leitendes Konstruktionsprinzip voraus, scheint aber implizit davon auszugehen, das parallele und effiziente Implementierungen auch keine Revision dieser Immunitat des symbolischen Kognitionskonzepts gegen Implementierung erzwingen konnten.
Archive | 1994
Helmut Schnelle
A discussion of basic properties of a new, neurobiologically motivated linguistics based formally on dynamical systems theorizing rather than on combinatorial symbolic structuring.
Konnektionismus in Artificial Intelligence und Kognitionsforschung. Proceedings 6. Österreichische Artificial Intelligence-Tagung (KONNAI) | 1990
Helmut Schnelle
Connectionism provides a challenge to current theorizing and modeling in theoretical cognitive linguistics. It has been acknowledged that cognitive linguistics must provide an integrated description and explanation of the knowledge (a) of language structure, (b) of language use, (c) of language acquisition and (d) of the physical mechanisms implementing knowledge, use and acquisition. So far cognitive linguistics has mostly concentrated on the definition of the knowledge of language structure in terms of systems of rules or a set of principles (constraints). Language use is currently specified by algorithms for parsers etc. This is usually confined to uses of regular language. Basic ideas have been advanced about how acquisition of language regularity could be specified in terms of a set of parameters for a system of basic principles for natural languages. No ideas have been given about the empirically justifiable or falsifiable structure of embodiment in physical mechanisms, though it has been acknowledge that we would be concerned with the structure of a biological system.
meeting of the association for computational linguistics | 1984
Helmut Schnelle
This contribution attempts a conceptual and practical introduction into the principles of wiring or constructing special machines for language processing tasks instead of programming a universal machine. Construction would in principle provide higher descriptive adequacy in computationally based linguistics. After all, our heads do not apply programs on stored symbol arrays but are appropriately wired for understanding or producing language.
international conference on computational linguistics | 1990
Helmut Schnelle; Rolf Wilkens
Description of a connectionist implementation of an Earley parser.
Acta Linguistica Hafniensia | 1997
Helmut Schnelle
international conference on computational linguistics | 1988
Helmut Schnelle; Garry Cottrell; Paradip Dey; Joachim Diederich; Peter A. Reich; Lokendra Shastri; Akinori Yonezawa