Skip to main content

Yves Kodratoff

Followers

987

Following

34

Co-authors

11

Public Views

I am a retired CNRS ‘research director’ who actually directed some 68 PhD during my 45 years career at CNRS.I have been very active since the 75’s in order to promote AI in France and Machine Learning in Europe. Since my retirement in 2008, my research field has been split in two (almost) independent topics. For one, I go on working on AI by trying to analyze complex interacting systems made of ‘symbiotic’ modules (i. e. one module cannot work properly in isolation of the other modules). These have to evolve by themselves (or sometimes under some human supervision) until a solution ‘emerges’ out of the many tries and failure recovery performed by the whole system. This can be labelled as ‘AI for scientific creativity’. This work is done in close collaboration with my former student, Marta Franova, who thus became my leader since she started this research long ago.For two, I got involved into the understanding of a literature dubbed as an ‘Old Norse’ one. I try to develop a fundamentally emic approach inspired by the strong connection underlined by C. G. Jung between myths and what he calls ‘collective unconscious’. This amounts to work at detecting the (by Jung’s definition) unconscious archetypes hidden in Old Norse literature with, so to say, as much empathy as possible for the characters met in the myth.

less

Emanuele Covino

Cristian Negrescu

University of Calgary

Centre National de la Recherche Scientifique / French National Centre for Scientific Research

Andreas Weiermann

Ghent University

Sylvain Schmitz

Université Paris Diderot

Victor Yodaiken

Michael Norrish

CSIRO

InterestsView All (7)

Uploads

Machine Learning: Papers published before 2008 by Yves Kodratoff

Comparing Machine Learning and Knowledge Discovery in DataBases: An Application to Knowledge Discovery in Texts

This chapter has two goals. The first goal is to compare Machine Learning (ML) and Knowledge Disc... more This chapter has two goals. The first goal is to compare Machine Learning (ML) and Knowledge Discovery in Data (KDD, also often called Data Mining, DM) insisting on how much they actually differ. In order to make my ideas somewhat easier to understand, and as an illustration, I will include a description of several research topics that I find relevant to KDD and to KDD only. The second goal is to show that the definition I give of KDD can be almost directly applied to text analysis, and that will lead us to a very restrictive definition of Knowledge Discovery in Texts (KDT). I will provide a compelling example of a real-life set of rules obtained by what I call KDT techniques.

Machine Learning and Knowledge Acquisition: Integrated Approaches

by Yves Kodratoff and Gheorghe Tecuci

Machine Learning for Object Recognition and Scene Analysis

International Journal of Pattern Recognition and Artificial Intelligence, 1994

Learning Based on Conceptual Distance

IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988

In this paper we present a new approach to concept learning from examples and concept learning by... more In this paper we present a new approach to concept learning from examples and concept learning by observation, which is based on an intuitive notion of conceptual distance between examples (concepts) and combines symbolical and numerical methods. Our approach is supported by the observation that very different examples generalize to an expression that is very far from each of them, while identical examples generalize to themselves. Therefore, a generalization of two examples, as well as the process of obtaining this generalization, represents indications of the conceptual distance between the examples. Following this idea we propose some domain independent and intuitively justified estimates for the conceptual distance. Usually however, a set of examples may be characterized by several generalizations, each suggesting a certain conceptual distance. The minimum of these is taken as the estimation of the real conceptual distance. Moreover, the corresponding generalization is recommended as the one to be made by the learning system because this generalization has the desirable property of reflecting the greatest number of common features of the examples. We also present a hierarchical conceptual clustering algorithm which groups objects so that to maximize the cohesiveness (a reciprocal of the conceptual distance) of the clusters. We further show that conceptual clustering may improve learning from complex examples describing objects and the relations between them. The idea is that learning good generalizations of such examples requires matching the most similar objects which, in turn, requies a clustering of these objects. Finally we present a methodology of learning hierarchies of prototype objects which is a step toward automating the construction of knowledge bases for expert systems.

Knowledge Discovery in Texts: A Definition, and Applications

The first part of this paper will give a general view of Knowledge Discovery in Data (KDD) in ord... more The first part of this paper will give a general view of Knowledge Discovery in Data (KDD) in order to insist on how much it differs from the fields it stems from, and in some cases, how much it opposes them. The second part will a definition of Knowledge Discovery in Texts (KDT), as opposed to what is known presently under the name of information retricval, information extraction, or knowledge extraction. I will provide an example of a real-life set of rules obtained by what I want to define as KDT techniques.

Panel: Four Stances on Knowledge Acquisition and Machine Learning

by Yves Kodratoff, Ramon Lopez De Mantaras, Katharina Morik, and Enric Plaza

This paper consists of four condensed position papers given as an initial statement by five membe... more This paper consists of four condensed position papers given as an initial statement by five members of a panel session on Knowledge Acquisition and Machine Learning. This panel session was instigated by Yves Kodratoff in order to cover an issue that was raised in 1987 at the first workshop on knowledge acqusition for knowledge-based systems held at Reading (UK). It was at this meeting that those involved in machine learning and knowledge acquisition came to realise that a cultural gulf existed between two groups of people; people who seemed to be working in the same area. It has been the objective over the years at the meetings of both the European Working Sessions on Learning (EWSL) and the European Knowledge Acquisition Workshops (EKAW) to bridge the gap between these two cultures. This panel is intended to be a further step towards achieving this objective.

Noise and Knowledge Acquisition

In this paper we analyse how noise can affect Knowledge Acquisition from a Machine Learning persp... more In this paper we analyse how noise can affect Knowledge Acquisition from a Machine Learning perspective. We present some methods to detect and treat noise that goes beyond modulating numerical coefficients and show that noise cannot be viewed as a single entity. There are several different types of noise and noise is not only wrong information.

DISCIPLE1: Interactive Apprentice System in Weak Theory Fields

rj A WEAK THEORY DOMAIN The paper presents an interactive approach to learning apprentice sys te... more rj A WEAK THEORY DOMAIN The paper presents an interactive approach to learning apprentice sys tems for weak theory domains. The approach consists of a combination of teaming by analogy and learning by generalizing instances. One main point of this approach is that it uses the explanations drawn from an example, both to reduce the version space of me rules

Induction of Decision Trees from Complex Structured Data

Apprenticeship Learning in Imperfect Domain Theories

... As one may notice, the structure of General Rule 1 in Figure 19-6 is idcntic:.ll with the str... more ... As one may notice, the structure of General Rule 1 in Figure 19-6 is idcntic:.ll with the structure of Example 1 in Figure 19-5. Therefore, rule learning is reduced to learning the features that the objects &amp;amp;#x27;x&amp;amp;#x27;, &amp;amp;#x27;y&amp;amp;#x27;. and &amp;amp;#x27;z&amp;amp;#x27; should have so that the attach ment of &amp;amp;#x27;x&amp;amp;#x27; and &amp;amp;#x27;y ... (z ISA adhesive) &amp;amp;amp; ...

The Central Role of Explannations in DISCIPLE

DISCIPLE is a Knowledge Acquisition system that contains several learning mechanisms as recognize... more DISCIPLE is a Knowledge Acquisition system that contains several learning mechanisms as recognized by Machine Learning. The central mechanism in DICIPLE is the one of explanations which is used in all the learning modes of DISCIPLE. When using the Explanation-Based mode of learning, an explanation points at the most relevant features of the examples. When using the Analogy-Based mode of learning, the explanations are used to generate instances analogous to those provided by the user. When using the Similarity-Based mode of learning, the explanations are "examples" among which similarities are looked for. The final result of DISCIPLE is the description of the validity domain of the variables contained in the rules. Since the users always provides totally instantiated rules, the system must automatically variabilize them, and then must find the validity domain of these variables by asking "clever" questions to the user. Given a particular (instantiated) rule by its user, the system will look in its Knowledge Base for possible explanations of this rule, and ask the user to validate them. The set of explanations validated by the user is then used as a set of (almost) sufficient conditions for the application of the instantiated rule.

Careful Generalization for Concept Learning

Generalization and Noise

International Journal of Human-computer Studies / International Journal of Man-machine Studies, 1987

ABSTRACT This paper describes a research project which aims at applying Machine Learning (ML) tec... more ABSTRACT This paper describes a research project which aims at applying Machine Learning (ML) techniques to ease Knowledge Acquisition (KA) for Knowledge Based systems. Since noise in real life data has a drastic effect on ML, we examine in detail problems connected with noise. The learning system integrates two apparently distinct approaches: the numeric approach and the symbolic approach. It uses a filtering mechanism that is driven by statistical information and by comparison between several sources of knowledge (multi-expertise and experts-users “cross-examination” of input). The system also attempts to generate concepts which are resilient to noise and to improve the language of description. While it is usually thought that noise prevents using ML techniques in real applications, we attempt to show that on the contrary existing techniques can be stretched to cope with noise and to obtain better results than traditional KA techniques.

Introduction to machine learning

Can machine learning solve my problem

Applied Artificial Intelligence, 1994

... &amp;amp;amp;quot;No taako defined&amp;amp;amp;#x27; PRODIGY GAUL femJly ofsystems Fi... more ... &amp;amp;amp;quot;No taako defined&amp;amp;amp;#x27; PRODIGY GAUL femJly ofsystems Figure 5. ML KBS adaptive use. ... KBG computadln 1IfAKEY classifiers&amp;amp;amp;quot; FOIL sequence APT AQ CLINT Identity predicates NewlDMARVIN or relevant features C4.5 8om.aolae OverlapPlna I. ...

Building a machine learning toolbox

Some Techniques for Recursion Removal from Recursive Functions

ACM Transactions on Programming Languages and Systems, 1982

... If f is transformed into g, then either f and g compute the Same defined value or f and g bot... more

Knowledge Acquisition and Learning by Experience-The Role of Case-Specific Knowledge

by Gheorghe Tecuci and Yves Kodratoff

As knowledge-based systems are addressing increasingly complex domains, their roles are shifting ... more As knowledge-based systems are addressing increasingly complex domains, their roles are shifting from classical expert systems to interactive assistants. To develop and maintain such systems, an integration of thorough knowledge acquisition procedures and sustained learning from experience is called for. A knowledge level modeling perspective has shown to be useful for analyzing the various types of knowledge related to a particular domain and set of tasks, and for constructing the models of knowledge contents needed in an intelligent system. To be able to meet the requirements of future systems with respect to robust competence and adaptive learning behavior, particularly in open and weak theory domains, a stronger emphasis should be put on the combined utilization of casespecific and general domain knowledge. In this chapter we present a framework for integrating KA and ML methods within a total knowledge modeling cycle, favoring an iterative rather than a top down approach to system development. Recent advances in the area of case-based reasoning provide a suitable basis. Focus is put on the knowledge maintenance part, for which a case-based method for learning from experience is described and exemplified by existing systems. Our own approach to integration of casespecific and general domain knowledge (the CREEK system) is briefly sketched, and used as a context for discussing case-based approaches to knowledge modeling in general.

Text and Web Mining Approaches in Order to Build Specialized Ontologies

Journal of Digital Information, 2009

The Role of Thematic and Concept Texture In Scientific Text: Comparing Native and Non-Native Writers of English* Publication Details

en.scientificcommons.org

Comparing Machine Learning and Knowledge Discovery in DataBases: An Application to Knowledge Discovery in Texts

This chapter has two goals. The first goal is to compare Machine Learning (ML) and Knowledge Disc... more This chapter has two goals. The first goal is to compare Machine Learning (ML) and Knowledge Discovery in Data (KDD, also often called Data Mining, DM) insisting on how much they actually differ. In order to make my ideas somewhat easier to understand, and as an illustration, I will include a description of several research topics that I find relevant to KDD and to KDD only. The second goal is to show that the definition I give of KDD can be almost directly applied to text analysis, and that will lead us to a very restrictive definition of Knowledge Discovery in Texts (KDT). I will provide a compelling example of a real-life set of rules obtained by what I call KDT techniques.

Machine Learning and Knowledge Acquisition: Integrated Approaches

by Yves Kodratoff and Gheorghe Tecuci

Machine Learning for Object Recognition and Scene Analysis

International Journal of Pattern Recognition and Artificial Intelligence, 1994

Learning Based on Conceptual Distance

IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988

In this paper we present a new approach to concept learning from examples and concept learning by... more In this paper we present a new approach to concept learning from examples and concept learning by observation, which is based on an intuitive notion of conceptual distance between examples (concepts) and combines symbolical and numerical methods. Our approach is supported by the observation that very different examples generalize to an expression that is very far from each of them, while identical examples generalize to themselves. Therefore, a generalization of two examples, as well as the process of obtaining this generalization, represents indications of the conceptual distance between the examples. Following this idea we propose some domain independent and intuitively justified estimates for the conceptual distance. Usually however, a set of examples may be characterized by several generalizations, each suggesting a certain conceptual distance. The minimum of these is taken as the estimation of the real conceptual distance. Moreover, the corresponding generalization is recommended as the one to be made by the learning system because this generalization has the desirable property of reflecting the greatest number of common features of the examples. We also present a hierarchical conceptual clustering algorithm which groups objects so that to maximize the cohesiveness (a reciprocal of the conceptual distance) of the clusters. We further show that conceptual clustering may improve learning from complex examples describing objects and the relations between them. The idea is that learning good generalizations of such examples requires matching the most similar objects which, in turn, requies a clustering of these objects. Finally we present a methodology of learning hierarchies of prototype objects which is a step toward automating the construction of knowledge bases for expert systems.

Knowledge Discovery in Texts: A Definition, and Applications

The first part of this paper will give a general view of Knowledge Discovery in Data (KDD) in ord... more The first part of this paper will give a general view of Knowledge Discovery in Data (KDD) in order to insist on how much it differs from the fields it stems from, and in some cases, how much it opposes them. The second part will a definition of Knowledge Discovery in Texts (KDT), as opposed to what is known presently under the name of information retricval, information extraction, or knowledge extraction. I will provide an example of a real-life set of rules obtained by what I want to define as KDT techniques.

Panel: Four Stances on Knowledge Acquisition and Machine Learning

by Yves Kodratoff, Ramon Lopez De Mantaras, Katharina Morik, and Enric Plaza

This paper consists of four condensed position papers given as an initial statement by five membe... more This paper consists of four condensed position papers given as an initial statement by five members of a panel session on Knowledge Acquisition and Machine Learning. This panel session was instigated by Yves Kodratoff in order to cover an issue that was raised in 1987 at the first workshop on knowledge acqusition for knowledge-based systems held at Reading (UK). It was at this meeting that those involved in machine learning and knowledge acquisition came to realise that a cultural gulf existed between two groups of people; people who seemed to be working in the same area. It has been the objective over the years at the meetings of both the European Working Sessions on Learning (EWSL) and the European Knowledge Acquisition Workshops (EKAW) to bridge the gap between these two cultures. This panel is intended to be a further step towards achieving this objective.

Noise and Knowledge Acquisition

In this paper we analyse how noise can affect Knowledge Acquisition from a Machine Learning persp... more In this paper we analyse how noise can affect Knowledge Acquisition from a Machine Learning perspective. We present some methods to detect and treat noise that goes beyond modulating numerical coefficients and show that noise cannot be viewed as a single entity. There are several different types of noise and noise is not only wrong information.

DISCIPLE1: Interactive Apprentice System in Weak Theory Fields

rj A WEAK THEORY DOMAIN The paper presents an interactive approach to learning apprentice sys te... more rj A WEAK THEORY DOMAIN The paper presents an interactive approach to learning apprentice sys tems for weak theory domains. The approach consists of a combination of teaming by analogy and learning by generalizing instances. One main point of this approach is that it uses the explanations drawn from an example, both to reduce the version space of me rules

Induction of Decision Trees from Complex Structured Data

Apprenticeship Learning in Imperfect Domain Theories

... As one may notice, the structure of General Rule 1 in Figure 19-6 is idcntic:.ll with the str... more ... As one may notice, the structure of General Rule 1 in Figure 19-6 is idcntic:.ll with the structure of Example 1 in Figure 19-5. Therefore, rule learning is reduced to learning the features that the objects &amp;amp;#x27;x&amp;amp;#x27;, &amp;amp;#x27;y&amp;amp;#x27;. and &amp;amp;#x27;z&amp;amp;#x27; should have so that the attach ment of &amp;amp;#x27;x&amp;amp;#x27; and &amp;amp;#x27;y ... (z ISA adhesive) &amp;amp;amp; ...

The Central Role of Explannations in DISCIPLE

DISCIPLE is a Knowledge Acquisition system that contains several learning mechanisms as recognize... more DISCIPLE is a Knowledge Acquisition system that contains several learning mechanisms as recognized by Machine Learning. The central mechanism in DICIPLE is the one of explanations which is used in all the learning modes of DISCIPLE. When using the Explanation-Based mode of learning, an explanation points at the most relevant features of the examples. When using the Analogy-Based mode of learning, the explanations are used to generate instances analogous to those provided by the user. When using the Similarity-Based mode of learning, the explanations are "examples" among which similarities are looked for. The final result of DISCIPLE is the description of the validity domain of the variables contained in the rules. Since the users always provides totally instantiated rules, the system must automatically variabilize them, and then must find the validity domain of these variables by asking "clever" questions to the user. Given a particular (instantiated) rule by its user, the system will look in its Knowledge Base for possible explanations of this rule, and ask the user to validate them. The set of explanations validated by the user is then used as a set of (almost) sufficient conditions for the application of the instantiated rule.

Careful Generalization for Concept Learning

Generalization and Noise

International Journal of Human-computer Studies / International Journal of Man-machine Studies, 1987

ABSTRACT This paper describes a research project which aims at applying Machine Learning (ML) tec... more ABSTRACT This paper describes a research project which aims at applying Machine Learning (ML) techniques to ease Knowledge Acquisition (KA) for Knowledge Based systems. Since noise in real life data has a drastic effect on ML, we examine in detail problems connected with noise. The learning system integrates two apparently distinct approaches: the numeric approach and the symbolic approach. It uses a filtering mechanism that is driven by statistical information and by comparison between several sources of knowledge (multi-expertise and experts-users “cross-examination” of input). The system also attempts to generate concepts which are resilient to noise and to improve the language of description. While it is usually thought that noise prevents using ML techniques in real applications, we attempt to show that on the contrary existing techniques can be stretched to cope with noise and to obtain better results than traditional KA techniques.

Introduction to machine learning

Can machine learning solve my problem

Applied Artificial Intelligence, 1994

... &amp;amp;amp;quot;No taako defined&amp;amp;amp;#x27; PRODIGY GAUL femJly ofsystems Fi... more ... &amp;amp;amp;quot;No taako defined&amp;amp;amp;#x27; PRODIGY GAUL femJly ofsystems Figure 5. ML KBS adaptive use. ... KBG computadln 1IfAKEY classifiers&amp;amp;amp;quot; FOIL sequence APT AQ CLINT Identity predicates NewlDMARVIN or relevant features C4.5 8om.aolae OverlapPlna I. ...

Building a machine learning toolbox

Some Techniques for Recursion Removal from Recursive Functions

ACM Transactions on Programming Languages and Systems, 1982

... If f is transformed into g, then either f and g compute the Same defined value or f and g bot... more

Knowledge Acquisition and Learning by Experience-The Role of Case-Specific Knowledge

by Gheorghe Tecuci and Yves Kodratoff

As knowledge-based systems are addressing increasingly complex domains, their roles are shifting ... more As knowledge-based systems are addressing increasingly complex domains, their roles are shifting from classical expert systems to interactive assistants. To develop and maintain such systems, an integration of thorough knowledge acquisition procedures and sustained learning from experience is called for. A knowledge level modeling perspective has shown to be useful for analyzing the various types of knowledge related to a particular domain and set of tasks, and for constructing the models of knowledge contents needed in an intelligent system. To be able to meet the requirements of future systems with respect to robust competence and adaptive learning behavior, particularly in open and weak theory domains, a stronger emphasis should be put on the combined utilization of casespecific and general domain knowledge. In this chapter we present a framework for integrating KA and ML methods within a total knowledge modeling cycle, favoring an iterative rather than a top down approach to system development. Recent advances in the area of case-based reasoning provide a suitable basis. Focus is put on the knowledge maintenance part, for which a case-based method for learning from experience is described and exemplified by existing systems. Our own approach to integration of casespecific and general domain knowledge (the CREEK system) is briefly sketched, and used as a context for discussing case-based approaches to knowledge modeling in general.

Text and Web Mining Approaches in Order to Build Specialized Ontologies

Journal of Digital Information, 2009

The Role of Thematic and Concept Texture In Scientific Text: Comparing Native and Non-Native Writers of English* Publication Details

en.scientificcommons.org

Cartesian Systemic Emergence and its Resonance Thinking Facet: Why and How?

by Marta Franova and Yves Kodratoff

International Journal On Advances in Systems and Measurements, 2020

Cartesian Systemic Emergence (CSE) is concerned with strategic aspects relative to the conception... more Cartesian Systemic Emergence (CSE) is concerned with strategic aspects relative to the conception of Symbiotic Recursive Pulsative Systems intended to solve real-world problems handling control and prevention in incomplete domains. This work is performed to prepare fundamentals for designing automated tools that help to perform this complex task. This paper recalls the fundamental notions of CSE and presents the most important features of one particular way of thinking present in CSE. We call it ‘Resonance Thinking’. Resonance Thinking takes care of generating and handling experiments during CSE. The work presented is related to systems design, cognitive, and computation models of human creative reasoning mechanisms as well as to the ML approach called “Ultra-Strong Learning” for computer-assisted learning of CSE and RT.
Keywords— Cartesian Systemic Emergence; Symbiotic Recursive Pulsative Systems; Resonance Thinking; systems design; implementation of human reasoning mechanisms; Ultra-Strong Learning.

Resonance Thinking and Inductive Machine Learning

Proc. of The Fourteenth International Conference on Systems, 2019

This paper presents one of the symbiotic parts deemed necessary to complete our theory of compute... more This paper presents one of the symbiotic parts deemed necessary to complete our theory of computer systems design devoted to incomplete domains, here called Cartesian Systemic Emergence (CSE). CSE is dealing with one version of the concept of (semi)-automated creativity. It explains the links between our (formal) approach and the one of Michie's "Ultra strong Machine Learning" as recently implemented by S. Muggleton, U. Schmid et al.
Best Paper Award ICONS 2019
.

A Model of Pulsation for Evolutive Formalizing Incomplete Intelligent Systems, 2018

Abstract— The notion of pulsation concerns a possibility of a particular kind of intelligent cont... more Abstract— The notion of pulsation concerns a possibility of a particular kind of intelligent controlled and secured evolution in dynamic real-world systems. It is therefore related to fundamentals in intelligent systems and applications as well as to the topic of intelligence by design. In this paper we present a model of pulsation based on Ackermann’s function. This brings more clarity to understanding Symbiotic Recursive Pulsative Systems that are important, for instance, for designing and implementing intelligent security systems or for automating robots’ programming in incomplete domains and unknown environments. One particular application for these systems is our Constructive Matching Methodology for automating program synthesis from formal specifications in incomplete domains.
Keywords-pulsation; Symbiotic Recursive Pulsative Systems; intelligent systems; intelligence by design; Ackermann's function; control; security; progress; practical completeness.

Symbiosis paper Intelli 2017

by Marta Franova and Yves Kodratoff

INTELLI 2017, The Sixth International Conference on Intelligent Systems and Applications, 2017

— The notion of pulsation concerns a possibility of a particular kind of intelligent controlled a... more — The notion of pulsation concerns a possibility of a particular kind of intelligent controlled and secured evolution in dynamic real-world systems. It is therefore related to fundamentals in intelligent systems and applications as well as to the topic of intelligence by design. In this paper we present a model of pulsation based on Ackermann's function. This brings more clarity to understanding Symbiotic Recursive Pulsative Systems that are important, for instance, for designing and implementing intelligent security systems or for automating robots' programming in incomplete domains and unknown environments. One particular application for these systems is our Constructive Matching Methodology for automating program synthesis from formal specifications in incomplete domains.

On Computational Creativity, ‘Inventing’ Theorem Proofs

Lecture Notes in Computer Science, 2009

We provide a precise illustration of what can be the idea of "computational creativity", that is,... more We provide a precise illustration of what can be the idea of "computational creativity", that is, the whole set of the methods by which a computer may simulate creativity. This paper is centered on the relationship between computational creativity and theorem proving. The basic tool for this kind of computational creativity is what we call an 'asset generator' a specification of which is given in section 5, followed by a short description of our methodology for the generation of assets in theorem proving. In a sense, our 'asset generation methodology' relies essentially on making explicit the logician's good sense while performing a recursion constructive proof. Our contribution is making explicit this good sense and making a systematic methodology of it.

Predicate synthesis from formal specifications: Using mathematical induction for finding the preconditions of theorems

Lecture Notes in Computer Science, 1993

On the mythical reality of 'quiet' Scandinavian draugar

On the mythical reality of ‘quiet’ Scandinavian draugar, 2019

There are not so many academic oriented works about 'quiet' Scandinavian ghosts. They are also of... more There are not so many academic oriented works about 'quiet' Scandinavian ghosts. They are also often called "mound-dwellers" instead of draugar who are famous and terrible. Both are nevertheless ghosts and our study tries to produce a categorization of 'the' quiet draugar by looking for features common to these quiet ghosts who did not terrorized their surroundings or stopped to did it.

Daring Hervör visits her fathers' burial mound

Hervör’s mytical life

This tells the most striking aspects of Hervör’s life at it is told in Hervarar saga ok Heiðreks.

Andvari, Loki and Fáfnir curse Nibelung's Gold

The curses of Nibelung’s gold, 2019

This paper provides a detailed Old Norse information related to the ‘curses’ assotiated to the Ni... more

English translation of Völuspá

Here is an nth translation of the Edda poem Völuspá. It is different from the other translations ... more Here is an nth translation of the Edda poem Völuspá. It is different from the other translations by three aspects.
1. I provide a literal translation that reveals my grammatical and vocabulary choices. It may look somewhat obscure but it is followed by a translation in a less acrobatic English that should clarify the intended meaning. From time to time two irreducible versions are possible and I will give them. All these various possible versions are presented from a ‘Heathen-centric’ point of view, i. e. the point of view of a committed Heathen.

2. In particular, some 20 specific comments associated to the relevant stanza, are devoted to counter or simply criticize the most significant claims to ‘Christian influences’ on Völuspá.

Hávamál English translation

English translation of Hávamál with Comments, 2019

Here is an English translation of Hávamál coming with lengthy comments associated to each stanza.... more Here is an English translation of Hávamál coming with lengthy comments associated to each stanza.
That nobody has been able to detect in it Christian influences vouched by the academic community makes of this poem a primary witness of Old Norse civilization.
Our own analysis insists on the existence in this poem of recursive thinking underlining the important function, for each individual, to analyze his/her own thought not only for the sake of pure introspection but mostly as a source of creativity for each of us.

Traduction commentée du Hávamál

Traduction commentée du Hávamál

Une traduction française qui contient de nombreux commentaires quant au sens du poème et au choix... more Une traduction française qui contient de nombreux commentaires quant au sens du poème et au choix des traductions possibles. Le fait que personne n’ait pu détecter d’influences chrétiennes résistant à la critique académique sur le Hávamál en fait un témoin de premier plan des caractères spécifiques à la culture norroise ancienne. Notre analyse insiste sur l’expression d’une pensée de nature récursive soulignant l’importance de l’analyse par l’individu de sa propre pensée qui ne soit pas une simple introspection mais une source créativité pour chacun.

Citations anglo-saxonnes du Wyrd

Une vingtaine de citations en contexte et commentées.

L'enlèvement de Iðunn : kenningar et heiti des strophes 2-13 du Haustlöng

Cet article fournit une analyse précise des kenningar contenus dans les strophes 2-13 du Haustlön... more Cet article fournit une analyse précise des kenningar contenus dans les strophes 2-13 du Haustlöng. Ils sont présentés deux fois : d’abord dans le contexte d poème où il apparaissent, ensuite dans le contexte de la mythologie norroise ancienne. Nos choix grammaticaux sont exposés et les connaissances inattendues qu’ils apportent au sujet de Þjazi et de Iðunn sont expliquées.

Iðunn's abduction: kenningar and heiti in Haustlöng stanzas 2-13

This paper provide a precise analysis of the kenningar contained in stanzas 2-13 of Haustlöng. Th... more This paper provide a precise analysis of the kenningar contained in stanzas 2-13 of Haustlöng. They are twice analyzed : first in the context of the poem, second in the context of old Norse mythology. The grammatical choices done are discussed and the unexpected knowledge they provide about Þjazi and Iðunn is put in evidence.

Galdr des Corbeaux d'Óðinn : HRAFNAGALDUR ÓÐINS

Une traduction nouvelle de Hrafnagaldur Óðins, complétée par de nombreuses notes quant au choix g... more

HRAFNAGALDUR ÓĐINS, (Óðinn's Ravens Galdr)

This is a new translation of Hrafnagaldur Óðins, with many comments relative to the grammatical c... more

VolusPagaFr.docx

Voici une nième traduction du poème Völuspá de l’Edda poétique. Elle est différente des autres ... more Voici une nième traduction du poème Völuspá de l’Edda poétique. Elle est différente des autres traductions par deux aspects.
1. Je vous donne une traduction mot à mot – elle-même souvent difficile à comprendre, je dois l’avouer – suivie d’une traduction en français normal qui donne au mieux le sens exact de la traduction mot à mot, ou bien qui explique les différentes versions possibles.
2. …les différentes versions possibles dans une perspective paganocentrique (j’ai inventé ce mot) c’est-à-dire que le point de vue présenté est celui d’un païen convaincu. Ceci change tout : les anglophones parlent d’une vision ‘christocentrique’ pour désigner un point vue pseudo athée (ou « objectif ») qui fait appel à des concepts fermement définis dans le cadre de la chrétienté.

Les Mabinogion traduits par Joseph Loth

Edition partielle de l'oeuvre publiée en 1913 (dernière version corrigée) de Joseph Loth : Les Ma... more

auraiceptFr.docx

Une version française de l'auraceipt na n-éces traduit par Calder (with an indirection to an Engl... more

Snorri’s Edda in its intellectual environment

This work attempts to re-open (with new arguments) a seemingly unwelcome topic within present da... more This work attempts to re-open (with new arguments) a seemingly unwelcome topic within present day academic research. While there is large flow of papers relating to the influence of Christianity on skaldic poetry, almost no one argues about its independence relative to Christianity.
We present new arguments to this old debate described by Faulkes (1993) as: “Discussion of the sources of Skáldskaparmál in the past has mainly been concerned by ... the accuracy with which Snorri reproduces pre-Christian tradition ... and the extent to which his work is influenced by the Christian, Latin thought."

La destinée et les façonnages dans l'Edda poétique

Introduction J'ai entrepris d'analyser avec un peu plus de précision ces deux concepts à partir d... more Introduction J'ai entrepris d'analyser avec un peu plus de précision ces deux concepts à partir des textes originaux qui en parlent. Cela suppose un gros travail de collection d'information. Je n'ai pas encore eu le temps d'analyser en détail la façon dont les sagas en parlent, ce qui reste donc une tâche à venir. J'ai utilisé l'ensemble du corpus eddique, et pour les sagas, 96 sagas incluant toutes celles classiques. La quantité de documents disponibles est assez importante pour qu'on puisse en tirer des conclusions claires. Les traductions sont les miennes, mais je les ai comparées à des traductions existantes qui tendent à oublier l'importance de la magie dans la représentation du monde (worldview) des anciens norrois. Les citations des mots örlög et sköp sont toujours placées en contexte, avec les vers du poème qui permettent de comprendre leur sens. Pour mes traductions, j'ai utilisé le dictionnaire étymologique de de Vries (publié en allemand seulement

Destiny and 'shapings' in the poetic Edda

We analyze the meaning of a little more than 60 stanzas belonging to the Poetic Edda. Our conclu... more

Huginn and Muninn or a few Archetypes of Old Norse Collective Unconscious

The polysemy of Óðinn's ravens names, Huginn and Muninn led us to the discovery of five archetype... more

Huginn et Muninn ou Quelques archétypes de l'inconscient collectif des anciens norrois

Une analyse détaillée de la polysémie des noms des deux corbeaux d'Óðinn nous a permis de découvr... more

La relation inconsciente mère-fils entre Frigg et Baldr

Not published, 2017

Dans le mythe de la mort de Baldr, trois 'détails' sont étranges. D'abord, les deux erreurs de Fr... more Dans le mythe de la mort de Baldr, trois 'détails' sont étranges. D'abord, les deux erreurs de Frigg : 1. avoir omis de faire prêter serment au gui de ne pas blesser Baldr sous prétexte « qu'il (le gui) était trop jeune », 2. avoir accepté de confier à une sorcière semble-t-il inconnue que Baldr n'était pas protégé contre le gui. Ensuite, le troisième détail est celui de l'erreur gravissime de Baldr qui s'est, apparemment de façon idiote, laissé aller à faire un jeu (enfantin) de la quasi-immortalité que sa mère lui avait offerte. Dans son ouvrage « Symbols of Transformations » p. 258-9, C. G. Jung se sert du mythe pour illustrer son argumentation au sujet d'un archétype des relations mère-fils.

An unconscious mother-son's relation between Frigg and Baldr

Not published

In the myth of Baldr’s death, three ‘details’ are hardly understandable. At first, Frigg’s two ... more In the myth of Baldr’s death, three ‘details’ are hardly understandable. At first, Frigg’s two mistakes:
1. omitting to ask the mistletoe to swear that it would not wound Baldr because (as she states) “it (mistletoe) was too young,”
2. agreeing to entrust an unknown witch with the knowledge that Baldr was not protected from mistletoe.
The third awkwardness is that Baldr could organize this childish show of quasi-immortality, granted to him by his mother.
The paper uses Jung's interpretation in order to explain these three details.

Introspection within Hávamál

Introspection within Hávamál, 2020

Stanzas 18, 20, 28, 63, 95 and 138 use words that evoke some kinds of introspective attitudes.

Quotations containing the word wyrd

These 21 citations all belong to web available Anglo-Saxon (AS) texts, they may be somewhat incom... more These 21 citations all belong to web available Anglo-Saxon (AS) texts, they may be somewhat incomplete. The translations I could consult looked to me so far from a literal translation that I took the freedom to restore them, while keeping their basic meaning, as given by an AS specialist. Word wyrd is presented in its context, together with a few words of the poem enabling to precise its particular meaning in this context.

Citations contenant le mot wyrd

Mes recherches n'ont porté que sur les textes anglo-saxons faciles à obtenir sur le web et donc e... more Mes recherches n'ont porté que sur les textes anglo-saxons faciles à obtenir sur le web et donc elles ne sont sans doute pas totalement exhaustives. Certaines des traductions que j'ai trouvées sont à mon sens vraiment très loin du mot à mot et je me suis permis de rétablir ce mot à mot, tout en conservant le sens général de la traduction faite par un spécialiste. Une vingtaine citation des mots wyrd sont placées en contexte, avec les vers du poème qui permettent de comprendre leur sens

Modèle de Pulsation pour une formalisation évolutive des systèmes intelligents incomplets

International audienceThe notion of pulsation concerns a possibility of a particular kind of inte... more International audienceThe notion of pulsation concerns a possibility of a particular kind of intelligent controlled and secured evolution in dynamic real-world systems. It is related to fundamentals in intelligent systems and applications as well as to the topic of intelligence by design. In this paper we present a model of pulsation based on Ackermann’s function. This brings more clarity to understanding Symbiotic Recursive Pulsative Systems that are important, for instance, for designing and implementing intelligent security systems or for automating robots’ programming in incomplete domains and unknown environments. One particular application for these systems is our Constructive Matching Methodology for automating program synthesis from formal specifications in incomplete domains

Cartesian Systemic Emergence - Tackling Underspecified Notions in Incomplete Domains

Pulsation is a model for a particular evolutive improvement that handles control and prevention i... more Pulsation is a model for a particular evolutive improvement that handles control and prevention in intelligent systems built following this ‘pulsative’ model in incomplete domains. Cartesian systemic emergence presented in this paper answers the question of strategic aspects of building such systems. The main problem to tackle here is working with informal specifications while aiming at their formalization in a way that is driven by the pulsation model

Resonance Thinking and Inductive Machine Learning

This paper presents one of the symbiotic parts deemed necessary to complete our theory of compute... more This paper presents one of the symbiotic parts deemed necessary to complete our theory of computer systems design devoted to incomplete domains, here called Cartesian Systemic Emergence (CSE). CSE is dealing with one version of the concept of (semi)-automated creativity. We call systems implementing this kind of creativity Symbiotic Recursive Pulsative Systems (SRPS). SRPS are intended to contribute to solving real-world problems in incomplete domains requiring control and prevention. CSE is concerned with strategic aspects of the conception of such SPRS. Each component of a SPRS has to be symbiotically linked to all the other components. This requirement is not very usual in Computer Science, hence we have to introduce notions that are not yet present in scientific vocabulary. This paper is devoted to the most important features of one particular way of thinking present in CSE. We call it ‘Resonance Thinking’ (RT). RT takes care of generating and handling experiments during CSE. We...

Le Défi fouilles de textes : quels paradigmes pour la reconnaissance d'auteurs ?

Les campagnes d'evaluation en traitement automatique du langage naturel et en informatique do... more Les campagnes d'evaluation en traitement automatique du langage naturel et en informatique documentaire sont devenues un passage oblige pour la reconnaissances des differentes techniques employees. Le Defi Fouille de Texte a pour objectif de permettre aux chercheurs du monde francophone de confronter leurs travaux avec un probleme, plus que primer une equipe, une methode, ou un outil. Dans cet article nous evoquons les diverses problematiques de la fouille de texte, a savoir la recherche d'information, l'extraction ou l'enrichissement de connaissances, la classificiation/categorisation de documents, la segmentation de textes, le profilage. La reconnaissance d'auteur, objet de ce premier defi, est une tâche complexe et composite qui necessite de traiter simultanement de la segmentation, de la categorisation et du profilage. L'idee generale est que la mise en place des defis est un outil de cartographie des diverses avancees en fouille de textes, et egalement u...

A Model of Pulsation for Evolutive Formalizing Incomplete Intelligent Systems

The notion of pulsation concerns a possibility of a particular kind of intelligent controlled and... more The notion of pulsation concerns a possibility of a particular kind of intelligent controlled and secured evolution in dynamic real-world systems. It is related to fundamentals in intelligent systems and applications as well as to the topic of intelligence by design. In this paper we present a model of pulsation based on Ackermann’s function. This brings more clarity to understanding Symbiotic Recursive Pulsative Systems that are important, for instance, for designing and implementing intelligent security systems or for automating robots’ programming in incomplete domains and unknown environments. One particular application for these systems is our Constructive Matching Methodology for automating program synthesis from formal specifications in incomplete domains.

Cartesian Systemic Emergence and its Resonance Thinking Facet: Why and How?

Cartesian Systemic Emergence (CSE) is concerned with strategic aspects relative to the conception... more Cartesian Systemic Emergence (CSE) is concerned with strategic aspects relative to the conception of Symbiotic Recursive Pulsative Systems intended to solve real-world problems handling control and prevention in incomplete domains. This work is performed to prepare fundamentals for designing automated tools that help to perform this complex task. This paper recalls the fundamental notions of CSE and presents the most important features of one particular way of thinking present in CSE. We call it 'Resonance Thinking'. Resonance Thinking takes care of generating and handling experiments during CSE. The work presented is related to systems design, cognitive, and computation models of human creative reasoning mechanisms as well as to the ML approach called "Ultra-Strong Learning" for computer-assisted learning of CSE and RT.

Cartesian Handling Informal Specifications in Incomplete Frameworks

This paper introduces and illustrates a fundament al notion, namely informal specification, for c... more This paper introduces and illustrates a fundament al notion, namely informal specification, for creating tools developed as symbiotic recursive pulsating systems (SRPS), in the framework of Inductive Theorem Proving and Intelligent Systems. It illustrates the use of this fundamental notion in scientific systemic creativity relative to theorem proving. We deal simultaneously with the meta-level design of a system that proves theorems automatically. Keywords-informal specification; intelligence by design; inductive theorem proving; Cartesian Intuitionism; symbiotic recursive systems; Constructive Matching Methodology.

Algorithmic Conceptualization of Tools for Proving by Induction « Unwinding » Theorems – A Case Study

Information flow security is an interesting and challenging problem in which the use of inductive... more Information flow security is an interesting and challenging problem in which the use of inductive theorem proving is required for dealing with the so-called ‘unwinding theorems’. This paper examines whether Constructive Matching Methodology (CMM) for proving theorems by induction can be used also in this industry oriented framework. We show here that nothing more than a minor extension is to be considered, namely handling non recursive functions and predicates. But many solutions for this problem are already available. However dealing with possibly incomplete recursive theories and their formulation is still a problem. This paper illustrates that CMM is well suited in both automated and non-automated versions in the context of information flow security.

Machine Learning for Object Recognition and Scene Analysis

International Journal of Pattern Recognition and Artificial Intelligence, 1994

Learning is a critical research field for autonomous computer vision systems. It can bring soluti... more Learning is a critical research field for autonomous computer vision systems. It can bring solutions to the knowledge acquisition bottleneck of image understanding systems. Recent developments of machine learning for computer vision are reported in this paper. We describe several different approaches for learning at different levels of the image understanding process, including learning 2-D shape models, learning strategic knowledge for optimizing model matching, learning for adaptive target recognition systems, knowledge acquisition of constraint rules for labelling and automatic parameter optimization for vision systems. Each approach will be commented on and its strong and weak points will be underlined. In conclusion we will suggest what could be the “ideal” learning system for vision.

Knowledge discovery in texts: A definition, and applications

Lecture Notes in Computer Science, 1999

The first part of this paper will give a general view of Knowledge Discovery in Data (KDD) in ord... more The first part of this paper will give a general view of Knowledge Discovery in Data (KDD) in order to insist on how much it differs from the lields it stems from, and in some cases, how much it opposes them. The second part will a definition of Knowledge Discovery in Texts (KDT), as opposed to what is known presently under the name of information retrieval, information extraction, or knowledge extraction. I will provide an example of a reallife set of rules obtained by what I want to define as KDT techniques.

High Level Imitation Learning in a Treasure Hunting Task

High level learning by imitation consists of learning to emulate high level human cognitive proce... more High level learning by imitation consists of learning to emulate high level human cognitive processes. We have been able to solve this problem within a specific setting, searching for a treasure in a maze. The human behaviour can be recorded as time series of observable variables, representing the portion of the human's problem-solving strategy which is apparent to an outside observer. These can then be subjected to several layers of constructive generalisation, until the strategy is abstracted from the original settings and can thus be applied to new ones.

Human Heuristics for a Team of Mobile Robots

2007 IEEE International Conference on Research, Innovation and Vision for the Future, 2007

This paper is at the crossroad of Cognitive Psychology and AI Robotics. It reports a cross-discip... more This paper is at the crossroad of Cognitive Psychology and AI Robotics. It reports a cross-disciplinary project concerned about implementing human heuristics within autonomous mobile robots. In the following, we address the problem of relying on human-based heuristics to endow a group of mobile robots with the ability to solve problems such as target finding in a labyrinth. Such heuristics may provide an efficient way to explore the environment and to decompose a complex problem into subtasks for which specific heuristics are efficient. We first present a set of experiments conducted with group of humans looking for a target with limited sensing capabilities solving. Then we describe the heuristics extracted from the observation and analysis of their behavior. Finally we implemented these heuristics within khepera-like autonomous mobile robots facing the same tasks. We show that the control architecture can be experimentally validated to some extent thanks to this approach.

Mechanical construction of a new efficient FLATTEN

ACM Lisp Bulletin, 1979

An efficient function FLATTEN is "well-known" and due to Mc CARTHY[5]. We give below a ... more

Reducing Uncertainty Through Learning

DISCIPLE: An integrated expert and learning system for weak theory domains

Disciple: An interative approach to learning apprentice systems

On the Simulation of an Art Work by a Markov Chain with the Aid of a Digital Computer

Leonardo, 1973

The author discusses mathematical analysis of works in the plastic arts by a Markov chain within ... more The author discusses mathematical analysis of works in the plastic arts by a Markov chain within the frame of the Monte Carlo method. The two procedures are discussed with particular references to their interrelations and their use in attempting to analyze artistic creativity with the aid of computers. He demonstrates that the two procedures can be applied together. Illustrations of several art works that were made in this way are presented. L’auteur présente une analyse mathématique d’oeuvres des arts plastiques, sous forme d’une chaîne de Markov croissant dans le cadre d’une méthode de Monte Carlo. Les deux procédés sont expliqués, particulièrement en fonction de leurs rapports mutuels et de la manière dont on peut les utiliser pour tenter d’analyser la création artistique à l’aide d’ordinateurs. L’auteur démontre que les deux méthodes peuvent être appliquées conjointement. Il fournit plusieurs illustrations d’oeuvres d’art réalisées de cette manière.

Metamorphic Thinking in Cartesian Systemic Emergence

international conference on systems, Apr 18, 2021

Cartesian Systemic Emergence (CSE) is a theory developed in order to formalize the process of a h... more Cartesian Systemic Emergence (CSE) is a theory developed in order to formalize the process of a human creation relative to particular problem-solving systems. Its final aim is to enable to design (semi-) automated tools that favor this creation. The creation process considered here concerns the context of informally specified systems and working with underspecified notions in incomplete environments. The aim of this paper is to show that this nonstandard research approach is epistemically justified. In particular, the paper focuses on justifying inspirationconduciveness of the CSE-experiments-generation-andhandling process relative to the invention of primitive notions needed in order to create the intended problem-solving system.

Case-Based Reasoning

Machine Learning and Its Applications, 2001

ABSTRACT

Conceptual Distance-Based Learning