In the wake of responsible AI, interpretability methods, which attempt to provide an explanation for the predictions of neural models have seen rapid progress. In this work, we are concerned with explanations that are applicable to natural language processing (NLP) models and tasks, and we focus specifically on the analysis of counterfactual, contrastive explanations. We note that while there have been several explainers proposed to produce counterfactual explanations, their behaviour can vary significantly and the lack of a universal ground truth for the counterfactual edits imposes an insuperable barrier on their evaluation. We propose a new back translation-inspired evaluation methodology that utilises earlier outputs of the explainer as ground truth proxies to investigate the consistency of explainers. We show that by iteratively feeding the counterfactual to the explainer we can obtain valuable insights into the behaviour of both the predictor and the explainer models, and infer patterns that would be otherwise obscured. Using this methodology, we conduct a thorough analysis and propose a novel metric to evaluate the consistency of counterfactual generation approaches with different characteristics across available performance indicators.

Neural network models that can handle inputs of variable lengths are powerful, but often hard to interpret. The lack of transparency hinders their adoption in many domains. Explanation techniques are essential for improving transparency. However, existing model-agnostic general explanation techniques do not consider the variable lengths of input data points, which limits their effectiveness. To address this limitation, we propose ReX, a general framework for adapting various explanation techniques to models that process variable-length inputs, expanding explanation coverage to data points of different lengths. Our approach adds temporal information to the explanations generated by existing techniques without altering their core algorithms. We instantiate our approach on two popular explanation techniques: LIME and Anchors. To evaluate the effectiveness of ReX, we apply our approach to three models in two different tasks. Our evaluation results demonstrate that our approach significantly improves the fidelity and understandability of explanations.

Smart home environments are designed to provide services that help improve the quality of life for the occupant via a variety of sensors and actuators installed throughout the space. Many automated actions taken by a smart home are governed by the output of an underlying activity recognition system. However, activity recognition systems may not be perfectly accurate and therefore inconsistencies in smart home operations can lead users reliant on smart home predictions to wonder "why did the smart home do that?" In this work, we build on insights from Explainable Artificial Intelligence (XAI) techniques and introduce an explainable activity recognition framework in which we leverage leading XAI methods to generate natural language explanations that explain what about an activity led to the given classification. Within the context of remote caregiver monitoring, we perform a two-step evaluation: (a) utilize ML experts to assess the sensibility of explanations, and (b) recruit non-experts in two user remote caregiver monitoring scenarios, synchronous and asynchronous, to assess the effectiveness of explanations generated via our framework. Our results show that the XAI approach, SHAP, has a 92% success rate in generating sensible explanations. Moreover, in 83% of sampled scenarios users preferred natural language explanations over a simple activity label, underscoring the need for explainable activity recognition systems. Finally, we show that explanations generated by some XAI methods can lead users to lose confidence in the accuracy of the underlying activity recognition model. We make a recommendation regarding which existing XAI method leads to the best performance in the domain of smart home automation, and discuss a range of topics for future work to further improve explainable activity recognition.

Significant attention has been paid to enhancing recommender systems (RS) with explanation facilities to help users make informed decisions and increase trust in and satisfaction with the RS. Justification and transparency represent two crucial goals in explainable recommendation. Different from transparency, which faithfully exposes the reasoning behind the recommendation mechanism, justification conveys a conceptual model that may differ from that of the underlying algorithm. An explanation is an answer to a question. In explainable recommendation, a user would want to ask questions (referred to as intelligibility types) to understand results given by the RS. In this paper, we identify relationships between Why and How explanation intelligibility types and the explanation goals of justification and transparency. We followed the Human-Centered Design (HCD) approach and leveraged the What-Why-How visualization framework to systematically design and implement Why and How visual explanations in the transparent Recommendation and Interest Modeling Application (RIMA). Furthermore, we conducted a qualitative user study (N=12) to investigate the potential effects of providing Why and How explanations together in an explainable RS on the users' perceptions regarding transparency, trust, and satisfaction. Our study showed qualitative evidence confirming that the choice of the explanation intelligibility types depends on the explanation goal and user type.

Algorithms play a crucial role in many technological systems that control or affect various aspects of our lives. As a result, providing explanations for their decisions to address the needs of users and organisations is increasingly expected by laws, regulations, codes of conduct, and the public. However, as laws and regulations do not prescribe how to meet such expectations, organisations are often left to devise their own approaches to explainability, inevitably increasing the cost of compliance and good governance. Hence, we envision Explainability-by-Design, a holistic methodology characterised by proactive measures to include explanation capability in the design of decision-making systems. The methodology consists of three phases: (A) Explanation Requirement Analysis, (B) Explanation Technical Design, and (C) Explanation Validation. This paper describes phase (B), a technical workflow to implement explanation capability from requirements elicited by domain experts for a specific application context. Outputs of this phase are a set of configurations, allowing a reusable explanation service to exploit logs provided by the target application to create provenance traces of the application's decisions. The provenance then can be queried to extract relevant data points, which can be used in explanation plans to construct explanations personalised to their consumers. Following the workflow, organisations can design their decision-making systems to produce explanations that meet the specified requirements. To facilitate the process, we present a software architecture with reusable components to incorporate the resulting explanation capability into an application. Finally, we applied the workflow to two application scenarios and measured the associated development costs. It was shown that the approach is tractable in terms of development time, which can be as low as two hours per sentence.

We propose a novel approach to logic-based learning which generates assumption-based argumentation (ABA) frameworks from positive and negative examples, using a given background knowledge. These ABA frameworks can be mapped onto logic programs with negation as failure that may be non-stratified. Whereas existing argumentation-based methods learn exceptions to general rules by interpreting the exceptions as rebuttal attacks, our approach interprets them as undercutting attacks. Our learning technique is based on the use of transformation rules, including some adapted from logic program transformation rules (notably folding) as well as others, such as rote learning and assumption introduction. We present a general strategy that applies the transformation rules in a suitable order to learn stratified frameworks, and we also propose a variant that handles the non-stratified case. We illustrate the benefits of our approach with a number of examples, which show that, on one hand, we are able to easily reconstruct other logic-based learning approaches and, on the other hand, we can work out in a very simple and natural way problems that seem to be hard for existing techniques.

Several recent studies have elucidated why knowledge distillation (KD) improves model performance. However, few have researched the other advantages of KD in addition to its improving model performance. In this study, we have attempted to show that KD enhances the interpretability as well as the accuracy of models. We measured the number of concept detectors identified in network dissection for a quantitative comparison of model interpretability. We attributed the improvement in interpretability to the class-similarity information transferred from the teacher to student models. First, we confirmed the transfer of class-similarity information from the teacher to student model via logit distillation. Then, we analyzed how class-similarity information affects model interpretability in terms of its presence or absence and degree of similarity information. We conducted various quantitative and qualitative experiments and examined the results on different datasets, different KD methods, and according to different measures of interpretability. Our research showed that KD models by large models could be used more reliably in various fields.

A significant drawback of eXplainable Artificial Intelligence (XAI) approaches is the assumption of feature independence. This paper focuses on integrating causal knowledge in XAI methods to increase trust and help users assess explanations' quality. We propose a novel extension to a widely used local and model-agnostic explainer that explicitly encodes causal relationships in the data generated around the input instance to explain. Extensive experiments show that our method achieves superior performance comparing the initial one for both the fidelity in mimicking the black-box and the stability of the explanations.

Online discussion moderators must make ad-hoc decisions about whether the contributions of discussion participants are appropriate or should be removed to maintain civility. Existing research on offensive language and the resulting tools cover only one aspect among many involved in such decisions. The question of what is considered appropriate in a controversial discussion has not yet been systematically addressed. In this paper, we operationalize appropriate language in argumentation for the first time. In particular, we model appropriateness through the absence of flaws, grounded in research on argument quality assessment, especially in aspects from rhetoric. From these, we derive a new taxonomy of 14 dimensions that determine inappropriate language in online discussions. Building on three argument quality corpora, we then create a corpus of 2191 arguments annotated for the 14 dimensions. Empirical analyses support that the taxonomy covers the concept of appropriateness comprehensively, showing several plausible correlations with argument quality dimensions. Moreover, results of baseline approaches to assessing appropriateness suggest that all dimensions can be modeled computationally on the corpus.

Over the past decade explainable artificial intelligence has evolved from a predominantly technical discipline into a field that is deeply intertwined with social sciences. Insights such as human preference for contrastive -- more precisely, counterfactual -- explanations have played a major role in this transition, inspiring and guiding the research in computer science. Other observations, while equally important, have received much less attention. The desire of human explainees to communicate with artificial intelligence explainers through a dialogue-like interaction has been mostly neglected by the community. This poses many challenges for the effectiveness and widespread adoption of such technologies as delivering a single explanation optimised according to some predefined objectives may fail to engender understanding in its recipients and satisfy their unique needs given the diversity of human knowledge and intention. Using insights elaborated by Niklas Luhmann and, more recently, Elena Esposito we apply social systems theory to highlight challenges in explainable artificial intelligence and offer a path forward, striving to reinvigorate the technical research in this direction. This paper aims to demonstrate the potential of systems theoretical approaches to communication in understanding problems and limitations of explainable artificial intelligence.