Graph Structures for knowledge Representation and Reasoning
From Graph Structures for knowledge Representation and Reasoning 2020
The development of effective techniques for knowledge representation and reasoning (KRR) is a crucial aspect of successful intelligent systems. Different representation paradigms, as well as their use in dedicated reasoning systems, have been extensively studied in the past. Nevertheless, new challenges, problems, and issues have emerged in the context of knowledge representation in Artificial Intelligence (AI), involving the logical manipulation of increasingly large information sets (see for example Semantic Web, BioInformatics, and so on). Improvements in storage capacity and performance of computing infrastructure have also affected the nature of KRR systems, shifting their focus towards representational power and execution performance. Therefore, KRR research is faced with the challenge of developing knowledge representation structures optimized for large-scale reasoning. This new generation of KRR systems includes graph-based knowledge representation formalisms such as Constraint Networks (CNs), Bayesian Networks (BNs), Semantic Networks (SNs), Conceptual Graphs (CGs), Formal Concept Analysis (FCA), CP-nets, GAI-nets, and Argumentation Frameworks, all of which have been successfully used in a number of applications. The goal of the workshop series on Graph Structures for Knowledge Representation and Reasoning (GKR) is to bring together researchers involved in the development and application of graph-based knowledge representation formalisms and reasoning techniques. This volume contains extended and revised selected papers of the sixth edition of GKR, under the auspices of ScaDS.AI — Center for Scalable Data Analytics and Artificial Intelligence Dresden/Leipzig, which took place jointly with ECAI 2020, the 24th European Conference on Artificial Intelligence, which was supposed to be held in Santiago de Compostela, Spain. Like ECAI, GKR had to be re-shaped into a virtual edition, given the global pandemic. This was a first, compared to previous editions of GKR held in Pasadena, USA (2009), Barcelona, Spain (2011), Beijing, China (2013), Buenos Aires, Argentina (2015), and Melbourne, Australia (2017). Still, like before, thanks to the association with a major international AI conference, the workshop provided the perfect venue for a rich and valuable exchange. As usual, the workshop submissions underwent single-blind reviewing by the program committee, each receiving between two and three reviews. On top of the extended workshop papers, this volume also contains two invited additional contributions from core GKR community members. The scientific program of this workshop included many topics related to graph-based knowledge representation and reasoning, from sub-disciplines as diverse as conceptual graphs, formal concept analysis, graphical models, graph neural networks, concept diagrams, and others. Application areas included Smart Homes, Education, Team Formation, Enterprise Architectures, and Usage Pattern Analysis, demonstrating the wide applicability of graph-based KR methods. All in all, the sixth edition of the GKR workshop was very successful despite the unusual circumstances. The papers coming from diverse fields all addressed various issues for knowledge representation and reasoning and the common graph-theoretic background helped to bridge the gap between the different communities. This made it possible for the participants to gain new insights and inspiration. The organizers are very grateful for the support of ECAI and we would also like to thank the program committee of the workshop for their hard work in reviewing papers and providing valuable guidance to the contributors. But, of course, GKR 2020 would not have been possible without the dedicated involvement of the contributing authors and participants.
Enterprise Architecture (EA) metamodels align an organisation’s business, information and technology resources so that these assets best meet the organisation’s purpose. The Layered EA Development (LEAD) Ontology enhances EA practices by a metamodel with layered metaobjects as its building blocks interconnected by semantic relations. Each metaobject connects to another metaobject by two semantic relations in opposing directions, thus highlighting how each metaobject views other metaobjects from its perspective. While the resulting two directed graphs reveal all the multiple pathways in the metamodel, more desirable would be to have one directed graph that focusses on the dependencies in the pathways. Towards this aim, using CG-FCA (where CG refers to Conceptual Graph and FCA to Formal Concept Analysis) and a LEAD case study, we determine an algorithm that elicits the active as opposed to the passive semantic relations between the metaobjects resulting in one directed graph metamodel. We also identified the general applicability of our algorithm to any metamodel that consists of triples of objects with active and passive relations.
Artificial Intelligence applications often require to maintain a knowledge base about the observed environment. In particular, when the current knowledge is inconsistent with new information, it has to be updated. Such inconsistency can be due to erroneous assumptions or to changes in the environment. Here we considered the second case, and develop a knowledge update algorithm based on event logic that takes into account constraints according to which the environment can evolve. These constraints take the form of events that modify the environment in a well-defined manner. The belief update triggered by a new observation is thus explained by a sequence of events. We then apply this algorithm to the problem of locating people in a smart home and show that taking into account past information and move’s constraints improves location inference.
The need for software applications that can assist with mental disorders has never been greater. Individuals suffering from mental illnesses often avoid consultation with a psychotherapist, because they do not realize the need, or because they cannot or will not face the social and economic consequences, which can be severe. Between ideal treatment by a human therapist and self-help websites lies the possibility of a helpful interaction with a language-using computer. A model of empathic response planning for sentence generation in a forthcoming automated psychotherapist is described here. The model combines emotional state tracking, contextual information from the patient’s history and continuously updated therapeutic goals to form suitable conceptual graphs that may then be realized as suitable textual sentences.
Contribution of this work is to Define the Creative Composition Problem (CCP) for Human Well-being Optimization by Construction of Knowledge Graph using Knowledge Representation and logic-based Artificial Intelligence reasoning-planning where the computation of the Optimal Solution is achieved by Dynamic Programming or Logic Programming. The Creative Composition Problem is embedded within Cecilia: an architecture of a digital companion artificial intelligence agent system composer of dialogue scripts for Well-being and Mental Health. Where Cecilia Framework is instantiated in Well-being and Mental Health domain for optimal well-being development of first year university students. We define the ‘The Problem of Creating a Dialogue Composition (PCDC)’ and we propose a feasible and optimal solution of it. CCP is instantiated in this applied domain to solve PCDC optimizing the Mental Health and Well-being of the student. CCP as PCDC is applied to optimize maximizing the mental health of the student but also maximizing the smoothness, coherence, enjoyment and engagement each time the dialogue session is composed. Cecilia helps students to manage stress/anxiety to attempt the prevention of depression. Students can interact through the digital companion making questions and answers. While the system “learns” from the user it allows the user to learn from herself. Once the student discovers elements that were unnoticed by her, she will find a better way to improve when discovering her points of improvement.
This paper discusses set visualisations with concept lattices in the sense of Formal Concept Analysis (FCA) in contrast to visualisations with Euler diagrams. Both types of visualisations have advantages and disadvantages. Because of the connection between both fields and the body of knowledge that exists in both fields it is of interest to investigate whether results from either field can contribute to the other.
During the past years, the number of platforms that are introducing a subscription plan is steadily increasing. This phenomenon helps support the developers as well as continuing to provide quality content. Since not so many individuals are willing to spend money or some simply do not have the means, they resort to sharing an account that has a subscription plan. This behavior can, in some instances, be harmful for the developers and, even if it is not, any provider can benefit from knowing what type of clients they have. The solution depicted and explored in this article will focus on using data that is easily available and structuring it in a way that can provide insight into each account activity.
Our interest here lies in supporting important, but routine and time-consuming activities that underpin success in highly distributed, collaborative design and manufacturing environments; and how information structuring can facilitate this. To that end, we present a simple, yet powerful approach to team formation, partner selection, scheduling and communication that employs a different approach to the task of matching candidates to opportunities or partners to requirements (matchmaking): traditionally, this is approached using either an idea of ‘nearness’ or ‘best fit’ (metric-based paradigms); or by finding a subtree within a tree (data structure) (tree traversal). Instead, we prefer concept lattices to establish notions of ‘inclusion’ or ‘membership’: essentially, a topological paradigm. While our approach is substantive, it can be used alongside traditional approaches and in this way one could harness the strengths of multiple paradigms.
Large, heterogeneous datasets are characterized by missing or even erroneous information. This is more evident when they are the product of community effort or automatic fact extraction methods from external sources, such as text. A special case of the aforementioned phenomenon can be seen in knowledge graphs, where this mostly appears in the form of missing or incorrect edges and nodes. Structured querying on such incomplete graphs will result in incomplete sets of answers, even if the correct entities exist in the graph, since one or more edges needed to match the pattern are missing. To overcome this problem, several algorithms for approximate structured query answering have been proposed. Inspired by modern Information Retrieval metrics, these algorithms produce a ranking of all entities in the graph, and their performance is further evaluated based on how high in this ranking the correct answers appear.
In this work we take a critical look at this way of evaluation. We argue that performing a ranking-based evaluation is not sufficient to assess methods for complex query answering. To solve this, we introduce Message Passing Query Boxes (MPQB), which takes binary classification metrics back into use and shows the effect this has on the recently proposed query embedding method MPQE.
A pattern is a generic instance of a binary constraint satisfaction problem (CSP) in which the compatibility of certain pairs of variable-value assignments may be unspecified. The notion of forbidden pattern has led to the discovery of several novel tractable classes for the CSP. However, for this field to come of age it is time for a theoretical study of the algebra of patterns. We present a Galois connection between lattices composed of sets of forbidden patterns and sets of generic instances, and investigate its consequences. We then extend patterns to augmented patterns and exhibit a similar Galois connection. Augmented patterns are a more powerful language than flat (i.e. non-augmented) patterns, as we demonstrate by showing that, for any , instances with tree-width bounded by k cannot be specified by forbidding a finite set of flat patterns but can be specified by a finite set of augmented patterns. A single finite set of augmented patterns can also describe the class of instances such that each instance has a weak near-unanimity polymorphism of arity k (thus covering all tractable language classes).We investigate the power of forbidding augmented patterns and discuss their potential for describing new tractable classes.
コメント