Cases or examples can be often explained by the interplay of arguments in favour or against their outcomes. This paper addresses the problem of finding explanations for a collection of cases where an explanation is a labelled argumentation graph consistent with the cases, and a case is represented as a statement labelling. The focus is on semi-abstract argumentation graphs specifying attack and subargument relations between arguments, along with particular complete argument labellings taken from probabilistic argumentation where arguments can be excluded. (...) The framework and computational considerations lead to an iterated local iterative deepening depth-first search to collect ‘interesting’ semi-abstract argumentation graphs consistent with a collection of statement labellings. Finally, the framework is illustrated and evaluated with a proof-of-concept implementation. (shrink)

This paper proposes a formal top-level model of explaining the outputs of machine-learning-based decision-making applications and evaluates it experimentally with three data sets. The model draws on AI & law research on argumentation with cases, which models how lawyers draw analogies to past cases and discuss their relevant similarities and differences in terms of relevant factors and dimensions in the problem domain. A case-based approach is natural since the input data of machine-learning applications can be seen as cases. While the (...) approach is motivated by legal decision making, it also applies to other kinds of decision making, such as commercial decisions about loan applications or employee hiring, as long as the outcome is binary and the input conforms to this paper’s factor- or dimension format. The model is top-level in that it can be extended with more refined accounts of similarities and differences between cases. It is shown to overcome several limitations of similar argumentation-based explanation models, which only have binary features and do not represent the tendency of features towards particular outcomes. The results of the experimental evaluation studies indicate that the model may be feasible in practice, but that further development and experimentation is needed to confirm its usefulness as an explanation model. Main challenges here are selecting from a large number of possible explanations, reducing the number of features in the explanations and adding more meaningful information to them. It also remains to be investigated how suitable our approach is for explaining non-linear models. (shrink)

There are many benefits of using argumentation-based techniques in multi-agent systems, as clearly shown in the literature. Such benefits come not only from the expressiveness that argumentation-based techniques bring to agent communication but also from the reasoning and decision-making capabilities under conditions of conflicting and uncertain information that argumentation enables for autonomous agents. When developing multi-agent applications in which argumentation will be used to improve agent communication and reasoning, argumentation schemes are useful in addressing the requirements of the application domain (...) in regards to argumentation. In this work, we propose an argumentation framework that takes into account the particular structure of argumentation schemes at its core. This paper formally defines such a framework and experimentally evaluates its implementation for both argumentation-based reasoning and dialogues. (shrink)

argumentation, as originally defined by Dung, is a model that allows the description of certain information about arguments and relationships between them: in an abstract argumentation framework, the agent knows for sure whether a given argument or attack exists. It means that the absence of an attack between two arguments can be interpreted as “we know that the first argument does not attack the second one”. But the question of uncertainty in abstract argumentation has received much attention in the last (...) years. In this paper, we survey approaches that allow to express information like “There may be an attack between these arguments”. We describe the main models that incorporate qualitative uncertainty in abstract argumentation, as well as some applications of these models. We also highlight some open questions that deserve some attention in the future. (shrink)

We all seem to have a sense of what good and bad arguments are, and there is a long history—focusing on fallacies—of trying to provide objective standards that would allow a clear separation of good and bad arguments. This contribution discusses the limits of attempts to determine the quality of arguments. It begins with defining bad arguments as those that deviate from an established standard of good arguments. Since there are different conceptualizations of “argument”—as controversy, as debate, and as justification—and (...) since arguments in each of these senses can be used for different purposes, a first problem is that we would need a large variety of standards for “good” arguments. After this, the contribution focuses in particular on proposals made in the literature on how to assess the quality of arguments in the sense of justification. It distinguishes three problems of assessment: How to determine whether reasons are acceptable, whether reasons are sufficient to justify the conclusion, and how to identify arguments in real-world speech acts and texts? It is argued that limitations of argument assessment result from unavoidable relativism: The assessment of many—if not most—arguments depends on the epistemic situation of the evaluator. (shrink)

This work considers the formalisation of the merging process of stratified belief bases, where beliefs are stored in different layers or strata. Their strata are ranked, following a total order, employing the value the agent using the belief base assigns to these beliefs. The agent uses an argumentation mechanism to reason from the belief base and obtain the final inferences. We present two ways of merging stratified belief bases: the first is defined by merging two strata without belief preservation, and (...) the second by merging two strata with belief preservation. For these operators, as layers merge, it is possible that some information has to be deleted from the stratum to maintain consistency. The latter merge operator will be defined as performing their actions so that information removed from the stratum is preserved after the merging by moving it to a stratum containing information perceived as having a lower value and where those beliefs may trigger a new revision, possibly starting a cascading effect. (shrink)

Scientific research in the formal sciences comes in multiple degrees of formality: fully formal work; rigorous proofs that practitioners know to be formalizable in principle; and informal work like rough proof sketches and considerations about the advantages and disadvantages of various formal systems. This informal work includes informal and semi-formal debates between formal scientists, e.g. about the acceptability of foundational principles and proposed axiomatizations. In this paper, we propose to use the methodology of structured argumentation theory to produce a formal (...) model of such informal and semi-formal debates in the formal sciences. For this purpose, we propose ASPIC-END, an adaptation of the structured argumentation framework ASPIC+ which can incorporate natural deduction style arguments and explanations. We illustrate the applicability of the framework to debates in the formal sciences by presenting a simple model of some arguments about proposed solutions to the Liar paradox, and by discussing a more extensive—but still preliminary—model of parts of the debate that mathematicians had about the Axiom of Choice in the early twentieth century. (shrink)

This work formalizes an informant-based structured argumentation approach in a multi-agent setting, where the knowledge base of an agent may include information provided by other agents, and each piece of knowledge comes attached with its informant. In that way, arguments are associated with the set of informants corresponding to the information they are built upon. Our approach proposes an informant-based notion of argument strength, where the strength of an argument is determined by the credibility of its informant agents. Moreover, we (...) consider that the strength of an argument is not absolute, but it is relative to the resolution of the conflicts the argument is involved in. In other words, the strength of an argument may vary from one context to another, as it will be determined by comparison to its attacking arguments (respectively, the arguments it attacks). Finally, we equip agents with the means to express reasons for or against the consideration of any piece of information provided by a given informant agent. Consequently, we allow agents to argue about the arguments’ strength through the construction of arguments that challenge (respectively, defeat) or are in favour of their informant agents. (shrink)

In this paper we introduce hypersequent-based frameworks for the modelling of defeasible reasoning by means of logic-based argumentation and the induced entailment relations. These structures are an extension of sequent-based argumentation frameworks, in which arguments and the attack relations among them are expressed not only by Gentzen-style sequents, but by more general expressions, called hypersequents. This generalization allows us to overcome some of the known weaknesses of logical argumentation frameworks and to prove several desirable properties of the entailments that are (...) induced by the extended frameworks. It also allows us to incorporate as the deductive base of our formalism some well-known logics, which lack cut-free sequent calculi, and so are not adequate for standard sequent-based argumentation. We show that hypersequent-based argumentation yields robust defeasible variants of these logics, with many desirable properties. (shrink)

This paper aims at comparing and relating belief revision and argumentation as approaches to model reasoning processes. Referring to some prominent literature references in both fields, we will discuss their (implicit or explicit) assumptions on the modeled processes and hence commonalities and differences in the forms of reasoning they are suitable to deal with. The intended contribution is on one hand assessing the (not fully explored yet) relationships between two lively research fields in the broad area of defeasible reasoning and (...) on the other hand pointing out open issues and potential directions for future research. (shrink)

Structured argumentation systems, and their implementation, represent an important research subject in the area of Knowledge Representation and Reasoning. Structured argumentation advances over abstract argumentation frameworks by providing the internal construction of the arguments that are usually defined by a set of (strict and defeasible) rules. By considering the structure of arguments, it becomes possible to analyze reasons for and against a conclusion, and the warrant status of such a claim in the context of a knowledge base represents the main (...) output of a dialectical process. Computing such statuses is a costly process, and any update to the knowledge base could potentially have a huge impact if done naively. In this work, we investigate the case of updates consisting of both additions and removals of pieces of knowledge in the Defeasible Logic Programming (DeLP) framework, first analyzing the complexity of the problem and then identifying conditions under which we can avoid unnecessary computations—central to this is the development of structures (e.g. graphs) to keep track of which results can potentially be affected by a given update. We introduce a technique for the incremental computation of the warrant statuses of conclusions in DeLP knowledge bases that evolve due to the application of (sets of) updates. We present the results of a thorough experimental evaluation showing that our incremental approach yields significantly faster running times in practice, as well as overall fewer recomputations, even in the case of sets of updates performed simultaneously. (shrink)