Timed Argumentation Frameworks (TAFs) allow taking into account the availability of arguments and attacks in abstract argumentation. We propose a new reasoning approach for TAFs, where a standard Dung-style AF can be associated with each timepoint. We show that, although this framework is more expressive than Dung's framework, our approach does not lead to an increase in computational complexity for most reasoning problems and classical extension-based semantics.

Efficient computation of hard reasoning tasks is a key issue in abstract argumentation. One recent approach is to define approximate algorithms, i.e. methods that provide an answer that may not always be correct, but outperform the exact algorithms regarding the computation runtime. One such approach proposes to use the grounded semantics, which is polynomially computable, as a starting point for determining whether arguments are (credulously or skeptically) accepted with respect to various other extension-based semantics. In this paper, we push further this idea by defining a general family of approaches to evaluate the acceptability of arguments which are not in the grounded extension, neither attacked by it. These approaches rely on gradual semantics to evaluate these arguments. We also propose an approach using an heuristic based on the number of arguments attacked by or attacking an argument, and we show that this last approach, although seemingly different, is actually also an instance of our general family of approaches based on gradual semantics. We have implemented our approaches and provided an empirical study in which we discuss the results and compare our approach with the state-of-the-art approximate algorithms.

Automated negotiation aims at finding agreements between agents with conflicting goals. Existing utility-based approaches guarantee agents satisfaction with negotiation outcomes, especially in multi-issue negotiations where concession mechanisms lead to win-win results. However, they lack explainability and do not consider agents’ beliefs. On the other hand, argument-based approaches provide reasons for accepting or rejecting offers but do not include utility modeling for offers or enable concession mechanisms in multi-issue settings. We propose a novel hybrid approach combining both types of approaches. The utility-based component enables agents to make concessions on complex negotiation objects to achieve win-win outcomes, while the argumentation component ensures that accepted offers align with the agents' personal argumentation theories. These theories represent their beliefs, encoding various profiles, ethical considerations, social norms, or legal principles.

Argument Mining (AM) [1], the task of extracting arguments and their relations (e.g. attacks, supports) from text, has lead to potential applications of formal models of argumentation on real-world scenarios. However, AM methods largely depend on Machine Learning (ML) techniques, which require high-quality annotated data. This poses a significant limitation in domains like legal reasoning, where annotated corpora are extremely scarce. Current legal AM research relies almost entirely on a single dataset from the European Court of Human Rights (ECHR) [2], limiting the development of robust, argumentation-driven automated legal reasoning systems. Addressing this data scarcity is essential for advancing computational legal argumentation. To address this challenge, this paper briefly describes some preliminary results that demonstrate the promising potential of LLMs to assist legal experts in annotating additional legal corpora. This advancement paves the way for more robust, argumentation-driven automated reasoning systems in the legal domain. More details on our approach and experimental results are provided in [3].

We have developed a system, based on prioritized default logic, for representing legal rules and their exceptions. Our system can help the law practitioner’s reasoning in presence of scenarios that induce an exception to a rule. When there is no known exception to the rule r1 for a scenario, the system can help the user to adapt the exception r′2 to a general rule r2.

Assumption-Based Argumentation (ABA) is a prominent formalism for structured argumentation, widely applied in domains such as healthcare, law, and robotics. Despite its inherent computational complexity, ABA has seen the development of effective techniques that successfully address key tasks, including evaluating the acceptability of literals and computing framework extensions. These approaches typically involve translating the initial ABA framework into an intermediate formalism, such as an Answer Set Program or an Abstract Argumentation Framework, which is then encoded into a Boolean satisfiability (SAT) problem. However, this translation can lead to large and complex intermediate representations, posing challenges for state-of-the-art SAT solvers. In this work, we propose a Counterexample-Guided Abstraction Refinement (CEGAR) approach that bypasses the initial translation step, at the cost of incrementally discovering certain ABA constraints that are not explicitly captured in the initial SAT encoding. We analyze the performance of our method and demonstrate that it outperforms state-of-the-art approaches on specific problem classes, while remaining competitive with the best existing solvers more broadly.

Various approaches have been proposed for providing efficient computational approaches for abstract argumentation. Among them, neural networks have permitted to solve various decision problems, notably related to arguments (credulous or skeptical) acceptability. In this work, we push further this study in various ways. First, relying on the state-of-the-art approach AFGCN, we show how we can improve the performances of the Graph Convolutional Networks (GCNs) regarding both runtime and accuracy. Then, we show that it is possible to improve even more the efficiency of the approach by modifying the architecture of the network, using Graph Attention Networks (GATs) instead.

In fair division, local envy-freeness is a desirable property which has been thoroughly studied in recent years. In this paper, we study explanations which can be given to explain that no allocation of items can satisfy this criterion, in the house allocation setting where agents receive a single item. While Minimal Unsatisfiable Subsets (MUSes) are key concepts to extract explanations, they cannot be used as such: (i) they highly depend on the initial encoding of the problem; (ii) they are flat structures which fall short of capturing the dynamics of explanations; (iii) they typically come in large number and exhibit great diversity. In this paper we provide two SAT encodings of the problem which allow us to extract MUS when instances are unsatisfiable. We build a dynamic graph structure which allows to follow step-by-step the explanation. Finally, we propose several criteria to select MUSes, some of them being based on the MUS structure, while others rely on this original graphical explanation structure. We give theoretical bounds on these metrics, showing that they can vary significantly for some instances. Experimental results on synthetic data complement these results and illustrate the impact of the encodings and the relevance of our metrics to select among the many MUSes.

Incomplete Argumentation Frameworks (IAFs) enrich classical abstract argumentation with arguments and attacks whose actual existence is questionable. The usual reasoning approaches rely on the notion of completion, i.e. standard AFs representing “possible worlds” compatible with the uncertain information encoded in the IAF. Recently, extension-based semantics for IAFs that do not rely on the notion of completion have been defined, using instead new versions of conflict-freeness and defense that take into account the (certain or uncertain) nature of arguments and attacks. In this paper, we give new insights on both the “completion-based” and the “direct” reasoning approaches. First, we adapt the well-known grounded semantics to this framework in two different versions that do not rely on completions. After determining that our new semantics are polynomially computable, we provide a principle-based analysis of these semantics, as well as the “direct” semantics previously defined in the literature, namely the complete, preferred and stable semantics. Finally, we also provide new results regarding the satisfaction of principles by the classical “completion-based” semantics.

Abstract Incomplete Argumentation Frameworks (IAFs) have been defined to incorporate some qualitative uncertainty in abstract argumentation: information such as ‘I am not sure whether this argument exists’ or ‘I am not sure whether this argument attacks that one’ can be expressed. Reasoning with IAFs is classically based on a set of completions, i.e. standard argumentation frameworks (AFs) that represent the possible worlds encoded in the IAF. The number of these completions may be exponential with respect to the number of arguments in the IAF. This leads, in some cases, to an increase of the complexity of reasoning, compared to the complexity of standard AFs. In this paper, we follow an approach that was initiated for Partial Argumentation Frameworks (PAFs) (a subclass of IAFs), which consists in defining new forms of conflict-freeness and defense, the properties that underly the definition of Dung’s semantics for AFs. We generalize these semantics from PAFs to IAFs. We show that, among three possible types of admissibility, only two of them satisfy some desirable properties. We use them to define two new families of extension-based semantics. We study the properties of these semantics, and in particular, we show that their complexity remains the same as in the case of Dung’s AFs. Finally, we propose a logical encoding of these semantics, and we show experimentally that this encoding can be used efficiently to reason with IAFs, thanks to the power of modern SAT solvers.

Recently, Strength-based Argumentation Frameworks (StrAFs) have been proposed to model situations where some quantitative strength is associated with arguments. In this setting, the notion of accrual corresponds to sets of arguments that collectively attack an argument. Some semantics have already been defined, which are sensitive to the existence of accruals that collectively defeat their target, while their individual elements cannot. However, until now, only the surface of this framework and semantics have been studied. Indeed, the existing literature focuses on the adaptation of the stable semantics to StrAFs. In this paper, we push forward the study and investigate the adaptation of admissibility-based semantics. Especially, we show that the strong admissibility defined in the literature does not satisfy a desirable property, namely Dung’s fundamental lemma. We therefore propose an alternative definition that induces semantics that behave as expected. We then study computational issues for these new semantics, in particular we show that complexity of reasoning is similar to the complexity of the corresponding decision problems for standard argumentation frameworks in almost all cases. We then propose a translation in pseudo-Boolean constraints for computing (strong and weak) extensions. We conclude with an experimental evaluation of our approach which shows in particular that it scales up well for solving the problem of providing one extension as well as enumerating them all.

Preference-based argumentation and ranking semantics are two important research topics in the computational argumentation literature. Surprisingly, no study investigates to what extent preferences over arguments and ranking semantics can interact. This paper fills the gap between the relative priorities that one can express and the evaluation of arguments individual acceptability. More precisely, we propose a natural principle that should be satisfied by a ranking semantics for Preference-based Argumentation Frameworks. We show that although existing semantics do not satisfy this desirable principle, they can be used to define new ranking semantics that exhibit the expected behavior. Finally, we discuss an application of these semantics to the modeling of human reasoning.

Dynamics of argumentation is the family of techniques concerned with the evolution of an argumentation framework (AF), for instance to guarantee that a given set of arguments is accepted. This work proposes Control Argumentation Frameworks (CAFs), a new approach that generalizes existing techniques, namely normal extension enforcement, by accommodating the possibility of uncertainty in dynamic scenarios. A CAF is able to deal with situations where the exact set of arguments is unknown and subject to evolution, and the existence (or direction) of some attacks is also unknown. It can be used by an agent to ensure that a set of arguments is part of one (or every) extension whatever the actual set of arguments and attacks. A QBF encoding of reasoning with CAFs provides a computational mechanism for determining whether and how this goal can be reached. We also provide some results concerning soundness and completeness of the proposed encoding as well as complexity issues.

Properties of argumentation semantics have been widely studied in the last decades. However, there has been no investigation on the question of difference measures between semantics. Such measures turn helpful when the semantics associated to an argumentation framework may have to be changed, in a way that ensures that the new semantics is not too dissimilar from the old one. Three main notions of difference measures between semantics are defined in this paper. Some of these measures are shown to be distances or semi-distances.

In this paper we introduce a new approach for revising and merging consistent Horn formulae under minimal model semantics. Our approach is translation-based in the following sense: we generate a propositional encoding capturing both the syntax of the original Horn formulae (the clauses which appear or not in them) and their semantics (their minimal models). We can then use any classical revision or merging operator to perform belief change on the encoding. The resulting propositional theory is then translated back into a Horn formula. We identify some specific operators which guarantee a particular kind of minimal change. A unique feature of our approach is that it allows us to control whether minimality of change primarily relates to the syntax or to the minimal model semantics of the Horn formula. We give an axiomatic characterization of minimal change on the minimal model for this new setting, and we show that some specific translation-based revision and merging operators satisfy our postulates.

Nowadays, argumentation is a salient keyword in artificial intelligence. The use of argumentation techniques is particularly convenient for thematics such that multiagent systems, where it allows to describe dialog protocols (using persuasion, negotiation, ...) or on-line discussion analysis, it also allows to handle queries where a single agent has to reason with conflicting information (inference in the presence of inconsistency, inconsistency measure). This very rich framework gives numerous reasoning tools, thanks to several acceptability semantics and inference policies. On the other hand, the progress of SAT solvers in the recent years, and more generally the progress on Constraint Programming paradigms, lead to some powerful approaches that permit to tackle theoretically hard problems. The needs of efficient applications to solve the usual reasoning tasks in argumentation, together with the capabilities of modern Constraint Programming solvers, lead us to study the encoding of usual acceptability semantics into logical settings. We propose diverse use of Constraint Programming techniques to develop a software library dedicated to argumentative reasoning. We present a library which offers the advantages to be generic and easily adaptable. We finally describe an experimental study of our approach for a set of semantics and inference tasks, and we describe the behaviour of our solver during the First International Competition on Computational Models of Argumentation.