Intention Recognition

Personalization and recommendation systems have become a cornerstone of modern digital experiences, providing tailored content to users and enhancing engagement across various industries. The integration of artificial intelligence (AI) and machine learning (ML) in recommendation systems has revolutionized how businesses interact with consumers by analyzing vast amounts of user data to generate highly relevant content suggestions. This paper explores the critical components of recommendation systems, focusing on their architecture, algorithmic implementation, management strategies, and the challenges they encounter. By addressing fundamental issues such as data sparsity, scalability, privacy, and algorithmic bias, organizations can develop more efficient and ethical AI-driven recommendation frameworks that improve user experiences while maintaining trust and transparency.

Stable Model Semantics and Well Founded Semantics have been shown to be very useful in several applications of non-monotonic reasoning. However, Stable Models presents a high computational complexity, whereas Well Founded Semantics is easy to compute and provides an approximation of Stable Models. Efficient engines exist for both semantics of logic programs. This work presents a computational integration of two of such systems, namely XSB and SMODELS. The resulting system is called XNMR, and provides an interactive system for the exploration of both semantics. Aspects such as modularity can be exploited in order to ease debugging of large knowledge bases with the usual Prolog debugging techniques and an interactive environment. Besides, the use of a full Prolog system as a front-end to a Stable Models engine augments the language usually accepted by such systems.

Following a deep digital transformation in which the banking industry has engaged itself and this digitization has led to the development of customer data as a key strategic asset in the industry. Nevertheless, the challenge of data fragmentation is ongoing caused by legacy systems, decentralized data, and even the growing variety of digital interactions. This article discusses the case study of strategic solution of Unified Customer Data Platforms (CDPs) that can help to de-platform data silos in the banking institutions. This paper is the synthesis of existing industry practice and academic literature, examining how unified customer profiles can become a very powerful element in credit decisioning and risk management. The paper covers integration strategies, organizational implications, and emerging technologies that will define the future of CDPs through the case-based analysis and theoretical application. It leaves the research with policy recommendations and a prospective outlook on the emerging future of customer data in financial services. I.

This paper shows how moral decisions can be drawn computationally by using prospective logic programs. These are employed to model moral dilemmas, as they are able to prospectively look ahead at the consequences of hypothetical moral judgments. With this knowledge of consequences, moral rules are then used to decide the appropriate moral judgments. The whole moral reasoning is achieved via a priori constraints and a posteriori preferences on abductive stable models, two features available in prospective logic programming. In this work we model various moral dilemmas taken from the classic trolley problem and employ the principle of double effect as the moral rule. Our experiments show that preferred moral decisions, i.e. those following the principle of double effect, are successfully delivered. Additionally we consider another moral principle, the principle of triple effect, in our implementation. We show that our prospective logic programs allow us to explain computationally different moral judgments that are drawn from these two slightly but distinctively different moral principles.

Personal debt in the United States has reached critical levels, creating widespread economic strain and limiting opportunities for financial mobility. This article presents a comprehensive AI-driven ecosystem designed to proactively identify financially distressed individuals and connect them with personalized debt relief resources through advanced machine learning and realtime data engineering. The framework integrates multiple AI models, including risk classification algorithms, propensity scoring systems, natural language processing for intent detection, and recommender systems for tailored program matching. Built on a scalable infrastructure utilizing Apache Kafka and Spark for stream processing, the system aggregates behavioral signals from diverse sources while maintaining privacy compliance. The multichannel engagement strategy encompasses on-site personalization, targeted digital remarketing, connected television campaigns, and direct communication channels to ensure inclusive reach across demographics. Through a structured five-phase journey from crisis identification to financial empowerment, the framework demonstrates significant improvements in program participation rates, debt reduction outcomes, credit score rehabilitation, and reduction in financial anxiety. The system's architecture enables nationwide deployment across varied populations and regions, offering a transformative solution to address economic inequality and promote sustainable financial recovery. This technological innovation represents a convergence of artificial intelligence, behavioral science, and social impact, providing a blueprint for large-scale financial wellness initiatives that serve the public good while advancing the field of applied AI in economic contexts.

Axelrod's work on the prisoner's dilemma is one of the most discussed models of social cooperation. While many aspects of his computer simulations have been debated, their evolutionary mechanism has not yet received the same attention. We know people do not differ only in the way they act, but also in how they change their behavior -some may like safe routines, others risk with the new. Yet in formal models cultural evolution is taken to be an homogeneous process, such as the imitation of successful peers. In this paper we challenge this view and we propose an agent-based model that takes into account heterogeneity among individuals' learning strategies. The evolutionary dynamic is an adaptation of the so-called consumat approach, originally developed by Wander Jager and Marco Janssen in order to integrate different models of individuals behavior.

Axelrod's work on the prisoner's dilemma is one of the most discussed models of social cooperation. While many aspects of his computer simulations have been debated, their evolutionary mechanism has not yet received the same attention. We know people do not differ only in the way they act, but also in how they change their behavior -some may like safe routines, others risk with the new. Yet in formal models cultural evolution is taken to be an homogeneous process, such as the imitation of successful peers. In this paper we challenge this view and we propose an agent-based model that takes into account heterogeneity among individuals' learning strategies. The evolutionary dynamic is an adaptation of the so-called consumat approach, originally developed by Wander Jager and Marco Janssen in order to integrate different models of individuals behavior.

The evolution of cooperation is a major question in the biological and behavioral sciences. While most theoretical studies model cooperation in the context of an isolated interaction (e.g., a Prisoner’s Dilemma), humans live in heterogeneous social environments, characterized by large variations in fitness interdependence—the extent to which one’s fitness is affected by others. Theoretical and experimental work indicates that humans can infer, and respond to, variations in interdependence. In a heterogeneous ancestral environment, these psychological mechanisms to infer fitness interdependence could have provided a selective advantage, allowing individuals to maximize their fitness by deciding when and with whom to cooperate. Yet, to date, the link between cognitive inference, variation in fitness interdependence, and cooperation remains unclear. Here we introduce a theoretical framework to study the evolution of inference and cooperation in heterogeneous social environments, where individuals experience interactions with varying levels of corresponding interests. Using a combination of evolutionary game theory and agent-based modeling, we model the evolution of adaptive agents, who incur a cost to infer interdependence, in populations of fixed-behavior agents who always cooperate or defect. Our results indicate that natural selection could promote the evolution of psychological mechanisms to infer fitness interdependence, provided that there is enough variation in fitness interdependence to offset the cost of inference. Under certain conditions, the fixation of adaptive agents results in higher levels of cooperation. This depends crucially on the type of inference performed and the features of the interdependence landscape.

Modern consumers engage across multiple digital and physical channels, creating fragmented data trails that challenge marketers in delivering cohesive and effective campaigns. Artificial Intelligence (AI) provides a unifying force for omnichannel marketing by enabling the integration of diverse data sources and accurate identification of high-value customer intent. At a macro level, AI facilitates data fusion from email interactions, website behavior, mobile app usage, in-store visits, social media activity, and third-party datasets to build a unified customer profile. This integration supports real-time intent recognition through techniques such as behavioral pattern mining, sequence modeling, and attention-based neural networks. By recognizing signals such as purchase hesitancy, repeat engagement, or comparative browsing, AI models classify customer intent and predict potential conversion trajectories. Marketers can then target segments with personalized offers, optimized timing, and content tailored to the user's journey stage. AI also enhances attribution modeling and campaign effectiveness by continuously learning from multi-touchpoint engagements. Furthermore, these systems assist in prioritizing high-value customers-those likely to exhibit loyalty or higher lifetime value-enabling resource-efficient allocation of marketing budgets. Challenges include managing data privacy, ensuring consistent messaging across platforms, and maintaining transparency in automated decision-making. Nevertheless, the strategic integration of AI into omnichannel marketing ecosystems empowers organizations to deliver personalized, predictive, and performance-driven campaigns across the full customer journey.

o direito, perpétuo e sem limites geográficos, de arquivar e publicar esta dissertação através de exemplares impressos reproduzidos em papel ou de forma digital, ou por qualquer outro meio conhecido ou que venha a ser inventado, e de a divulgar através de repositórios científicos e de admitir a sua cópia e distribuição com objectivos educacionais ou de investigação, não comerciais, desde que seja dado crédito ao autor e editor. First of all, I would like to express my deepest gratitude and appreciation to my supervisor, Prof. Luís Moniz Pereira, who has worked closely with me all these years. For he is knowledgeable in so many research areas, I cannot recall how much I have learnt from him through an enormous number of discussions (not only in this PhD, but also during my Master study). His genuine enthusiasm for new research ideas and his endless desire to sort them out show me an example of an extraordinarily dedicated scientist, which, I believe, will strongly influence my research approach and attitudes for the rest of my career. I truly enjoy and appreciate the many long discussions that he always generously arranges whenever I wish. Furthermore, I am profoundly indebted to his constant (and always speedily) support in all matters, not only in my studies, but also in the everyday life ones. Besides my supervisor, I am most grateful to Francisco Santos, who has been, not only providing me with the significant guidance for the evolutionary part of my PhD research, but also exhibiting essential support and warm friendship. I truly appreciate all the precious scientific connections Francisco has helped to establish, which are undoubtedly crucial for my career, and our several interesting discussions and collaborations during these years. I am very grateful to the members of the Thesis Plan Advisory Committee, Carlos Damásio, Fariba Sadri, and Francisco Santos, who have provided me with essential suggestions to converge into an appropriate plan for my PhD research. I also thank them, plus Tom Lenaerts, for useful suggestions as members of my thesis program examination committee. I thank the course lecturers during my PhD studies, Nuno Marques, Carlos Damásio, Luís Caires, José Alferes, João Leite, for providing me with essential scientific background. I also thank Luís Caires, Nuno Correia, Pedro Barahona, and José Alferes, who were always of great help with respect to the administration issues. I acknowledge the Portuguese national foundation of science and technologies (FCT-MCTES) for giving me the PhD grant (ref. SFRH/BD/62373/2009), which enables me to focus on doing research. I acknowledge the

Abstract. The present paper shows how the mechanism of tabling can be brought to fruition in abductive logic programming, but requiring an innovative usage of said tabling mechanisms. Abductive solutions tabled in one context can thereby be reused in a different abductive context consistent with them, and need not be recomputed. Given an abductive logic program P, a query to P is evaluated with respect to a program transformation of P, which does not require a metainterpreter. The transformation introduces tabled predicates ...

This study presents a novel approach to enhancing chatbot intent classification through an optimized data preparation combined with Active Learning. We applied the clustering mechanism using a state-of-the-art sentence-transformers model with cosine similarity for cluster detection in order to categorize messages. This process was further refined through a dedicated Active Learning Pipeline, which focused on the most essential observations for labeling. Incorporating externally sourced labeled data from Scale AI, the labeling process was fine-tuned iteratively, until the model's performance stabilized. This approach shows promise for various datasets and tasks, suggesting a scalable solution for preparing data for supervised modeling and achieving optimal model performance in real-world commercial chatbot scenarios.

Purpose of the study: This study presents an approach for improving the performance of natural language processing (NLP) models in pseudo-labeling tasks, with a particular focus on enhancing chatbot model intent recognition for business use cases.
Methodology: The employed case study approach explores the pseudo-labeling technique and demonstrates a practical and efficient way to iteratively expand the original set of labeled data for the purpose of refining model training to achieve superior intent recognition accuracy in chatbots.
Main Findings: The approach results in notable increases in macro-averaged F1 score and overall accuracy, particularly by iteratively retraining the model with progressively larger datasets. While enhancing the model's ability to generalize through difficult cases was effective, the study found that incorporating a full range of examples, including easy ones, yielded the best results. This comprehensive approach made the model better suited for real-world applications.
Applications of the study: As chatbots are increasingly deployed in various sectors, including business, customer service, healthcare, and education, it becomes crucial for research to examine their long-term impact, scalability, and adaptability to ensure their effectiveness and sustainability in diverse contexts. Therefore, building more accurate chatbots, capable of understanding a wide range of user intents, is particularly valuable in real-world applications where chatbots need to respond to diverse, often complex and unpredictable user queries.
Novelty/Originality of the study: Unlike traditional approaches, this study introduces a novel strategy of filling low-density regions in the dataset with pseudolabels, allowing the model to better separate classes and handle semantically similar but varied messages. These advancements contribute to a more effective and scalable chatbot solution across diverse industries.

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

L'observation d'une activité pour comprendre des comportements particuliers nécessite de discerner ce qui relève ou non du contexte. En intelligence artificielle la notion de contexte est une approche modale du raisonnement et plusieurs études ont été faites pour proposer des modèles génériques de représentation des connaissances contextuelles. La question de la découverte interactive des connaissances contextuelles et de leur inscription dans un modèle générique n'est toutefois que rarement examinée dans la littérature. Cette communication propose une approche exploitant le potentiel de la représentation de l'activité par ses traces modélisées d'interaction. Elle consiste à assister la découverte du contexte explicatif par une approche interactive associant l'analyse des données, les connaissances de l'analyste et leur mise en situation dans les traces modélisés. Cette approche interactive facilite l'élicitation de ce qui est contexte ou non. De plus, nous montrons qu'il est alors possible de représenter le contexte découvert sous une forme générique telle que proposée dans la littérature spécialisée. La méthode a été implémentée et a permis la découverte de contextes explicatifs pour la consommation en carburant lors de l'activité de conduite automobile. Les expérimentations, les données et les analyses ont été menées à l'IFSTTAR dans des conditions de conduite réelles.

This paper presents a method for realising abduction in artificial neural networks (ANNs) by generalising existing neurosymbolic approaches from normal logic programs to abductive logic programs (ALPs) in order to provide a more expressive formalism for representing and reasoning about partial knowledge and integrity constraints. The aim is to develop a massively-parallel methodology for abduction that can also be integrated with connectionist learning approaches to offer a finer degree of control over which assumptions can and cannot be made in learning. Existing methods for abduction in neural networks are not well suited to this task as they only apply to a restricted a class of abduction problems or they do not adequately address the issues of local minima and multiple solutions. This paper proposes an algorithm for translating ALPs into ANNs whereby no restrictions are imposed on the underlying programs and, if required, the network can systematically compute all abductive explanations or provide a guarantee when none exist. Moreover, since the topology of the network mirrors the structure of the program, it can be acquired and revised by standard neuro-symbolic training techniques and can also be exploited to impose a preference on the order in which the solutions are found.

Humans cooperate in groups in which mutual monitoring is common, and this provides the possibility of third-party arbitration. Third-party arbitration stabilizes reciprocity in at least two ways: first, when it is accurate, it reduces the frequency of misunderstandings resulting from perception errors, and second, even when it is inaccurate, it provides a public signal that allows pairs to align their expectations about how to behave after errors occur. Here, we describe experiments that test for these two effects. We find that in an iterated, sequential Prisoner's Dilemma game with errors, players with the highest average payoffs are those who make use of third-party arbitration and who also employ forgiving strategies. The combination of these two behaviours reduces the detrimental effects of errors on reciprocity, resulting in more cooperation.

Given its ubiquity, scale and complexity, few problems have created the combined interest of so many unrelated areas as the evolution of cooperation. Using the tools of evolutionary game theory, here we address, for the first time, the role played by intention recognition in the final outcome of cooperation in large populations of self-regarding individuals. By equipping individuals with the capacity of assessing intentions of others in the course of repeated Prisoner's Dilemma interactions, we show how intention recognition opens a window of opportunity for cooperation to thrive, as it precludes the invasion of pure cooperators by random drift while remaining robust against defective strategies. Intention recognizers are able to assign an intention to the action of their opponents based on an acquired corpus of possible intentions. We show how intention recognizers can prevail against most famous strategies of repeated dilemmas of cooperation, even in the presence of errors. Our approach invites the adoption of other classification and pattern recognition mechanisms common among Humans, to unveil the evolution of complex cognitive processes in the context of social dilemmas.

This paper presents a computational model, via Logic Programming (LP), of counterfactual reasoning with applications to agent morality. Counterfactuals are conjectures about what would have happened, had an alternative event occurred. In the first part, we show how counterfactual reasoning, inspired by Pearl’s structural causal model of counterfactuals, is modeled using LP, by benefiting from LP abduction and updating. In the second part, counterfactuals are applied to agent morality, resorting to this LP-based approach. We demonstrate its potential for specifying and querying moral issues, by examining viewpoints on moral permissibility via classic moral principles and examples taken from the literature. Finally, we discuss some potential extensions of our LP approach to cover other aspects of counterfactual reasoning and show how these aspects are relevant in modeling agent morality.

Abstract: One way to address preferences in Multi-Agent negotiation scenarios, is to devise a strategy for removal of contradictions, which may arise when the preferences of all agents are put together. This paper contributes towards that goal by proposing a methodology, based on declarative diagnosis, that minimizes the losses of all agents (ie the number of preferences they have to yield in order to find a consensus) as negotiations are iterated; in other words, it will bias negotiations towards cooperation. Instead of considering explicit ...

Les erreurs humaines sont responsables de 70% des incidents et accidents dans les systèmes de transports. Ce papier propose de prendre en compte la stabilité humaine dont l'objectif principal est d'accroître le contrôle des risques. Un rappel des définitions des différents termes (robustesse, résilience...) est proposé pour positionner la stabilité humaine par rapport à ces termes. Ensuite, une définition de la stabilité humaine est proposée sous forme d'états de stabilité et de transition. Un premier formalisme est proposé et appliqué à une système de transport : le contrôle de trafic aérien.

Les erreurs humaines sont responsables de 70% des incidents et accidents dans les systèmes de transports. Ce papier propose de prendre en compte la stabilité humaine dont l'objectif principal est d'accroître le contrôle des risques. Un rappel des définitions des différents termes (robustesse, résilience...) est proposé pour positionner la stabilité humaine par rapport à ces termes. Ensuite, une définition de la stabilité humaine est proposée sous forme d'états de stabilité et de transition. Un premier formalisme est proposé et appliqué à une système de transport : le contrôle de trafic aérien.

Partant de l'idée que les activités sociales étaient mues par un système de dynamiques, nous avons observé que les propriétés de ce système déterminaient la nature des cadres organisationnels dans lesquels une activité allait se dérouler. L'étude des activités de production des biens immatériels, nous a également permis de montrer que ces propriétés découlaient des interactions entre les acteurs qui cherchent à influer sur le cours de ces dynamiques.

Abstract—In abductive logic programming, abductive solutions are typically computed without attending to the abductive context. These abductive solutions can actually be reused in a different abductive context. In this paper we employ a tabling mechanism and propose a tabled abduction mechanism, that consists of a transformation from abductive normal logic programs into tabled dual programs, by tabling abductive solution entries and without requiring any meta-interpreter. Recomputation of abductive solutions for a different context ...

This work proposes the application of preferences over abductive logic programs as an appealing declarative formalism to model choice situations. In particular, both a priori and a posteriori handling of preferences between abductive extensions of a theory are addressed as complementary and essential mechanisms in a broader framework for abductive reasoning. Furthermore, both of these choice mechanisms are combined with other formalisms for decision making, like economic decision theory, resulting in theories ...

L'analyse comportementale des vehicules auto-nomes represente un challenge majeur dans le monde de l'automobile. Dans le but d'assurer une conduite securisee et fluide, plusieurs me-thodes peuvent etre appliquees, notamment la simulation informatique. L'analyse, cependant, doit tenir compte de tres nombreux parametres evoluant en temps reel. Dans ce contexte, nous etudions le changement de voie d'un vehicule a differents niveaux d'abstraction. Cela nous conduit a determiner les proprietes qui peuvent etre validees avec un niveau d'abstraction plus haut, ou l'espace des parametres est reduit, per-mettant une analyse plus efficace. Abstract The behavioral analysis of autonomous vehicles represents a major challenge in the automotive world. In order to ensure safe and fluid driving, several methods can be applied, including computer simulation. The analysis, however, must take into account a large number of parameters evolving in real time. In this conte...

Individuals make commitments towards others in order to influence others to behave in certain ways. Most commitments may depend on some incentive that is required to ensure that the action is in the agent's best interest and thus, should be carried out to avoid eventual penalties. Similarly, individuals may ground their decision on an accurate assessment of the intentions of others. Hence, both commitments and intention recognition go side by side in behavioral evolution. Here, we analyze the role played by the co-evolution of intention recognition plus the emergence of commitments, in the framework of the evolution of cooperative behavior. We resort to tools of evolutionary game theory in finite populations, showing how the combination of these two aspects of human behavior can enhance the emergent fraction of cooperative acts under a broad spectrum of configurations.

Deliberation plays an important role in the design of rational agents embedded in the real-world. In particular, deliberation leads to the formation of intentions, i.e., plans of action that the agent is committed to achiev ing. In this paper, we present a branching time possible-worlds model for representing and reasoning about, beliefs, goals, inten tions, time, actions, probabilities, and pay offs. We compare this possible-worlds ap proach with the more traditional decision tree representation and provide a transfor mation from decision trees to possible worlds. Finally, we illustrate how an agent can per form deliberation using a decision-tree rep resentation and then use a possible-worlds model to form and reason about his inten tions.

We present a method for incremental intention recognition by means of incrementally constructing a Bayesian Network (BN) model as more actions are observed. It is achieved based on a knowledge base of easily maintained and constructed fragments of BNs, connecting intentions to actions. The simple structure of the fragments enables to easily and efficiently acquire the knowledge base, either from domain experts or automatically from a plan corpus. We show experimental results improvement for the Linux Plan Corpus. In addition, we create a new, so-called IPD Plan Corpus, for strategies in the iterated Prisoner’s Dilemma and show the experimental results for it.

Evolution of cooperation among self-interested agents is revisited in this paper in the context of globalization and localization. A globalized society is characterized by disentrenchment-or routine interactions between strangers across subcultures. Such interactions are rich in novelty, but also have high levels of distrust and insecurity. A localized society is comprised of clusters of subcultures where most social interactions happen. Each tightly knit subculture is rich in mutual familiarity and trust, but not conducive to the spread of novel ideas. A second dimension is that of utilitarian knowledge. Historically, social acquaintances were the primary (if not the only) source of utilitarian knowledge. With technologies like the internet, diffusion of utilitarian knowledge in a society is no longer modulated by acquaintance networks. This leads us to two different forms of (dis)entrenchment: (dis)entrenchment of knowledge and (dis)entrenchment of acquaintance, leading to four societal configurations. This paper asks how each of the configurations fares with respect to the evolution of cooperation. Entrenchment is represented using well-known network models from the literature, and evolution of cooperation is modeled by the evolutionary version of the iterated prisoners' dilemma game. Based on simulation runs, we note that acquaintance and knowledge are characteristically different aspects. We find that disentrenched knowledge is more conducive for evolution of cooperation in networks rather than disentrenched acquaintances.

This paper proposes a behavioral strategy called expectation of cooperation strategy with which cooperation in the prisoner's dilemma game emerges in agent networks by incorporating Q-learning. The proposed strategy is simple and easy to implement but nevertheless can evolve and maintain cooperation in all agent networks under certain conditions. We conducted a number of experiments to clarify these conditions, and the results indicate that cooperation emerged.

This work proposes the application of preferences over abductive logic programs as an appealing declarative formalism to model choice situations. In particular, both a priori and a posteriori handling of preferences between abductive extensions of a theory are addressed as complementary and essential mechanisms in a broader framework for abductive reasoning. Furthermore, both of these choice mechanisms are combined with other formalisms for decision making, like economic decision theory, resulting in theories ...

Abstract. In previous work, we have proposed a multi-level agent model with at least a meta-level aimed at meta-reasoning and meta-control. In agents, these aspects are strongly related with time and therefore we retain that they can be expressed by means of temporal-logic-like rules. In this paper, we propose an interval temporal logic inspired by METATEM, that allows properties to be verified in specific time interval situated either in the past or in the future. We adopt this logic for definition and run-time verification of properties which ...

This special issue of the IfColog Journal of Logics and their Applications "Frontiers of Abduction" is based on a selection of papers concerning abduction that are situated at the crossroad of logic, epistemology, and cognitive science. The status of abduction-that Hintikka considered the fundamental problem of contemporary epistemology [7]-is very controversial. When dealing with abductive reasoning, misinterpretations and equivocations are common. What are the differences between abduction and induction? What are the differences between abduction and the well-known hypothetico-deductive method? What did Peirce mean when he considered abduction a kind of inference? Does abduction involve only the generation of hypotheses or their evaluation too? Are the criteria for the best explanation in abductive reasoning epistemic, or pragmatic, or both? How many kinds of abduction are there? The papers presented in this issue aim at increasing knowledge-logical, cognitive, epistemological-about creative and expert inferences. Analyzing various reasoning and modeling practices provides a way of specifying the nature of some abductive reasoning processes. Following [12; 14] we have to distinguish between sentential, model-based, and manipulative abduction. In the past decades many attempts have been made to model abduction by developing some formal tools in order to illustrate its computational properties and the relationships with the different forms of deductive reasoning (see, for example, [4]). Some of these formal models of abductive reasoning are based on the theory of the epistemic state of an agent [3], where the epistemic state of an individual is modeled as a consistent set of beliefs that can change by expansion and contraction (belief revision framework). These formal tools, often related to the tradition of nonmonotonic logic, logic programming, semantic tableaux, dynamic and adaptive logic, etc., are usually devoted to illustrating sentential abduction and present some limitations [13]. It may be said that the traditional logical accounts of abduction certainly illustrate much of what is important in abductive reasoning, especially the objective of selecting a set of hypotheses (diagnoses, causes) that are able to dispense good (preferred) explanations of data (observations), but fail in accounting for many cases of explanations occurring in science or in everyday reasoning. The logical papers

The mechanisms of emergence and evolution of cooperation in populations of abstract individuals with diverse behavioural strategies in co-presence have been undergoing mathematical study via Evolutionary Game Theory, inspired in part on Evolutionary Psychology. Their systematic study resorts as well to implementation and simulation techniques, thus enabling the study of aforesaid mechanisms under a variety of conditions, parameters, and alternative virtual games. The theoretical and experimental results have continually been surprising, rewarding, and promising. Recently, in our own work we have initiated the introduction, in such groups of individuals, of cognitive abilities inspired on techniques and theories of Artificial Intelligence, namely those pertaining to both Intention Recognition and to Commitment (separately and jointly), encompassing errors in decision-making and communication noise. As a result, both the emergence and stability of cooperation become reinforced comparatively to the absence of such cognitive abilities. This holds separately for Intention Recognition and for Commitment, and even more when they are engaged jointly. The present paper aims to sensitize the reader to these Evolutionary Game Theory based studies and issues, which are accruing in importance for the modelling of

The aim of this article is to take into account the explanations given by people involved in road accident (drivers, passengers and witnesses) so as to consider preconisation susceptible to improve road safety. Testimonies from 205 reports of the French "Gendarmerie Nationale" were analysed. The results show the existence of actor (driver)/observer (passenger and witness) asymmetry in attribution. In fact, observers give as many internal explanations as external explanations, when actors give more external explanations than internal explanations. This result can be interpreted in term of self-serving bias (in so far as we note an effect of the severity of the outcomes on the drivers' explanations) and also can be interpreted according to the theory of defensive attribution (even if we fail to show that observers' explanations can be influenced by accident severity). Actors and observers explain events with different categories (actor/observer asymmetry bias), but u...

Most real-world applications inevitably face the issue of persistence, generally understood as how to design, maintain and interact with a database. The standard approach relies on the mature technology of relational databases, with interaction specified through SQL embedded in the host programming language. Attempts to raise the level of the error-prone interaction code have been in the direction of object-oriented databases or deductive databases, with simpler queries but less mature technology on the database part. Here we present a different high-level approach, close in spirit to natural language, using variable-free conceptual expressions that are quite concise, natural and easy to read and understand, promoting much better code maintenance than the alternative approaches. This is achieved by using the flexible operator syntax and the deductive capabilities of logic programming in two ways, first to compile the database scheme from a modular structural description into a clausal representation, and then to translate (using the compiled scheme) terms expressing queries and commands into SQL statements. The approach relies crucially on the use of attributes, whose inheritance and composition avoid many explicit joins. Expressions become natural by choosing the right noun phrases (rather than verbal) for the attributes. A useful feature is the use of global parameters for implicit current values. Our deductive handle on the scheme allows the query translation to automatically split the needed joins into inner and outer joins. The abstraction power is further raised by having manifold attributes, whose values actually vary along a parametric domain, the main examples being the handling of temporal and multilingual data. Commands can also be very high-level, for example the simple statement of update of an identity value (part of the primary key) of an individual results in its replacement in all tuples of all concepts where the individual was referenced, this being done in the correct order to prevent violation of foreign keys. We present the ideas incrementally, with examples along with a rigorous account of the used syntax and semantics.

Most real-world applications inevitably face the issue of persistence, generally understood as how to design, maintain and interact with a database. The standard approach relies on the mature technology of relational databases, with interaction specified through SQL embedded in the host programming language. Attempts to raise the level of the error-prone interaction code have been in the direction of object-oriented databases or deductive databases, with simpler queries but less mature technology on the database part. Here we present a different high-level approach, close in spirit to natural language, using variable-free conceptual expressions that are quite concise, natural and easy to read and understand, promoting much better code maintenance than the alternative approaches. This is achieved by using the flexible operator syntax and the deductive capabilities of logic programming in two ways, first to compile the database scheme from a modular structural description into a clausal representation, and then to translate (using the compiled scheme) terms expressing queries and commands into SQL statements. The approach relies crucially on the use of attributes, whose inheritance and composition avoid many explicit joins. Expressions become natural by choosing the right noun phrases (rather than verbal) for the attributes. A useful feature is the use of global parameters for implicit current values. Our deductive handle on the scheme allows the query translation to automatically split the needed joins into inner and outer joins. The abstraction power is further raised by having manifold attributes, whose values actually vary along a parametric domain, the main examples being the handling of temporal and multilingual data. Commands can also be very high-level, for example the simple statement of update of an identity value (part of the primary key) of an individual results in its replacement in all tuples of all concepts where the individual was referenced, this being done in the correct order to prevent violation of foreign keys. We present the ideas incrementally, with examples along with a rigorous account of the used syntax and semantics.

As we face the real possibility of modelling programs that are capable of nondeterministic self-evolution, we are confronted with the problem of having several different possible futures for a single such program. It is desirable that such a system be somehow able to look ahead, prospectively, into such possible futures, in order to determine the best courses of evolution from its own present, and then to prefer amongst them. This is the objective of the ACORDA, a prospective logic programming system. We start from a ...

The aim of AI planning is to solve the problems with no exact solution available. These problems usually have a big search space, and planning may not find plans with the least actions and in the shortest time. Recent researches show that using suitable heuristics can help to find desired plans. In planning problems specified formally through graph transformation system (GTS), there are dependencies between applied rules (actions) in the search space. This fact motivates us to solve the planning problem for a small goal (instead of the main goal), extract dependencies from the searched space, and use these dependencies to solve the planning problem for the main goal. In GTS based systems, the nodes of a state (really is a graph) can be grouped due to their type. To create a small (refined) goal, we use a refinement technique to remove the predefined percent of nodes from each group of the main goal. Bayesian Optimization Algorithm (BOA) is then used to solve the planning problem for the refined goal. BOA is an Estimation of Distribution Algorithm (EDA) in which Bayesian networks are used to evolve the solution populations. Actually, a Bayesian network is learned from the current population, and then this network is employed to generate the next population. Since the last Bayesian network learned in BOA has the knowledge about dependencies between applied rules, this network can be used to solve the planning problem for the main goal. Experimental results on four well-known planning domains confirm that the proposed approach finds plans with the least actions and in the lower time compared with the state-of-the-art approaches.

The aim of AI planning is to solve the problems with no exact solution available. These problems usually have a big search space, and planning may not find plans with the least actions and in the shortest time. Recent researches show that using suitable heuristics can help to find desired plans. In planning problems specified formally through graph transformation system (GTS), there are dependencies between applied rules (actions) in the search space. This fact motivates us to solve the planning problem for a small goal (instead of the main goal), extract dependencies from the searched space, and use these dependencies to solve the planning problem for the main goal. In GTS based systems, the nodes of a state (really is a graph) can be grouped due to their type. To create a small (refined) goal, we use a refinement technique to remove the predefined percent of nodes from each group of the main goal. Bayesian Optimization Algorithm (BOA) is then used to solve the planning problem for the refined goal. BOA is an Estimation of Distribution Algorithm (EDA) in which Bayesian networks are used to evolve the solution populations. Actually, a Bayesian network is learned from the current population, and then this network is employed to generate the next population. Since the last Bayesian network learned in BOA has the knowledge about dependencies between applied rules, this network can be used to solve the planning problem for the main goal. Experimental results on four well-known planning domains confirm that the proposed approach finds plans with the least actions and in the lower time compared with the state-of-the-art approaches.

This thesis wished to investigate the existence of stereotypes pertaining to older drivers (Experiment 1 and 2) and their influence on the simulated driving performance of older adults (Experiment 3). In Experiments 1 and 2, participants viewed 12 clips of driving reactions created using a driving simulator (i.e., four clips of older drivers‘ unsafe behaviours, four clips of younger adults‘ unsafe behaviours, and four clips of appropriate driving behaviours). In Experiment 1, 61 participants rated how representative the behaviours were of a typical younger, middle-aged or older driver. Older drivers‘ unsafe behaviours were rated as more representative of the older driver, while young adults‘ unsafe behaviours were perceived as more representative of the younger driver than the other two age groups. Appropriate driving behaviours were rated as more representative of the middle-aged driver than the young and older driver. Consequently, older drivers‘ stereotypes include specific unsaf...