Abstract Modelling the interactions and reasoning processes of multiple agents in a dynamic environment presents a significant challenge, requiring tools that effectively capture diverse interaction types (such as persuasion and deliberation) while supporting agents in decision-making and consensus-building. We extend the Timed Concurrent Language For Argumentation (TCLA) to support the specification of agents equipped with local argument memories and private knowledge reasoning. This extension enables the full formalization of Symmetric Strategic Argumentation Dialogues and Multi-Agent Decision Making with Privacy Preserved problems within TCLA, for which we also introduce general translation functions to automatically obtain TCLA programs. To demonstrate practical applications of TCLA, we provide examples that model the two studied problems and make use of the translation functions.

This paper presents fast and exact methods for computing the probability of an argument’s acceptance using Dung’s semantics in the Constellation paradigm of Abstract Argumentation. For (directed) Singly-Connected Graphs (SCGs), the problem can now be solved in linearithmic time instead of being exponential in the number of attacks, as reported in the literature. Moreover, in the more general case of Directed Acyclic Graphs (DAGs), we provide an algorithm whose time complexity is linearithmic in the product of the out-degree of dependent arguments, i.e., arguments reaching the argument considered for acceptance through multiple paths in the graph. We theoretically show that this complexity is lower than the lower bound of the (exact) Constellation method, which is also supported by empirical results. Our approach to DAGs is also compared with the (approximate) Monte-Carlo method, which is stopped when exact results are obtained. Within this time constraint, Monte-Carlo still outputs significant errors, underlying the fast computation of our approach.

Abstract The study of Dung-style Argumentation Frameworks in recent years has focused on incorporating time. For example, availability intervals have been added to arguments and relations, resulting in different outputs of Dung semantics over time. This paper examines the probability distribution of arguments over time intervals. Using this temporal probabilistic model, the study explores how these frameworks can be transformed into a probabilistic argumentation according to the constellation approach and how they can be interpreted within the epistemic approach. The epistemic approach relies on the notion of defeat to select significant conflicts based on probability distributions. The study also introduces the temporal acceptability of arguments based on the concept of defence, allowing for more precise results over time. Finally, the models (constellation and epistemic) are extended to account for events that have a duration, i.e. that can occur for several consecutive instants of time.

Abstract The timed concurrent language for argumentation (tcla) is a framework to model concurrent interactions between communicating agents that reason and take decisions through argumentation processes, also taking into account the temporal duration of the performed actions. Time is a crucial factor when dealing with dynamic environments in real-world applications, where agents must act in a coordinated fashion to reach their own goals. However, modelling complex interactions and concurrent processes may be challenging without the help of proper languages and tools. In this paper, we discuss the use of tcla for practical purposes and provide a working implementation of the language, endowed with a user interface available online, that serves the dual purpose of aiding the research in this field and facilitating the development of multi-agent systems based applications.

Abstract Computational Argumentation provides tools for both modelling and reasoning with controversial information. Abstract Argumentation Frameworks represent the building blocks in this field and allow one to model the relationships between arguments with the ultimate goal of establishing their acceptability. Arguments can be evaluated through sets of criteria, called semantics, which distinguish among various justification states. For example, an argument may be accepted, rejected, ignored or even marked as undecided. This paper considers Weighted Argumentation Frameworks and proposes a novel labelling semantics that differentiates four states and generalizes existing approaches.

Abstract Time is a crucial factor in modelling dynamic behaviours of intelligent agents: activities have a determined temporal duration in a real-world environment, and previous actions influence agents’ behaviour. In this paper, we propose a language for modelling concurrent interaction between agents that also allows the specification of temporal intervals in which particular actions occur. Such a language exploits a timed version of Abstract Argumentation Frameworks to realise a shared memory used by the agents to communicate and reason on the acceptability of their beliefs with respect to a given time interval. An interleaving model on a single processor is used for basic computation steps, with maximum parallelism for time elapsing. Following this approach, only one of the enabled agents is executed at each moment. To demonstrate the capabilities of the language, we also show how it can be used to model interactions such as debates and dialogue games taking place between intelligent agents. Lastly, we present an implementation of the language that can be accessed via a web interface.

The black box model used in Machine Learning is considered one of the major problems in the application of Artificial Intelligence techniques [1] as it makes machine decisions non-transparent and often incomprehensible even to experts or developers themselves.In this paper, we provide an argumentative interpretation of both the training process and the results predicted.The goal is to build a Bipolar Argumentation Framework (BAF) [2] showing the dialectical reasoning behind the assignment of a certain class to a given record.Since we make assumptions neither on the dataset nor on the algorithm used, the presented procedure can be applied to existing models without the need for further adjustments.To illustrate our proposal, we use the Titanic dataset from www.kaggle.com,which contains records relating to people involved in the Titanic disaster.We consider three categorical features, namely Survived (the class to predict, with value 1 if the person survived or 0, otherwise), Pclass (ticket class among 1, 2 and 3) and sex (0 for woman and 1 for man), and two numerical features: Age (passenger age, ranging from 0.17 to 76) and Fare (passenger fare with values from 0 to 512).In the following, we describe the step our procedure goes through in order to find an explanation for the class Survived=1.Dataset Clustering.In the first step, starting from the input dataset, we create a new clustered dataset in which numerical features are split into categories that group ranges of values to obtain a more appropriate and concise explanation. BAF Generation.Then we build a BAF based on the correlation matrix computed among the features.By construction, the obtained BAF only has symmetric relations. Breaking Complete Symmetry.Given the correlation matrix, we apply a procedure that removes symmetric edges from the BAF to establish a causal relationship between features.In particular, we use the conditional probability [3] computed for arguments which attack/support each other.We choose the minimum values possible that keep the graph connected. Computing Extensions.To identify the set of arguments which are more likely to be accepted, we compute the semi-stable extensions [4] of the previously obtained 1 This work has been partially supported by: GNCS-INdAM, CUP E55F22000270001;

The requirement of explainability is gaining more and more importance in Artificial Intelligence applications based on Machine Learning techniques, especially in those contexts where critical decisions are entrusted to software systems (think, for example, of financial and medical consultancy). In this paper, we propose an Argumentation-based methodology for explaining the results predicted by Machine Learning models. Argumentation provides frameworks that can be used to represent and analyse logical relations between pieces of information, serving as a basis for constructing human tailored rational explanations to a given problem. In particular, we use extension-based semantics to find the rationale behind a class prediction.

Abstract Argumentation Theory provides tools for both modelling and reasoning with controversial information and is a methodology that is often used as a way to give explanations to results provided using machine learning techniques. In this context, labelling-based semantics for Abstract Argumentation Frameworks (AFs) allow for establishing the acceptability of sets of arguments, dividing them into three partitions: in, out and undecidable (instead of classical Dung acceptable and not acceptable sets). This kind of semantics have been studied only for classical AFs, while the more powerful weighted and preference-based frameworks have not been studied yet. In this paper, we define a novel labelling semantics for Weighted Argumentation Frameworks (WAFs), extending and generalizing the crisp one, and we provide some insights towards a definition of strong admissibility for WAFs.

The paper provides an initial study on how ranking semantics in argumentation have to be handled when leaving the purely abstract setting. We employ claim-augmented frameworks where each argument is associated to a claim it stands for. We propose liftings from argument- to claim-level in two veins: for desired properties and for actual rankings. Our main contribution is to investigate whether the satisfaction of properties by argument-based ranking semantics carries over to the lifted, claim-based, variants of the corresponding properties and semantics.

ConArg is a suite of tools that offers a wide series of applications for dealing with argumentation problems. In this work, we present the advances we made in implementing a ranking-based semantics, based on computational choice power indexes, within ConArg. Such kind of semantics represents a method for sorting the arguments of an abstract argumentation framework, according to some preference relation. The ranking-based semantics we implement relies on Shapley, Banzhaf, Deegan-Packel and Johnston power index, transferring well know properties from computational social choice to argumentation framework ranking-based semantics.

The issue of handling dynamics is a central problem in Argumentation Theory. In order to understand how dynamics work, we extended the ConArg suite with Rob, a tool that is able to display Abstract Argumentation Frameworks and their corresponding sets of extensions, in a way suitable to understand what happens to the semantics when a modification to the graph occurs. In particular, Rob allows to inspect for a particular framework all the corresponding sets of extensions, and for every extension all the frameworks which admit it for some semantics.

In Probabilistic Abstract Argumentation, arguments and attacks (nodes and edges) in a graph instance are associated with a probability value. These probabilities can be interpreted in different ways: for instance, in the constellation approaches, the probabilities introduce uncertainty in the topology of the graph. In this paper we use MetaProbLog, a ProbLog framework where facts in a logic program are annotated by probabilities; the purpose is to compute the probability of possible worlds of arguments. The tool is integrated in the web interface of ConArg, a constraint-programming based tool aimed to solve different problems in Abstract Argumentation.