# The Noos Representation Language

## Phd Contents

Available copy of all the Phd [Postscript]

## 1 Introduction

Chapter 1 presents The motivations, goals, and main contributions of the thesis.

#### 1.1 | #### Motivation |

#### 1.2 | #### Goals and contributions of the thesis |

#### 1.3 | #### Structure of the thesis |

## 2 Background

Chapter 2 reviews the main research relevant to our thesis and discusses
their contributions and limitations. We present research on knowledge
modeling methodologies for analyzing and developing knowledge systems;
research on integrated architectures for exploring the relationships among
problem solving, learning, and knowledge representation; research on
reflective representation languages for providing introspection
capabilities on knowledge systems; and finally, we discuss the role of
reflection in learning.

#### 2.1 | #### Knowledge modeling frameworks |

#### 2.2 | #### Integrated architectures |

#### 2.3 | #### Reflective representation languages |

#### 2.4 | #### Introspective learning |

#### 2.5 | #### Conclusions |

## 3 The Noos Approach

Chapter 3 presents the Noos representation language. The language is introduced
incrementally. First, The Noos modeling framework is presented. Next, the
basic elements of the language such as descriptions, refinement,
references, methods, and the basic inference are described. Then, the
reflective capabilities of Noos are described introducing elements such as
metalevels, tasks, reflective operations, reification, and reinstantiation.
Next, a declarative mechanism for decision making about sets of
alternatives, called *preferences*, is presented. Finally, the complete
Noos inference engine is described.

#### 3.1 | #### The Noos modeling framework |

#### 3.2 | #### The Noos language |

#### 3.3 | #### Reflection |

#### 3.4 | #### Preferences |

#### 3.5 | #### Inference in Noos |

#### 3.6 | #### Summary |

## 4 Memory, Experience, and Learning

Chapter 4 presents the Noos capabilities for reasoning about experience and the
integration of learning and problem solving. First, the notion of
*episodic memory* is introduced. Then, introspective mechanisms
such as *retrieval* and *perspectives* are presented. Next,
three different families of learning techniques such as case-based
reasoning, inductive learning, and analytical learning, are presented with
examples of how they have been integrated in Noos.

#### 4.1 | #### Episodic knowledge in Noos |

#### 4.2 | #### Retrieval |

#### 4.3 | #### Prespectives |

#### 4.4 | #### Reasoning and learning |

#### 4.5 | #### Case-based reasoning |

#### 4.6 | #### Inductive learning |

#### 4.7 | #### Analytical learning |

#### 4.8 | #### Summary |

## 5 Noos Formalization

Chapter 5 presents feature terms, a formalism for describing the Noos language.
Feature terms are introduced using a syntax notation based on the Lambda N
Calculus. Then, using the work on feature structures a semantics based on
the notion of partial descriptions is presented. The results obtained by
the research on feature structures are also adapted for providing several
equivalent representations of feature terms.
Chapter 5 introduces a formalism for describing preferences based on the
notion of pre-orders. Two kinds of basic operations are defined over
preferences: *preference operations* , that take a set of source
elements and an ordering criterion and build a preference (a partially
ordered set), and *preference combination operations* , that take
two preferences and a combination criterion and build a new preference.

Next, using feature terms *perspectives* are defined. Perspectives
are formalized as second order feature terms that denote sets of terms.

Finally, we describe formally the inference in Noos using Descriptive Dynamic
Logic, a propositional dynamic logic for describing and comparing
reflective knowledge systems.

#### 5.1 | #### Basic notions of Lambda N Calculus |

#### 5.2 | #### Noos formal syntax |

#### 5.3 | #### Translation rules from Noos to Lambda N |

#### 5.4 | #### Using variables in feature terms |

#### 5.5 | #### Semantics |

#### 5.6 | #### Term subsumption |

#### 5.7 | #### Representing feature terms as labeled graphs |

#### 5.8 | #### Understanding feature terms as clauses |

#### 5.9 | #### Evaluable feature terms |

#### 5.10 | #### Preferences |

#### 5.11 | #### Perspectives |

#### 5.12 | #### Descriptive Dynamic Logic |

#### 5.13 | #### Modeling Noos inference using DDL |

#### 5.14 | #### Summary |

## 6 Applications

Chapter 6 provides a set of examples of how diverse applications have been
developed using Noos by several persons at the IIIA. Specifically, the chapter
present six applications developed using Noos: CHROMA,
SPIN, SHAM, GYMEL, SaxEx,and
NoosWeb.

#### 6.1 | #### CHROMA |

#### 6.2 | #### SPIN |

#### 6.3 | #### SHAM |

#### 6.4 | #### GYMEL |

#### 6.5 | #### SAXEX |

#### 6.6 | #### NoosWeb |

#### 6.7 | #### Summary |

## 7 Conclusions and Future Work

Chapter 7 summarizes the main contributions of the thesis and discuss
further directions of research.

#### 7.1 | #### The Noos language and feature terms |

#### 7.2 | #### Memory and learning |

#### 7.3 | #### Methods and applications |

#### 7.4 | #### Future work |

## A The Noos Development Environment

Appendix A presents the Noos development environment including development
facilities such as browsing, tracing, and the extension of the built-in
methods.

#### A.1 | #### Defining feature terms in Noos |

#### A.2 | #### The predefined sort hierarchy of Noos |

#### A.3 | #### Episodic memory |

#### A.4 | #### Browsing |

#### A.5 | #### Tracing |

#### A.6 | #### Extending built-in methods |

## B Glossary

Appendix B provides a glossary of the main concepts introduced in this
thesis.

## C The Noos Syntax

Appendix C presents the complete syntax of the Noos language, using BNF
notation, and the collection of compact descriptions for built-in methods.

## D Built-in Methods

Appendix D presents the complete list of Noos built-in methods describing their
required features and their functionality.