Refine
Year of publication
- 1999 (397) (remove)
Document Type
- Preprint (397) (remove)
Keywords
- Case-Based Reasoning (10)
- Fallbasiertes Schliessen (5)
- Location Theory (5)
- case-based problem solving (5)
- Abstraction (4)
- Fallbasiertes Schließen (4)
- Knowledge Acquisition (4)
- Internet (3)
- Knowledge acquisition (3)
- Maschinelles Lernen (3)
Faculty / Organisational entity
- Kaiserslautern - Fachbereich Informatik (206)
- Kaiserslautern - Fachbereich Mathematik (121)
- Kaiserslautern - Fachbereich Physik (51)
- Kaiserslautern - Fachbereich Elektrotechnik und Informationstechnik (9)
- Fraunhofer (ITWM) (6)
- Kaiserslautern - Fachbereich Wirtschaftswissenschaften (2)
- Kaiserslautern - Fachbereich Maschinenbau und Verfahrenstechnik (1)
- Universitätsbibliothek (1)
Abstract: We present experimental and theoretical results of a detailed study of laser-induced continuum structures (LICS) in the photoionization continuum of helium out of the metastable state 2s^1 S_0. The continuum dressing with a 1064 nm laser, couples the same region of the continuum to the 4s^1 S_0 state. The experimental data, presented for a range of intensities, show pronounced ionization suppression (by asmuch as 70% with respect to the far-from-resonance value) as well as enhancement, in a Beutler-Fano resonance profile. This ionization suppression is a clear indication of population trapping mediated by coupling to a contiuum. We present experimental results demonstrating the effect of pulse delay upon the LICS, and for the behavior of LICS for both weak and strong probe pulses. Simulations based upon numerical solution of the Schrödinger equation model the experimental results. The atomic parameters (Rabi frequencies and Stark shifts) are calculated using a simple model-potential method for the computation of the needed wavefunctions. The simulations of the LICS profiles are in excellent agreement with experiment. We also present an analytic formulation of pulsed LICS. We show that in the case of a probe pulse shorter than the dressing one the LICS profile is the convolution of the power spectra of the probe pulse with the usual Fano profile of stationary LICS. We discuss some consequences of deviation from steady-state theory.
An interrupter for use in a daisy-chained VME bus interrupt system has beendesigned and implemented as an asynchronous sequential circuit. The concur-rency of the processes posed a design problem that was solved by means of asystematic design procedure that uses Petri nets for specifying system and in-terrupter behaviour, and for deriving a primitive flow table. Classical designand additional measures to cope with non-fundamental mode operation yieldeda coded state-machine representation. This was implemented on a GAL 22V10,chosen for its hazard-preventing structure and for rapid prototyping in studentlaboratories.
Hexagonal BN films have been deposited by rf-magnetron sputtering with simultaneous ion plating. The elastic properties of the films grown on silicon substrates under identical coating conditions have been de-termined by Brillouin light scattering from thermally excited surface phonons. Four of the five independent elastic constants of the deposited material are found to be c11 = 65 GPa, c13 = 7 GPa, c33 = 92 GPa and c44 = 53 GPa exhibiting an elastic anisotropy c11/c33 of 0.7. The Young's modulus determined with load indenta-tion is distinctly larger than the corresponding value taken from Brillouin light scattering. This discrepancy is attributed to the specific morphology of the material with nanocrystallites embedded in an amorphous matrix.
We present an inference system for clausal theorem proving w.r.t. various kinds of inductivevalidity in theories specified by constructor-based positive/negative-conditional equations. The reductionrelation defined by such equations has to be (ground) confluent, but need not be terminating. Our con-structor-based approach is well-suited for inductive theorem proving in the presence of partially definedfunctions. The proposed inference system provides explicit induction hypotheses and can be instantiatedwith various wellfounded induction orderings. While emphasizing a well structured clear design of theinference system, our fundamental design goal is user-orientation and practical usefulness rather thantheoretical elegance. The resulting inference system is comprehensive and relatively powerful, but requiresa sophisticated concept of proof guidance, which is not treated in this paper.This research was supported by the Deutsche Forschungsgemeinschaft, SFB 314 (D4-Projekt)
Die Verwendung von existierenden Planungsansätzen zur Lösung von realen Anwendungs- problemen führt meist schnell zur Erkenntnis, dass eine vorliegende Problemstellung im Prinzip zwar lösbar ist, der exponentiell anwachsende Suchraum jedoch nur die Behandlung relativ kleiner Aufgabenstellungen erlaubt. Beobachtet man jedoch menschliche Planungsexperten, so sind diese in der Lage bei komplexen Problemen den Suchraum durch Abstraktion und die Verwendung bekannter Fallbeispiele als Heuristiken, entscheident zu verkleinern und so auch für schwierige Aufgabenstellungen zu einer akzeptablen Lösung zu gelangen. In dieser Arbeit wollen wir am Beispiel der Arbeitsplanung ein System vorstellen, das Abstraktion und fallbasierte Techniken zur Steuerung des Inferenzprozesses eines nichtlinearen, hierarchischen Planungssystems einsetzt und so die Komplexität der zu lösenden Gesamtaufgabe reduziert.
We describe a hybrid architecture supporting planning for machining workpieces. The archi- tecture is built around CAPlan, a partial-order nonlinear planner that represents the plan already generated and allows external control decision made by special purpose programs or by the user. To make planning more efficient, the domain is hierarchically modelled. Based on this hierarchical representation, a case-based control component has been realized that allows incremental acquisition of control knowledge by storing solved problems and reusing them in similar situations.
We describe a hybrid case-based reasoning system supporting process planning for machining workpieces. It integrates specialized domain dependent reasoners, a feature-based CAD system and domain independent planning. The overall architecture is build on top of CAPlan, a partial-order nonlinear planner. To use episodic problem solving knowledge for both optimizing plan execution costs and minimizing search the case-based control component CAPlan/CbC has been realized that allows incremental acquisition and reuse of strategical problem solving experience by storing solved problems as cases and reusing them in similar situations. For effective retrieval of cases CAPlan/CbC combines domain-independent and domain-specific retrieval mechanisms that are based on the hierarchical domain model and problem representation.
While most approaches to similarity assessment are oblivious of knowledge and goals, there is ample evidence that these elements of problem solving play an important role in similarity judgements. This paper is concerned with an approach for integrating assessment of similarity into a framework of problem solving that embodies central notions of problem solving like goals, knowledge and learning.
Contrary to symbolic learning approaches, which represent a learned concept explicitly, case-based approaches describe concepts implicitly by a pair (CB; sim), i.e. by a measure of similarity sim and a set CB of cases. This poses the question if there are any differences concerning the learning power of the two approaches. In this article we will study the relationship between the case base, the measure of similarity, and the target concept of the learning process. To do so, we transform a simple symbolic learning algorithm (the version space algorithm) into an equivalent case- based variant. The achieved results strengthen the hypothesis of the equivalence of the learning power of symbolic and case-based methods and show the interdependency between the measure used by a case-based algorithm and the target concept.
Im Bereich der Expertensysteme ist das Problemlösen auf der Basis von bekannten Fallbeispielen ein derzeit sehr aktuelles Thema. Auch für Diagnoseaufgaben gewinnt der fallbasierte Ansatz immer mehr an Bedeutung. In diesem Papier soll der im Rahmen des Moltke -Projektes1 an der Universität Kaiserslautern entwickelte fallbasierte Problemlöser Patdex/22 vorgestellt werden. Ein erster Prototyp, Patdex/1, wurde bereits 1988 entwickelt.
Patdex is an expert system which carries out case-based reasoning for the fault diagnosis of complex machines. It is integrated in the Moltke workbench for technical diagnosis, which was developed at the university of Kaiserslautern over the past years, Moltke contains other parts as well, in particular a model-based approach; in Patdex where essentially the heuristic features are located. The use of cases also plays an important role for knowledge acquisition. In this paper we describe Patdex from a principal point of view and embed its main concepts into a theoretical framework.
Formalismen und Anschauung
(1999)
In der Philosophie ist es selbstverständlich, daß Autoren, die Erkenntnisse früherer Philosophen weitergeben oder kommentieren, die Originalliteratur kennen und sich in ihrer Argumentation explizit auf bestimmte Stellen in den Originaldarstellungen beziehen. In der Technik dagegen ist es allgemein akzeptierte Praxis, daß Autoren von Lehrbüchern, in denen Erkenntnisse früherer Forscher dargestellt oder kommentiert werden, nicht die Originaldarstellungen zugrunde legen, sondern sich mit den Darstellungen in der Sekundärliteratur begnügen. Man denke an die Erkenntnisse von Boole oder Maxwell, die in sehr vielen Lehrbüchern der Digitaltechnik bzw. der theoretischen Elektrotechnik vermittelt werden, ohne daß die Autoren dieser Lehrbücher auf die Originalschriften von Boole oder Maxwell Bezug nehmen. Dagegen wird man wohl kaum ein Buch über Erkenntnisse von Aristoteles oder Kant finden, dessen Autor sich nicht explizit auf bestimmte Stellen in den Schriften dieser Philosophen bezieht.
Die systemtheoretische Begründung für die Einführung des Zustandsbegriffs findet man im Mosaik-stein "Der Zustandsbegriff in der Systemtheorie". Während sich die dortige Betrachtung sowohl mitkontinuierlichen als auch mit diskreten Systemen befaßt, wird hier die Betrachtung auf diskrete Sy-steme beschränkt.
Umgangssprachlich wurde das Wort Daten schon gebraucht, lange bevor der Computer erfundenwurdeund die AbkürzungEDV für "Elektronische Datenverarbeitung" in die Alltagssprache gelangte.So sagte beispielsweise der Steuerberater zu seinem Klienten: "Bevor ich Ihre Steuererklärung fertigmachen kann, brauche ich von Ihnen noch ein paar Daten." Oder der Straßenbaureferent einer Stadtschrieb an den Oberbürgermeister: "Für die Entscheidung, welche der beiden in Frage stehenden Stra-ßen vorrangig ausgebaut werden soll, müssen wir noch eine Datenerhebung durchführen." Bei diesenDaten ging es zwar oft um Zahlen - Geldbeträge, Anzahl der Kinder, Anzahl der Beschäftigungsmo-nate, gezählte Autos - , aber eine Gleichsetzung von Daten mit Zahlen wäre falsch. Zum einen wärenZahlen ohne mitgelieferte Wörter wie Monatseinkommen, Kinderzahl u.ä. für den Steuerberater nutz-los, zum anderen will das Finanzamt u.a. auch den Arbeitgeber des Steuerpflichtigen wissen, und dazumuß eine Adresse angegeben werden, aber keine Zahl.
Für die Systemtheorie ist der Begriff Zustand ein sehr zentraler Begriff. Das Wort "Zustand" wird um-gangssprachlich recht häufig verwendet, aber wenn man die Leute fragen würde, was sie denn meinen,wenn sie das Wort Zustand benützen, dann würde man sicher nicht die präzise Definition bekommen,die man für die Systemtheorie braucht.
In diesem Aufsatz geht es um eine Klassifikation von Programmen nach zwei orthogonalen Kriterien.Programm und Software werden dabei nicht als Synonyme angesehen; Programm sein wird hiergleichgesetzt mit ausführbar sein, d.h. etwas ist dann und nur dann ein Programm, wenn man die Fragebeantworten kann, was es denn heißen solle, dieses Etwas werde ausgeführt. Es gibt durchaus Softwa-regebilde, bezüglich derer diese Frage keinen Sinn hat und die demzufolge auch keine Programme sind - beispielsweise eine Funktions - oder eine Klassenbibliothek.Klassifikation ist von Nutzen, wenn sie Vielfalt überschaubarer macht - die Vielfalt der Schüler einergroßen Schule wird überschaubarer, wenn die Schüler "klassifiziert" sind, d.h. wenn sie in ihren Klas-senzimmern sitzen. Die im folgenden vorgestellte Klassifikation soll die Vielfalt von Programmenüberschaubarer machen.
Bei der Programmierung geht es in vielfältiger Form um Identifikation von Individuen: Speicherorte,Datentypen, Werte, Klassen, Objekte, Funktionen u.ä. müssen definierend oder selektierend identifiziert werden.Die Ausführungen zur Identifikation durch Zeigen oder Nennen sind verhältnismäßig kurz gehalten,wogegen der Identifikation durch Umschreiben sehr viel Raum gewidmet ist. Dies hat seinen Grunddarin, daß man zum Zeigen oder Nennen keine strukturierten Sprachformen benötigt, wohl aber zumUmschreiben. Daß die Betrachtungen der unterschiedlichen Formen funktionaler Umschreibungen soausführlich gehalten sind, geschah im Hinblick auf ihre Bedeutung für die Begriffswelt der funktionalen Programmierung. Man hätte zwar die Formen funktionaler Umschreibungen auch im Mosaikstein "Programmzweck versus Programmform" im Kontext des dort dargestellten Konzepts funktionaler Programme behandeln können, aber der Autor meint, daß der vorliegende Aufsatz der angemessenerePlatz dafür sei.
One of the problems of autonomous mobile systems is the continuous tracking of position and orientation. In most cases, this problem is solved by dead reckoning, based on measurement of wheel rotations or step counts and step width. Unfortunately dead reckoning leads to accumulation of drift errors and is very sensitive against slippery. In this paper an algorithm for tracking position and orientation is presented being nearly independent from odometry and its problems with slippery. To achieve this results, a rotating range-finder is used, delivering scans of the environmental structure. The properties of this structure are used to match the scans from different locations in order to find their translational and rotational displacement. For this purpose derivatives of range-finder scans are calculated which can be used to find position and orientation by crosscorrelation.
A map for an autonomous mobile robot (AMR) in an indoor environment for the purpose ofcontinuous position and orientation estimation is discussed. Unlike many other approaches, this map is not based on geometrical primitives like lines and polygons. An algorithm is shown , where the sensordata of a laser range finder can be used to establish this map without a geometrical interpretation of the data. This is done by converting single laser radar scans to statistical representations of the environ-ment, so that a crosscorrelation of an actu al converted scan and this representative results into the actual position and orientation in a global coordinate system. The map itsel f is build of representative scansfor the positions where the AMR has been, so that it is able to find its position and orientation by c omparing the actual scan with a scan stored in the map.
We tested the GYROSTAR ENV-05S. This device is a sensor for angular velocity. There- fore the orientation must be calculated by integration of the angular velocity over time. The devices output is a voltage proportional to the angular velocity and relative to a reference. The test where done to find out under which conditions it is possible to use this device for estimation of orientation.
Starting from the uniqueness question for mixtures of distributions this review centers around the question under which formally weaker assumptions one can prove the existence of SPLIFs, in other words perfect statistics and tests. We mention a couple of positive and negative results which complement the basic contribution of David Blackwell in 1980. Typically the answers depend on the choice of the set theoretic axioms and on the particular concepts of measurability.
We study a model for learning periodic signals in recurrent neural networks proposed by Doya and Yoshizawa [7] that can be considered as a model for temporal pattern memory in animal motoric systems. A network receives an external oscillatory input and adjusts its weights so that this signal can be reproduced approximately as the network output after some time. We use tools from adaptive control theory to derive an algorithm for weight matrices with special structure. If the input is generated by a network of the same structure the algorithm converges globally and does not exhibit the deficiencies of the back-propagation based approach of Doya and Yoshizawa under a persistency of excitation condition. This simple algorithm can also be used for open loop identification under quite restructive assumptions. The persistency of excitation condition cannot be proven even for the matrices with special structure but for a 3d system. For higher dimensional systems we give connections to the theory of linear time-varying systems where this condition is generically true (under assumption which are also needed in the time-invariant case). However, we cannot show that the linearized system related to the nonlinear neural network fulfills these generic assumptions.
Problem specifications for classical planners based on a STRIPS-like representation typically consist of an initial situation and a partially defined goal state. Hierarchical planning approaches, e.g., Hierarchical Task Network (HTN) Planning, have not only richer representations for actions but also for the representation of planning problems. The latter are defined by giving an initial state and an initial task network in which the goals can be ordered with respect to each other. However, studies with a specification of the domain of process planning for the plan-space planner CAPlan (an extension of SNLP) have shown that even without hierarchical domain representation typical properties called goal orderings can be identified in this domain that allow more efficient and correct case retrieval strategies for the case-based planner CAPlan/CbC. Motivated by that, this report describes an extension of the classical problem specifications for plan-space planners like SNLP and descendants. These extended problem specifications allow to define a partial order on the planning goals which can interpreted as an order in which the solution plan should achieve the goals. These goal ordering can theoretically and empirically be shown to improve planning performance not only for case-based but also for generative planning. As a second but different way we show how goal orderings can be used to address the control problem of partial order planners. These improvements can be best understood with a refinement of Barrett's and Weld's extended taxonomy of subgoal collections.
Real world planning tasks like manufacturing process planning often don't allow to formalize all of the relevant knowledge. Especially, preferences between alternatives are hard to acquire but have high influence on the efficiency of the planning process and the quality of the solution. We describe the essential features of the CAPlan planning architecture that supports cooperative problem solving to narrow the gap caused by absent preference and control knowledge. The architecture combines an SNLP-like base planner with mechanisms for explict representation and maintenance of dependencies between planning decisions. The flexible control interface of CAPlan allows a combination of autonomous and interactive planning in which a user can participate in the problem solving process. Especially, the rejection of arbitrary decisions by a user or dependency-directed backtracking mechanisms are supported by CAPlan.
Software development is becoming a more and more distributed process, which urgently needs supporting tools in the field of configuration management, software process/w orkflow management, communication and problem tracking. In this paper we present a new distributed software configuration management framework COMAND. It offers high availabilit y through replication and a mechanism to easily change and adapt the project structure to new business needs. To better understand and formally prove some properties of COMAND, we have modeled it in a formal technique based on distributed graph transformations. This formalism provides an intuitive rule-based description technique mainly for the dynamic behavior of the system on an abstract level. We use it here to model the replication subsystem.
If \(A\) generates a bounded cosine function on a Banach space \(X\) then the negative square root \(B\) of \(A\) generates a holomorphic semigroup, and this semigroup is the conjugate potential transform of the cosine function. This connection is studied in detail, and it is used for a characterization of cosine function generators in terms of growth conditions on the semigroup generated by \(B\). This characterization relies on new results on the inversion of the vector-valued conjugate potential transform.
The thermal equilibrium state of a bipolar, isothermal quantum fluid confined to a bounded domain \(\Omega\subset I\!\!R^d,d=1,2\) or \( d=3\) is the minimizer of the total energy \({\mathcal E}_{\epsilon\lambda}\); \({\mathcal E}_{\epsilon\lambda}\) involves the squares of the scaled Planck's constant \(\epsilon\) and the scaled minimal Debye length \(\lambda\). In applications one frequently has \(\lambda^2\ll 1\). In these cases the zero-space-charge approximation is rigorously justified. As \(\lambda \to 0 \), the particle densities converge to the minimizer of a limiting quantum zero-space-charge functional exactly in those cases where the doping profile satisfies some compatibility conditions. Under natural additional assumptions on the internal energies one gets an differential-algebraic system for the limiting \((\lambda=0)\) particle densities, namely the quantum zero-space-charge model. The analysis of the subsequent limit \(\epsilon \to 0\) exhibits the importance of quantum gaps. The semiclassical zero-space-charge model is, for small \(\epsilon\), a reasonable approximation of the quantum model if and only if the quantum gap vanishes. The simultaneous limit \(\epsilon =\lambda \to 0\) is analyzed.
Convex Operators in Vector Optimization: Directional Derivatives and the Cone of Decrease Directions
(1999)
The paper is devoted to the investigation of directional derivatives and the cone of decrease directions for convex operators on Banach spaces. We prove a condition for the existence of directional derivatives which does not assume regularity of the ordering cone K. This result is then used to prove that for continuous convex operators the cone of decrease directions can be represented in terms of the directional derivatices . Decrease directions are those for which the directional derivative lies in the negative interior of the ordering cone K. Finally, we show that the continuity of the convex operator can be replaced by its K-boundedness.
In diesem Aufsatz wird die Arbeitsweise eines Werkzeuges dargestellt, mit dessen Hilfedie Analyse von Feature-Interaktionen in Intelligenten (Telefon-)Netzwerken unterstütztwird. Dieses Werkzeug basiert auf einem von uns entwickelten formalen Lösungsansatz, deraus einem geeigneten Spezifikationsstil, aus einem formalen Kriterium zur Erkennung vonFeature-Interaktionen und aus einer Methode zur Auflösung der erkannten Feature- Interaktionen besteht. Das Werkzeug führt eine statische Analyse von Estelle-Spezifikationendurch und erkennt dabei potientielle Feature-Interaktionen sowie nichtausführbare Transitionen. Darüberhinaus kann es die erkannten nichtausführbaren Transitionen zur Optimierung aus der Spezifikation entfernen. Wir erläutern zunächst kurz den zugrundeliegendenAnsatz und beschreiben danach die Anwendung auf Estelle anhand der Funktionsweisedes Werkzeuges.
Die Verfahren der Induktiven Logischen Programmierung (ILP) [Mug93] haben die Aufgabe, aus einer Menge von positiven Beispielen E+, einer Menge von negativen Beispielen E und dem Hintergrundwissen B ein logisches Programm P zu lernen, das aus einer Menge von definiten Klauseln C : l0 l1, : : : ,ln besteht. Da der Hypothesenraum für Hornlogik unendlich ist, schränken viele Verfahren die Hypothesensprache auf eine endliche ein. Auch wird oft versucht, die Hypothesensprache so einzuschränken, dass nur Programme gelernt werden können, für die die Konsistenz entscheidbar ist. Eine andere Motivation, die Hypothesensprache zu beschränken, ist, dass das Wissen über das Zielprogramm, das schon vorhanden ist, ausgenutzt werden soll. So sind für bestimmte Anwendungen funktionsfreie Hypothesenklauseln ausreichend, oder es ist bekannt, dass das Zielprogramm funktional ist.
We present detailed studies of the enhanced coercivity of exchange-bias bilayer Fe/MnPd, both experimentally and theoretically. We have demonstrated that the existence of large higher-order anisotropies due to exchange coupling between different Fe and MnPd layers can account for the large increase of coercivity in Fe/MnPd system. The linear dependence of coercivity on inverse Fe thickness are well explained by a phenomenological model by introducing higher-order anisotropy terms into the total free energy of the system.
The mathematical modelling of problems in science and engineering leads often to partial differential equations in time and space with boundary and initial conditions.The boundary value problems can be written as extremal problems(principle of minimal potential energy), as variational equations (principle of virtual power) or as classical boundary value problems.There are connections concerning existence and uniqueness results between these formulations, which will be investigated using the powerful tools of functional analysis.The first part of the lecture is devoted to the analysis of linear elliptic boundary value problems given in a variational form.The second part deals with the numerical approximation of the solutions of the variational problems.Galerkin methods as FEM and BEM are the main tools. The h-version will be discussed, and an error analysis will be done.Examples, especially from the elasticity theory, demonstrate the methods.
The asymptotic behaviour of a singular-perturbed two-phase Stefan problem due to slow diffusion in one of the two phases is investigated. In the limit the model equations reduce to a one-phase Stefan problem. A boundary layer at the moving interface makes it necessary to use a corrected interface condition obtained from matched asymptotic expansions. The approach is validated by numerical experiments using a front-tracking method.
The asymptotic analysis of IBVPs for the singularly perturbed parabolic PDE ... in the limit epsilon to zero motivate investigations of certain recursively defined approximative series ("ping-pong expansions"). The recursion formulae rely on operators assigning to a boundary condition at the left or the right boundary a solution of the parabolic PDE. Sufficient conditions for uniform convergence of ping-pong expansions are derived and a detailed analysis for the model problem ... is given.
Orderings on polynomial interpretations of operators represent a powerful technique for proving thetermination of rewriting systems. One of the main problems of polynomial orderings concerns thechoice of the right interpretation for a given rewriting system. It is very difficult to develop techniquesfor solving this problem. Here, we present three new heuristic approaches: (i) guidelines for dealingwith special classes of rewriting systems, (ii) an algorithm for choosing appropriate special polynomialsas well as (iii) an extension of the original polynomial ordering which supports the generation ofsuitable interpretations. All these heuristics will be applied to examples in order to illustrate theirpractical relevance.
An important property and also a crucial point ofa term rewriting system is its termination. Transformation or-derings, developed by Bellegarde & Lescanne strongly based on awork of Bachmair & Dershowitz, represent a general technique forextending orderings. The main characteristics of this method aretwo rewriting relations, one for transforming terms and the otherfor ensuring the well-foundedness of the ordering. The centralproblem of this approach concerns the choice of the two relationssuch that the termination of a given term rewriting system can beproved. In this communication, we present a heuristic-based al-gorithm that partially solves this problem. Furthermore, we showhow to simulate well-known orderings on strings by transformationorderings.
High frequency switching of single domain, uniaxial magnetic particles is discussed in terms of transition rates controlled by a small transverse bias field. It is shown that fast switching times can be achieved using bias fields an order of magnitude smaller than the effective anisotropy field. Analytical expressions for the switching time are derived in special cases and general configurations of practical interest are examined using numerical simulations.
Die Induktive Logische Programmierung (ILP) ist ein Forschungsgebiet, das Techniken aus dem Maschinellen Lernen und der Logischen Programmierung vereint. Sie untersucht das klassische Problem induktiven Lernens aus klassifizierten Beispielen im Rahmen der Hornlogik erster Stufe. Inzwischen gibt es eine grosse Zahl verschiedener Ansätze für dieses Lernproblem, die sich hauptsächlich in der Suchrichtung im Hypothesenraum, den Generalisierungs- und Spezialisierungsoperatoren und den verwendeten nichtlogischen Beschränkungen (Bias) unterscheiden. Der Vergleich und die Integration dieser verschiedenen Ansätze war die Hauptmotivation für die Entwicklung des Systems MILES. MILES ist eine Programmierumgebung für die ILP, die neben Mechanismen zur Repräsentation und Verwaltung von Beispielen, Hintergrundwissen und Hypothesen einen Werkzeugkasten mit einem Grossteil der bekannten Generalisierungs-, Spezialisierungs- und Reformulierungsoperatoren enthält. Eine generische Kontrolle erlaubt, verschiedene dieser Operatoren in einen spezifischen ILP-Algorithmus zu integrieren. In diesem Beitrag wird ein kurzer Überblick über die Repräsentation, die Operatoren und die Kontrolle von MILES gegeben.
It is generally agreed that one of the most challenging issues facing the case-based reasoning community is that of adaptation. To date the lion's share of CBR research has concentrated on the retrieval of similar cases, and the result is a wide range of quality retrieval techniques. However, retrieval is just the first part of the CBR equation, because once a similar case has been retrieved it must be adapted. Adaptation research is still in its earliest stages, and researchers are still trying to properly understand and formulate the important issues. In this paper I describe a treatment of adaptation in the context of a case-based reasoning system for software design, called Deja Vu. Deja Vu is particularly interesting, not only because it performs automatic adaptation of retrieved cases, but also because it uses a variety of techniques to try and reduce and predict the degree of adaptation necessary.
We compare different notions of differentiability of a measure along a vector field on a locally convex space. We consider in the L2-space of a differ entiable measure the analoga of the classical concepts of gradient, divergence and Laplacian (which coincides with the OrnsteinUhlenbeck operator in the Gaussian case). We use these operators for the extension of the basic results of Malliavin and Stroock on the smoothness of finite dimensional image measures under certain nonsmooth mappings to the case of non-Gaussian measures. The proof of this extension is quite direct and does not use any Chaos-decomposition. Finally, the role of this Laplacian in the procedure of quantization of anharmonic oscillators is discussed.
The multiple-view modeling of a product in a design context is discussed in this paper. We study the existing approaches for multiple-view modeling of a product and we give a brief analysis of them. Then we propose our approach which incorporates the multiple-model approach in STEP standard current works based on a single model. We propose a meta-model inspired by this approach for a multiple-view design environment. Next, we validate this meta-model with a case study. Finally we conclude and give some perspectives of this work. Keywords: product data modeling, multiple-view modeling, product data integration, STEP, functional model.