Agent supervision is a form of control/customization where a supervisor restricts the behavior of an agent to enforce certain requirements, while leaving the agent as much autonomy as possible. In this work, we investigate supervision of an agent that may acquire new knowledge about her environment during execution, for example, by sensing. Thus we consider an agent's online executions, where, as she executes the program, at each time point she must make decisions on what to do next based on what her current knowledge is. This is done in a setting based on the situation calculus and a variant of the ConGolog programming language. To reason about such agents, we first define a notion of online situation-determined agent which ensures that for any sequence of actions that the agent can perform online, the resulting agent configuration is unique. We then present our formalization of the online maximally permissive supervisor.

We propose a general framework for inconsistency-tolerant query answering within existential rule setting. This framework unifies the main semantics proposed by the state of art and introduces new ones based on cardinality and majority principles. It relies on two key notions: modifiers and inference strategies. An inconsistency-tolerant semantics is seen as a composite modifier plus an inference strategy. We compare the obtained semantics from a productivity point of view.

Reasoning with the maximally consistent subsets (MCS) of the premises is awell-known approach for handling contradictory information. We introduce two argumentation-based methods for doing so: a declarative approach that is related to Dung-style semantics for abstract argumentation, and a computational approach that is based on extensions of Gentzen-type proofs systems. This brings about a new perspective on reasoning with MCS which shows a strong link between the latter and argumentation systems, and which can be extended to related formalisms. A by-product of this is the introduction of a dynamic proof system for classical logic and rebuttal attacks, which is sound and complete with respect to Dung's stable semantics for the associated argumentation framework.

Learning linguistic and visual concepts from videos and textual the word blue is represented by a subset of the colour feature descriptions without having a predefined set of representations space). We will refer to the words that have visual representations is a challenging yet important task. For example, as concrete linguistic concepts (e.g. the word humans are born without the knowledge of how many representations blue has a representation in the colour space, therefore, blue for directions there are in the world, or how they is a concrete linguistic concept). We will refer to these visual are described in natural language. In some situations, it is representations as visual concepts (e.g. the blue colour better to use the 4 directions representation (front, right, left, in the colour feature space is a visual concept). Finally, we back), in others, one can use the 8 directions (front, front will use the term groundings to refer to the connections between right, right, etc.). Humans are capable of learning these different the different linguistic concepts and visual concepts.

Probabilistic models with weighted formulas, known as Markov models or log-linear models, are used in many domains. Recent models of weighted orderings between elements that have been proposed as flexible tools to express preferences under uncertainty, are also potentially useful in applications like planning, temporal reasoning, and user modeling. Their computational properties are very different from those of conventional Markov models; because of the transitivity of the “less than” relation, standard methods that exploit structure of the models, such as variable elimination, are not directly applicable, as there are no conditional independencies between the orderings within connected components. The best known algorithms for general inference inthese models are exponential in the number of statements. Here, we present the first algorithms that exploit the available structure. We begin with the special case of models in the form of chains; we present an exact O(n^3) algorithm, where n is the total number of elements. Next, we generalize this technique to models in which the set of statements are comprised of arbitrary sets of atomic weighted preference formulas (while the query and evidence are conjunctions of atomic preference formulas), and the resulting exact algorithm runs in time O(m * n^2 * n^c), where m is the number of preference formulas, n is the number of elements, and c is the maximum number of elements in a linear cut (which depends both on the structure of the model and the order in which the elements are processed)—therefore, this algorithm is tractable for cases in which c can be bounded to a low value. Finally, we report on the results of an empirical evaluation of both algorithms, showing how they scale with reasonably-sized models.

Subjective opinions can represent uncertain probabilistic information of any kind, minor or major A Bayesian network (BN) is a compact representation of a imprecision and even total ignorance about the probability joint probability distribution in the form of a directed acyclic distribution, by varying the uncertainty mass between 0 and graph (DAG) with random variables as nodes, and a set 1. By simply substituting every input conditional probability of conditional probability distributions associated with each distribution in a BN with a subjective opinion, we obtain node representing the probabilistic connection of the node what we call a subjective Bayesian network.

Model checking is one of the most powerful and widespread tools for system verification with applications in many areas of computer science and artificial intelligence. The large majority of model checkers deal with properties expressed in point-based temporal logics, such as LTL and CTL. However, there exist relevant properties of systems which are inherently interval-based. Model checking algorithms for interval temporal logics (ITLs) have recently been proposed to check interval properties of computations. As the model checking problem for full Halpern and Shoham's ITL (HS for short) turns out to be decidable, but computationally heavy, research has focused on its well-behaved fragments. In this paper, we provide an almost final picture of the computational complexity of model checking for HS fragments with modalities for (a subset of) Allen's relations meets , met by , starts , and ends .

Most approaches in the field of qualitative spatial reasoning (QSR) use constraint networks to encode spatial scenarios. The size of these networks is quadratic in the number of variables, which has severely limited the real-world application of QSR. In this paper, we propose another representation of spatial scenarios, in which each variable is associated with one or more rectangles. Instead of requiring these rectangles to define a solution of the corresponding constraint network, we construct sequences of rectangles that define partial solutions to progressively weaker constraint networks. We present experimental results that illustrate the effectiveness of this strategy.

This paper tackles the problem of evaluating the degree of inconsistency in spatial and temporal qualitative reasoning. We first introduce postulates to propose a formal framework for measuring inconsistency in this context. Then, we provide two inconsistency measures that can be useful in various AI applications. The first one is based on the number of constraints that we need to relax to get a consistent qualitative constraint network. The second inconsistency measure is based on variable restrictions to restore consistency. It is defined from the minimum number of variables that we need to ignore to recover consistency. We show that our proposed measures satisfy required postulates and other appropriate properties. Finally, we discuss the impact of our inconsistency measures on belief merging in qualitative reasoning.

In this paper, we focus on a recently introduced problem in the context of spatial and temporal qualitative reasoning, called the MAX-QCN problem. This problem involves obtaining a spatial or temporal configuration that maximizes the number of satisfied constraints in a qualitative constraint network (QCN). To efficiently solve the MAX-QCN problem, we introduce and study two families of encodings of the partial maximum satisfiability problem (PMAX-SAT). Each ofthese encodings is based on, what we call, a forbidden covering with regard to the composition table of the considered qualitative calculus. Intuitively, a forbidden covering allows us to express, in a more or less compact manner, the non-feasible configurations for three spatial or temporal entities.The experimentation that we have conducted with qualitative constraint networks from the Interval Algebra shows the interest of our approach.