High-level Knowledge Representation Research Articles

AI-Based Methods and Technologies to Develop Wearable Devices for Prosthetics and Predictions of Degenerative Diseases.

Neurodegenerative diseases, mainly amyotrophic lateral sclerosis, Parkinson, Alzheimer, and rarer diseases, have gained the attention of healthcare service providers due to their impact on the economy of countries where healthcare is a public service. These diseases increase with aging and affect the neuromotor cells and cognitive areas in the brain, causing serious disabilities in people affected by them.Early prediction of these syndromes is the first strategy to be implemented, then the developing of prostheses that rehabilitate motion and the primary cognitive functions. Prostheses could recover some important disabilities such as motion and aphasia, reduce the cost of assistance and increase the life quality of people affected by neurodegenerative diseases.Due to recent advances in the field of artificial intelligence (AI) (deep learning, brain-inspired computational paradigms, nonlinear predictions, neuro-fuzzy modeling), the early prediction of neurodegenerative diseases is possible using state-of-the-art computational technologies. The latest generation of artificial neural networks (ANNs) exploits capabilities such as online learning, fast training, high level knowledge representation, online evolution, learning by data and inferring rules.Wearable electronics is also developing rapidly and represents an important enabling technology to deploy physical and practical (noninvasive) devices using AI-based models for early prediction of neurodegenerative diseases and of intelligent prostheses.Here we describe how to apply advanced brain-inspired methods for inference and prediction, the evolving fuzzy neural network (EFuNN) paradigm and the spiking neural network (SNN) paradigm, and the system requirements to develop a wearable electronic prosthesis for functional rehabilitation.

Distributed graph queries over models@run.time for runtime monitoring of cyber-physical systems

Smart cyber-physical systems (CPSs) have complex interaction with their environment which is rarely known in advance, and they heavily depend on intelligent data processing carried out over a heterogeneous and distributed computation platform with resource-constrained devices to monitor, manage and control autonomous behavior. First, we propose a distributed runtime model to capture the operational state and the context information of a smart CPS using directed, typed and attributed graphs as high-level knowledge representation. The runtime model is distributed among the participating nodes, and it is consistently kept up to date in a continuously evolving environment by a time-triggered model management protocol. Our runtime models offer a (domain-specific) model query and manipulation interface over the reliable communication middleware of the Data Distribution Service (DDS) standard widely used in the CPS domain. Then, we propose to carry out distributed runtime monitoring by capturing critical properties of interest in the form of graph queries, and design a distributed graph query evaluation algorithm for evaluating such graph queries over the distributed runtime model. As the key innovation, our (1) distributed runtime model extends existing publish–subscribe middleware (like DDS) used in real-time CPS applications by enabling the dynamic creation and deletion of graph nodes (without compile time limits). Moreover, (2) our distributed query evaluation extends existing graph query techniques by enabling query evaluation in a real-time, resource-constrained environment while still providing scalable performance. Our approach is illustrated, and an initial scalability evaluation is carried out on the MoDeS3 CPS demonstrator and the open Train Benchmark for graph queries.

Functional and semantic roles in a high-level knowledge representation language

We describe in this paper a formalization of the notion of “role” that involves a clear separation between two very different sorts of roles. Semantic roles, like student or customer, are seen as (pre-defined) transitory properties that can be associated with (usually animate) entities. From a formal point of view, they can be represented as standard concepts to be placed into a specific branch of a particular ontology; they formalize the static and classificatory aspects of the notion of role. Functional roles must be used, instead, to model those pervasive and dynamic situations corresponding to events, activities, circumstances etc. that are characterized by spatio-temporal references; see, e.g., “John is now acting as a student”. They denote the specific function with respect to the global meaning of an event/situation/activity... that is performed by the entities involved in this event/situation... and formalize the dynamic and relational aspects of the notion of role. A functional role of the subject/agent/actor/protagonist... type is used to associate “John” with the notion of student or customer (semantic roles) during a specific time interval. Formally, functional roles are expressed as primitive symbols like subject, object, source, beneficiary. Semantic and functional roles interact smoothly when they are used to deal with challenging knowledge representation problems like the so-called “counting problem”, or when we need to set-up powerful inference rules whose atoms can directly denote complex situations. In this paper, the differentiation between semantic and functional roles will be illustrated from an narrative knowledge representation language (NKRL) point of view. NKRL is a high-level conceptual tool used for the computer-usable representation and management of the inner meaning of syntactically complex and semantically rich multimedia information. But, as we will see, the importance of this distinction goes well beyond its usefulness in a specific NKRL context. In particular, the use of functional roles is of paramount importance for the set-up of those evolved n-ary forms of knowledge representation that allow us to get rid from the limitations in expressiveness proper to the standard (binary) solutions.

Transfer metric learning for action similarity using high-level semantics

The goal of transfer learning is to exploit previous experiences and knowledge in order to improve learning in a novel domain. This is especially beneficial for the challenging task of learning classifiers that generalize well when only few training examples are available. In such a case, knowledge transfer methods can help to compensate for the lack of data. The performance and robustness against negative transfer of these approaches is influenced by the interdependence between knowledge representation and transfer type. However, this important point is usually neglected in the literature; instead the focus lies on either of the two aspects. In contrast, we study in this work the effect of various high-level semantic knowledge representations on different transfer types in a novel generic transfer metric learning framework. Furthermore, we introduce a hierarchical knowledge representation model based on the embedded structure in the semantic attribute space. The evaluation of the framework on challenging transfer settings in the context of action similarity demonstrates the effectiveness of our approach compared to state-of-the-art.

Representing and Reasoning About the Rules of General Games With Imperfect Information

A general game player is a system that can play previously unknown games just by being given their rules. For this purpose, the Game Description Language (GDL) has been developed as a high-level knowledge representation formalism to communicate game rules to players. In this paper, we address a fundamental limitation of state-of-the-art methods and systems for General Game Playing, namely, their being confined to deterministic games with complete information about the game state. We develop a simple yet expressive extension of standard GDL that allows for formalising the rules of arbitrary finite, n-player games with randomness and incomplete state knowledge. In the second part of the paper, we address the intricate reasoning challenge for general game-playing systems that comes with the new description language. We develop a full embedding of extended GDL into the Situation Calculus augmented by Scherl and Levesque's knowledge fluent. We formally prove that this provides a sound and complete reasoning method for players' knowledge about game states as well as about the knowledge of the other players.

Application of a Multi-Objective Approach and Sequential Covering Algorithm to the Fatigue Segment Classification Problem

In this paper, a multi-objective rule discovery approach is introduced to address the problem related to fatigue data editing. A set of rules are introduced to simplify the editing process, with which users can easily predict an appropriate level of damage for fatigue data segments subject to SAE1045 steel without need for a cycle-counting algorithm or stress analysis. A multi-objective approach called the elitist non-dominated sorting in genetic algorithm (NSGA-II) is modified and applied to find high-level knowledge representation IF-THEN rules in the eight-dimensional search space. Three possible outcomes represented with the labels very low, low and high are chosen for the rule consequent. Two conflicting classification objectives, the predictive accuracy and comprehensibility, are considered as the rule optimisation criteria. To accelerate the searching process, a sequential covering algorithm is applied in advance of the training session, which positively reduces data size and provides several simple rules. A set of seven high-accuracy and high-interpretability rules is discovered, and their uncomplicated implementation provides an alternative to the task of estimating the fatigue damage value and could be a part of a fatigue data editing process.

Active Knowledge for the Situation-driven Control of Product Definition

The current product model consists of features and unstructured contextual connections in order to relate features. The feature modifies the previous state of the product model producing contextual connections with previously defined features. Active knowledge is applied for the adaptive modification of product model features in the case of a changed situation or event. Starting from this state-of-the-art, the authors of this paper introduced a new method to achieve higher-level and more advanced active feature driven product model definition. As part of the related research program, new situation driven model definition processes and model entities are explained in this paper. Higher-level knowledge representation in the product model is motivated by a recent trend in industrial product modeling systems towards more advanced and efficient situation-based self- adaptive model generation. The proposed model represents one of the possible future ways of product model development for product lifecycle management (PLM) systems on the global or product level of decisions. Its implementation will be new application-oriented model entity generation and representation utilizing existing modeling resources in industrial PLM systems by use of application programming interfaces (API).

CoJACK: A high-level cognitive architecture with demonstrations of moderators, variability, and implications for situation awareness

Abstract We report a high-level architecture, CoJACK, that provides insights on behavior variability, situation awareness, and behavioral moderators. CoJACK combines Beliefs/Desires/Intentions (BDI) agents’ high-level knowledge representation and usability with several aspects of low-level cognitive architectures, including processing time predictions, errors, and traceability. CoJACK explores new areas for cognitive architectures, such as variability arising from moderators. It also allows aspects of situation awareness (SA) in a cognitive architecture to be explored. Its behavior and the effects of moderators on behavior are demonstrated in a simple adversarial environment. It provides lessons for other architectures including how to define, measure, and control variability due to individual and temporal aspects of cognition; the importance of SA and knowledge representations necessary to support complex SA; the potential for parameter sweeps and paths as measures of variability; and some of the complexities that will arise when aspects of moderators and SA are added to cognitive architectures.

A Cognition-Based Control System for Autonomous Robots

Most approaches to controlling autonomous systems require extensive pre-mission preparation, intensive effort by the human operator, or have strong limitations on the range of possible missions that can be accomplished. In this paper we describe an approach called Cognitive Patterns that promises to alleviate these challenges by replicating three key processes of human cognition—pattern generation, perception/action, and adaptation—and instantiating them in a new architecture which can then be embedded into an autonomous system. An early version of this approach connected high-level knowledge representations in an ontology with a robot's sensing and acting abilities. The advantages of this approach were then demonstrated in a simulation environment. A more refined version, based on lessons learned and called Cognitive Patterns Knowledge Generation, can deal with anomalies, unexpected events, and uncertainties, and is also described in terms of its components, their interactions, and benefits.

A General Game Description Language for Incomplete Information Games

A General Game Player is a system that can play previously unknown games given nothing but their rules. The Game Description Language (GDL) has been developed as a high-level knowledge representation formalism for axiomatising the rules of any game, and a basic requirement of a General Game Player is the ability to reason logically about a given game description. In this paper, we address the fundamental limitation of existing GDL to be confined to deterministic games with complete information about the game state. To this end, we develop an extension of GDL that is both simple and elegant yet expressive enough to allow to formalise the rules of arbitrary (discrete and finite) n-player games with randomness and incomplete state knowledge. We also show that this extension suffices to provide players with all information they need to reason about their own knowledge as well as that of the other players up front and during game play.

Linguistic quantifiers modeled by Sugeno integrals

Since quantifiers have the ability of summarizing the properties of a class of objects without enumerating them, linguistic quantification is a very important topic in the field of high level knowledge representation and reasoning. This paper introduces a new framework for modeling quantifiers in natural languages in which each linguistic quantifier is represented by a family of fuzzy measures, and the truth value of a quantified proposition is evaluated by using Sugeno's integral. This framework allows us to have some elegant logical properties of linguistic quantifiers. We compare carefully our new model of quantification and other approaches to linguistic quantifiers. A set of criteria for linguistic quantification was proposed in the previous literature. The relationship between these criteria and the results obtained in the present paper is clarified. Some simple applications of the Sugeno's integral semantics of quantifiers are presented.

Integrating cognition, perception and action through mental simulation in robots

We argue that many problems in robotics arise from the difficulty of integrating multiple knowledge representation and inference techniques. We describe an architecture that integrates disparate reasoning, planning, sensation and mobility algorithms by composing them from strategies for managing mental simulations. Since simulations are conducted by modules that include high-level knowledge representation and inference techniques in addition to algorithms for sensation and reactive mobility, cognition, perception and action are continually integrated. An implemented robot using this framework in object-tacking and human–robot interaction tasks demonstrates that knowledge representation and inference techniques enable more complex and flexible robot behavior.

ACCURACY, COMPREHENSIBILITY AND COMPLETENESS EVALUATION OF A FUZZY EXPERT SYSTEM

Using optimization tools such as genetic algorithms to construct a fuzzy expert system (FES) focusing only on its accuracy without considering the comprehensibility may not result in a system that produces understandable expressions. To exploit the transparency characteristics of FES for reasoning in a higher-level knowledge representation, a FES should provide high comprehensibility while preserving its accuracy. The completeness of fuzzy sets and rule structures should also be considered to guarantee that every data point has a response output. This paper proposes some quantitative measures for a FES to determine the degree of the accuracy, the comprehensibility of the fuzzy sets, and the completeness of fuzzy rule structure. These quantitative measures are then used as a fitness function for a genetic algorithm in optimally refining a FES.

Development of an object-oriented knowledge-based system for power electronic circuit design

A knowledge-based (expert) system (KBS) approach to power electronic circuit design is presented. The system, a power electronics and control tool, designated PECT, is written using Smalltalk-80, which integrates the artificial intelligence techniques of production rules for high-level knowledge representation. PECT embraces elements of power circuitry design and contains such features as selection of the circuit configuration, control facility, and power device best suited to a given application. It is interfaced with HSPICE, the commercial version of a general-purpose circuit simulation package, and a semiconductor power device library. Data retrieved from this library is converted into an object-oriented representation database. Analytical hierarchy process reasoning is performed on this data to aid the device selection process. System architecture as well as target design realization process are detailed. The criteria of development and the factors and requirements in building the PECT are also included. An example simulation is given representing a buck-boost converter using a gate turn-off thyristor (GTO) switch model.

An information presentation system for decision makers

This paper describes a new concept for interfacing operational decision makers with their information systems. It is intended to overcome the rigidities of current architectures by enabling the decision maker to direct, modify, and manipulate the presentation of information in real-time through a highly automated interface system. The current practice of specifying a decision maker's desires to a programmer, who must in turn produce a data gathering and formatting program that interfaces with some graphics system, is unsatisfactory for the real-time needs of an ever-changing decision environment.An experimental Information Presentation System has been implemented utilizing a high-level knowledge representation language. The system manipulates general descriptions of information presentation concepts such as coordinate systems, trnasformations, display regions, display objects, and the relationships between data objects and their depictions. This paper summarizes the current state of that system and the Information Presentation System concept.

An information presentation system for decision makers

This paper describes a new concept for interfacing operational decision makers with their information systems. It is intended to overcome the rigidities of current architectures by enabling the decision maker to direct, modify, and manipulate the presentation of information in real-time through a highly automated interface system. The current practice of specifying a decision maker's desires to a programmer, who must in turn produce a data gathering and formatting program that interfaces with some graphics system, is unsatisfactory for the real-time needs of an ever-changing decision environment.An experimental Information Presentation System has been implemented utilizing a high-level knowledge representation language. The system manipulates general descriptions of information presentation concepts such as coordinate systems, trnasformations, display regions, display objects, and the relationships between data objects and their depictions. This paper summarizes the current state of that system and the Information Presentation System concept.