ABSTRACT. High-utility action rules mining is a data mining task that aims to generate rules that would provide the user with the information on which actions might be the most profitable. The action rule's purpose is to suggest a change in attributes in the rule's antecedent that will consequently change (re-classify) the target attribute. The recommended action (change in attributes) may have some cost, and changing to the desired target may yield some revenue - these values can be used to determine which action rules are profitable.

In this paper, we report on a new extension to the ActionRules package written in Python that enriches action rules with utility. It allows setting a minimum utility threshold to get only the action rules that users are interested in. The extension uses the new utility formula proposed in this paper. The utility of a rule is calculated based on a utility table. The paper can serve as a guide for using the extension we have created or for researchers who would like to extend the package.

ABSTRACT. Knowledge graphs have become an essential technology for powering intelligent applications. Enriching the knowledge within knowledge graphs based on use case-specific requirements can be achieved using inference rules. Applying rules on knowledge graphs requires performant and scalable rule engines. Analyzing rule engines based on test cases covering various characteristics is crucial for identifying the optimal rule engine for a given use case. To this end, we present RUBEN: A Rule Engine Benchmarking Framework providing interfaces to benchmark rule engines based on given test cases. Besides a description of RUBEN's interfaces, we present a selection of test cases adopted from the OpenRuleBench, and an evaluation of four rule engines. In the future, we aim to benchmark existing rule engines regularly and encourage the community to propose new test cases and include other rule engines.

ABSTRACT. Answer Set Programming (ASP) has become a popular paradigm for declarative problem solving and is about to find its way into industry. This is due to its expressive yet easy knowledge representation language powered by highly performant (Boolean) solving technology. As with many other such paradigms before, the transition from academia to industry calls for more versatility. Hence, many real-world applications are not tackled by pure ASP but rather hybrid ASP. The corresponding ASP systems are usually augmented by foreign language constructs from which additional inferences can be drawn. Examples include linear equations or temporal formulas. For the design of "sound" systems, however, it is indispensable to provide semantic underpinnings right from the start. To this end, we will discuss the vital role of ASP's logical foundations, the logic of Here-and-There and its non-monotonic extension, Equilibrium Logic, in designing hybrid ASP systems and highlight some of the resulting industrial applications.

ABSTRACT. Ontologies provide the logical underpinning for the Semantic Web, but their consequences can sometimes be surprising and must be explained to users. A promising kind of explanations are proofs generated via automated reasoning. We report about a series of studies with the purpose of exploring how to explain such formal logical proofs to humans. We compare different representations, such as tree- vs. text-based visualizations, but also vary other parameters such as length, interactivity, and the shape of formulas. We did not find evidence to support our main hypothesis that different user groups can understand different proof representations better. Nevertheless, when participants directly compared proof representations, their subjective rankings showed some tendencies such as that most people prefer short tree-shaped proofs. However, this did not impact the user's understanding of the proofs as measured by an objective performance measure.

ABSTRACT. Ontologies based on Description Logic (DL) represent general background knowledge in a terminology (TBox) and the actual data in an ABox. Both human-made and machine-learned data sets may contain errors, which are usually detected when the DL reasoner returns unintuitive or obviously incorrect answers to queries. To eliminate such errors, classical repair approaches offer maximal subsets of the ABox not having the unwanted answers w.r.t. the TBox as repairs. It is, however, not always clear which of these classical repairs to use as the new, corrected data set. Error-tolerant semantics instead takes all repairs into account: cautious reasoning returns the answers that follow from all classical repairs whereas brave reasoning returns the answers that follow from some classical repair. It is inspired by inconsistency-tolerant reasoning and has been investigated for the DL EL.

In a series of papers, we have developed a repair approach that improves on classical repairs in that it preserves a maximal set of consequences (i.e., answers to queries) rather than a maximal set of ABox assertions. The repairs obtained by this approach are called optimal repairs. In the present paper, we investigate error-tolerant reasoning in the DL EL, but using optimal repairs rather than classical repairs as the underlying set of repairs. To be more precise, we consider a static EL TBox (which is assumed to be correct), represent the data by a quantified ABoxes (where some individuals may be anonymous), and use EL concepts as queries (instance queries). We show that brave entailment of instance queries can be decided in polynomial time. Cautious entailment can be decided by a coNP procedure, but is still in P if the TBox is empty.

ABSTRACT. LegalRuleML is a comprehensive XML-based representation framework for modelling and exchanging normative rules. The TPTP input and output formats, on the other hand, are general-purpose standards for the interaction with automated reasoning systems. In this paper we provide a bridge between the two communities by (i) defining a logic-pluralistic normative reasoning representation format in TPTP, (ii) providing translation schemes between relevant fragments of LegalRuleML and the presented TPTP format, and (iii) proposing a flexible architecture for automated normative reasoning based on this translation. We exemplarily instantiate and demonstrate the approach with three different normative logics.

ABSTRACT. Autonomous vehicles require highly sophisticated decision-making to determine their motion. This paper describes how such functionality can be achieved with a practical rule engine learned from expert driving decisions. We propose an algorithm to create and maintain a rule-based behaviour planner, using a two-layer rule-based theory. The first layer determines a set of feasible parametrized behaviours, given the perceived state of the environment. From these, a resolution function chooses the most conservative high-level maneuver, then the second layer reconciles the parameters into a single behaviour. To demonstrate the practicality of our approach, we report results of its implementation in a level-3 autonomous vehicle and its field test in an urban environment.

ABSTRACT. In this paper, we further advance a line of work aimed to formally model via epistemic logic (aspects of) the group dynamics of cooperative agents. In fact, we have previously proposed and here extend a particular logical framework (the Logic of "Inferable" L-DINF), where a group of cooperative agents can jointly perform actions. I.e., at least one agent of the group can perform the action, either with the approval of the group or on behalf of the group. So far, we have been able to take into consideration actions' costs and the preferences that each agent can have for what concerns performing each action, and agents' roles within a group. In this paper, we introduce the possibility of cooperation among different group: if a group has not an agent with the right role for performing an action, it can ask another group in order to have the action performed.

ABSTRACT. In this short system paper, we present our implementation of a prioritized rule-based language for representing actionable policies, in the context of developing cognitive assistants. The language is associated with a provably efficient deduction process, and owing it to its interpretation under an argumentative semantics it can naturally offer ante-hoc explanations on its drawn inferences. Relatedly, the language is associated with a knowledge acquisition process based on the paradigm of machine coaching, guaranteeing the probable approximate correctness of the acquired knowledge against a target policy. The paper focuses on demonstrating the implemented features of the representation language and its exposed APIs and libraries, and discusses some of its more advanced features that allow the calling of procedural code, and the computation of in-line operations when evaluating rules.