Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI

This paper presents an overview of the field of software systems requirements engineering (RE). It describes the main areas of RE practice, and highlights some key open research issues for the future.

/pdf/requirements-engineering-a-roadmap-3lcu5ikya8.pdf

Requirements engineering: a roadmap

Goals capture, at different levels of abstraction, the various objectives the system under consideration should achieve. Goal-oriented requirements engineering is concerned with the use of goals for eliciting, elaborating, structuring, specifying, analyzing, negotiating, documenting, and modifying requirements. This area has received increasing attention. The paper reviews various research efforts undertaken along this line of research. The arguments in favor of goal orientation are first briefly discussed. The paper then compares the main approaches to goal modeling, goal specification and goal-based reasoning in the many activities of the requirements engineering process. To make the discussion more concrete, a real case study is used to suggest what a goal-oriented requirements engineering method may look like. Experience, with such approaches and tool support are briefly discussed as well.

/pdf/goal-oriented-requirements-engineering-a-guided-tour-2uwx0ha66r.pdf

Goal-oriented requirements engineering: a guided tour

In the last years, Artificial Intelligence (AI) has achieved a notable momentum that may deliver the best of expectations over many application sectors across the field. For this to occur, the entire community stands in front of the barrier of explainability, an inherent problem of AI techniques brought by sub-symbolism (e.g. ensembles or Deep Neural Networks) that were not present in the last hype of AI. Paradigms underlying this problem fall within the so-called eXplainable AI (XAI) field, which is acknowledged as a crucial feature for the practical deployment of AI models. This overview examines the existing literature in the field of XAI, including a prospect toward what is yet to be reached. We summarize previous efforts to define explainability in Machine Learning, establishing a novel definition that covers prior conceptual propositions with a major focus on the audience for which explainability is sought. We then propose and discuss about a taxonomy of recent contributions related to the explainability of different Machine Learning models, including those aimed at Deep Learning methods for which a second taxonomy is built. This literature analysis serves as the background for a series of challenges faced by XAI, such as the crossroads between data fusion and explainability. Our prospects lead toward the concept of Responsible Artificial Intelligence, namely, a methodology for the large-scale implementation of AI methods in real organizations with fairness, model explainability and accountability at its core. Our ultimate goal is to provide newcomers to XAI with a reference material in order to stimulate future research advances, but also to encourage experts and professionals from other disciplines to embrace the benefits of AI in their activity sectors, without any prior bias for its lack of interpretability.

Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI.

Reasoning is generally seen as a means to improve knowledge and make better decisions. However, much evidence shows that reasoning often leads to epistemic distortions and poor decisions. This suggests that the function of reasoning should be rethought. Our hypothesis is that the function of reasoning is argumentative. It is to devise and evaluate arguments intended to persuade. Reasoning so conceived is adaptive given the exceptional dependence of humans on communication and their vulnerability to misinformation. A wide range of evidence in the psychology of reasoning and decision making can be reinterpreted and better explained in the light of this hypothesis. Poor performance in standard reasoning tasks is explained by the lack of argumentative context. When the same problems are placed in a proper argumentative setting, people turn out to be skilled arguers. Skilled arguers, however, are not after the truth but after arguments supporting their views. This explains the notorious confirmation bias. This bias is apparent not only when people are actually arguing, but also when they are reasoning proactively from the perspective of having to defend their opinions. Reasoning so motivated can distort evaluations and attitudes and allow erroneous beliefs to persist. Proactively used reasoning also favors decisions that are easy to justify but not necessarily better. In all these instances traditionally described as failures or flaws, reasoning does exactly what can be expected of an argumentative device: Look for arguments that support a given conclusion, and, ceteris paribus, favor conclusions for which arguments can be found.

/pdf/why-do-humans-reason-arguments-for-an-argumentative-theory-19genwgdwy.pdf

Why Do Humans Reason? Arguments for an Argumentative Theory

Logic-based formalizations of argumentation, that take pros and cons for some claim into account, have been extensively studied, and some basic principles have been established (for reviews see [1-3]). These formalizations assume a set of formulae and then exhaustively lay out arguments and counterarguments, where a counterargument either rebuts (i.e. negates the claim of the argument) or undercuts (i.e. negates the support of the argument). Recently attempts have been made to refine these formalizations by using techniques for selecting the more appropriate arguments and counterarguments by taking into account intrinsic factors (such as the degree of inconsistency between an argument and its counterarguments) and extrinsic factors (such as the impact of particular arguments on the audience and the beliefs of the audience). In this presentation, we consider the need to take intrinsic and extrinsic factors into account, and then consider ways that this can be done in logic in order to refine existing logic-based approaches to argumentation. These refinements offer interesting options for formalizations that may better capture practical argumentation for intelligent agents [3].

https://link.springer.com/content/pdf/10.1007%2F978-3-540-75256-1_3.pdf

Elements of argumentation

The development of most large and complex systems necessarily involves many people-each with their own perspectives on the system defined by their knowledge, responsibilities, and commitments. To address this we have advocated distributed development of specifications from multiple perspectives. However, this leads to problems of identifying and handling inconsistencies between such perspectives. Maintaining absolute consistency is not always possible. Often this is not even desirable since this can unnecessarily constrain the development process, and can lead to the loss of important information. Indeed since the real-world forces us to work with inconsistencies, we should formalize some of the usually informal or extra-logical ways of responding to them. This is not necessarily done by eradicating inconsistencies but rather by supplying logical rules specifying how we should act on them. To achieve this, we combine two lines of existing research: the ViewPoints framework for perspective development, interaction and organization, and a logic-based approach to inconsistency handling. This paper presents our technique for inconsistency handling in the ViewPoints framework by using simple examples. >

/pdf/inconsistency-handling-in-multiperspective-specifications-2t3q1t2e7h.pdf

Inconsistency handling in multiperspective specifications

In practical reasoning, it is common to have “too much” informationabout some situation. In other words, it is common for there to be classically inconsistent information in a practical reasoning database [Besnard et al., 1995]. The diversity of logics proposed for aspects of practical reasoning indicates the complexity of this form of reasoning. However, central to practical reasoning seems to be the need to reason with inconsistent information without the logic being trivialized [Gabbay and Hunter, 1991; Finkelstein et al., 1994]. This is the need to derive reasonable inferences without deriving the trivial inferences that follow the ex falso quodlibet proof rule that holds in classical logic.

Paraconsistent logics

http://www2.fiit.stuba.sk/~kvasnicka/CognitiveScience/8.prednaska/BesnardHunter_AI.pdf

A logic-based theory of deductive arguments

https://www.cs.ox.ac.uk/people/michael.wooldridge/pubs/aij2011a.pdf

Anthony Hunter

Papers

Elements of argumentation

Inconsistency handling in multiperspective specifications

Paraconsistent logics

A logic-based theory of deductive arguments

Weighted argument systems: Basic definitions, algorithms, and complexity results