The paper I wrote for the Knowledge Engineering Review was intended primarily to give a balanced review of different techniques developed in (or imported into) Artificial Intelligence to deal with uncertain knowledge.

The definitions above, separated by ten years, represent two very different conceptions of learning. For Simon learning depends on an internal change in representation, and for Kaelbling it is instead measured in terms of an external change in behaviour. Furthermore, Kaelbling's focus on the situatedness of the learning system being embedded in its environment reflects the recent experience gained by much direct experimentation with physical robots.

Psychological models have been used to simulate emotions within agents as part of the decision-making process. The body of this work has focussed on applying the process of decision making using emotions to social dilemmas, notably the Prisoner’s Dilemma. Previous work has focussed on agents which do not move around, with an initial analysis on how mobility and the environment can affect the decisions chosen. Additionally simulated mood has been introduced to the decision-making process. Exploring simulated emotions and mood to inform the decision-making process in multi-agent systems allows us to explore in further detail how outside influences can have an effect on different strategies. We expand and clarify aspects of how agents are affected by environmental differences. We show how emotional characters settle on an outcome without deviation by providing a formal proof. We validate how the addition of mood increases cooperation, while also showing how small groups achieve this quicker than large groups. Once pure defectors are added, to test the resilience of the cooperation achieved, we see that while agents with a low starting mood achieve a payoff closest to the pure defectors, they are reduced in numbers the most by the pure defectors.

Many real-world domains can benefit from adaptable decentralized task allocation through emergent specialization, especially in large teams of non-communicating agents. We begin with an existing bio-inspired response threshold reinforcement approach for decentralized task allocation and extend it to handle hierarchical task domains. We test the extension on self-deployment of a large team of non-communicating agents to patrolling a hierarchically defined set of areas. Results show near-ideal performance across all areas, while minimizing wasteful task switching through the development of specializations and subsequent respecializations when area demands change. A genetic algorithm is then used to evolve even more adaptable and stable task allocation behavior, by incorporating weight and power coefficients into agents’ response threshold reinforcement action probability calculations.

Solving a legal problem is a complex business. First, the lawyer—or whoever is doing the solving—has to assemble and marshall the facts of his case in as orderly a fashion as his knowledge of the relevant law will allow. Secondly, he must confront the range of pertinent legal sources and decompose these into manageable and applicable units. Thirdly, he will subsume the particular facts of his case under the more general terms of his decomposed units. And, finally, he draws inferences about the applicable legal provisions and their impact on the facts. Of course, assembly, decomposition, subsumption and inference may not occur in linear progression. What is more, problems can cccur at each stage and interim evaluations and judgments must be made.

Finance is a challenging yet appropriate domain for model-based reasoning, an area of research otherwise grounded in classical physics. Among the many features that suggest a model-based approach are that firms have formal internal structures, business entities have idealizable behaviours, and there is a history of formal analysis of business problems. This article discusses the motivations and foundations of the model-based approach, and surveys several existing artificial intelligence programs that exploit its advantages. The survey shows that there are ample opportunities for useful systems and significant research in this area. However, accomplishing either of these goals depends crucially upon moving beyond qualitative models based only on accounting information, which tend not to capture the actual complexities of the domain.

The preceding paper on the refinement of rule-based systems by Precce et al. reflects a recent, rapid growth of interest in verification and validation of knowledge based systems. Much of the literature on this subject is thinly spread over conference proceedings, professional periodicals and academic journals, so we have tried to collect together some of the more frequently cited work to assist readers interested in the area.

Patterns have been extensively used to extract hypernym relations from texts. The most popular patterns are Hearst’s patterns, formulated as regular expressions mainly based on lexical information. Experiences have reported good precision and low recall for such patterns. Thus, several approaches have been developed for improving recall. While these approaches perform better in terms of recall, it remains quite difficult to further increase recall without degrading precision. In this paper, we propose a novel 3-phase approach based on sequential pattern mining to improve pattern-based approaches in terms of both precision and recall by (i) using a rich pattern representation based on grammatical dependencies (ii) discovering new hypernym patterns, and (iii) extending hypernym patterns with anti-hypernym patterns to prune wrong extracted hypernym relations. The results obtained by performing experiments on three corpora confirm that using our approach, we are able to learn sequential patterns and combine them to outperform existing hypernym patterns in terms of precision and recall. The comparison to unsupervised distributional baselines for hypernym detection shows that, as expected, our approach yields much better performance. When compared to supervised distributional baselines for hypernym detection, our approach can be shown to be complementary and much less loosely coupled with training datasets and corpora.

The knowledge engineering practices of one experienced company are outlined, commenting on differences in emphasis in some of the knowledge elicitation and AI research literatures.

Ontologies provide potential support for knowledge and content management on a P2P platform. Although we can design ontologies beforehand for an application, it is argued that in P2P environments static or predefined ontologies cannot satisfy the ever-changing requirements of all users. So we propose every user should make proposals for what kind of ontology is the most apt to his need. Collecting all these proposals (or votes) helps the drift of ontologies. This paper introduces OntoVote, a scalable distributed vote-collecting mechanism based on application-level broadcast trees, and describes how OntoVote can be applied to ontology drift on a P2P platform by discussing several problems involved in the voting process.

In recent years, the capabilities of knowledge-based systems to communicate with their users have evolved from simple interactions to complex dialogues. With this evolution comes a need to understand what makes a good dialogue. In this paper, we are concerned with dialogue coherence. We review the notion of focus, which partly explains this property, and its use for user-system communication. First, we examine the major theories dealing with this notion. We describe what their contribution is and how they differ. Then, we illustrate the benefits of using the notion of focus and especially the improvement in text coherence. We pay particular attention to how the notion can concretely be implemented. Its integration with other techniques and theories is described. We conclude the paper by pointing out remaining issues in the understanding of the notion of focus. The contribution of this paper is to provide a classification of the theories of focus and to show the improvements they offer in elaborate user-system dialogues.

In this paper, we discuss the challenges posed by the NEUROWEB project, as a case study of ontological modeling at a knowledge interface between neurovascular medicine and genomics. The aim of the project is the development of a support system for association studies. We identify the notion of clinical phenotypes, that is, the pathological condition of a patient, as the central construct of the knowledge model. Clinical phenotypes are assessed through the diagnostic activity, performed by clinical experts operating within communities of practice; the different communities operate according to specific procedures, but they also conform to the minimal requirements of international guidelines, displayed by the adoption of a common standard for the patient classification. We develop a central model for the clinical phenotypes, able to reconcile the different methodologies into a common classificatory system. To bridge neurovascular medicine and genomics, we identify the general theory of biological function as the common ground between the two disciplines; therefore, we decompose the clinical phenotypes into elementary phenotypes with a homogeneous physiological background, and we connect them to the biological processes, acting as the elementary units of the genomic world.

Usually, humanoid walking gaits are only roughly distinguished between stable and unstable. The evaluation of a stable humanoid walking gait is difficult to quantify in scales. And, it is extremely hard to adjust humanoid robots in suitable a walking gait for different movement objectives such as fast walking, uneven floor walking, and so on. This paper proposes a stability margin constructed by center of pressure (COP) to evaluate the gait stability of humanoid walking. The stability margin is modeled by the COP regions that a humanoid robot needs for stable standing. We derive the mathematical model for COP position by dividing the walking gait into single and double support phases in order to measure the stability of the COP regions. An actual measuring system for the stable COP regions is designed and implemented. The measured COP trajectory of a walking gait is eventually evaluated with respect to the stable COP regions for the stability margins. The evaluation focuses on weak stability areas to be improved for robust walking gaits. To demonstrate the robustness of the improved walking gait, we replicate the experiment on three different terrains. The experiments demonstrate that the walking gaits developed based on stable COP region can be applied for different movement objectives.

The field of humanoid robotics is constantly evolving, with new advances creating exciting opportunities for research and development. Especially in the entertainment area, robotics applications show significant growth potential. To guide and track the progress of robotics research, good benchmark problems are crucial, but especially investigating human–robot interaction capabilities is difficult. This paper examines robot magic shows as a benchmark and research direction for entertainer robots and discusses flexible and versatile system architectures for robot magicians in the humanoid robot application challenge competition since 2017. The goal is a detailed analysis of magic tricks and magician robots presented in front of audiences. This paper reviews the hardware components and examines the robot platforms of participating teams in the robot magic show, the types of magic performances, the magic tricks and tools, the algorithms for the magic tricks, and the framework for humanoid robots. By providing a comprehensive analysis of these elements, we can gain insight into the capabilities of advanced humanoid robots for high-performance magic tricks. In conclusion, this paper highlights the exciting potential of humanoid robotics and entertainment fusion. In addition, we analyze the use of humanoid robots in magic shows, presenting the industrial potential of entertainer robots.