Research Papers

Why do explanations fail? A typology and discussion on failures in XAI

Clara Bove, Thibault Laugel, Marie-Jeanne Lesot +2

As Machine Learning models achieve unprecedented levels of performance, the XAI domain aims at making these models understandable by presenting end-users with intelligible explanations. Yet, some existing XAI approaches fail to meet expectations: several issues have been reported in the literature, generally pointing out either technical limitations or misinterpretations by users. In this paper, we argue that the resulting harms arise from a complex overlap of multiple failures in XAI, which existing ad-hoc studies fail to capture. This work therefore advocates for a holistic perspective, presenting a systematic investigation of limitations of current XAI methods and their impact on the interpretation of explanations. % By distinguishing between system-specific and user-specific failures, we propose a typological framework that helps revealing the nuanced complexities of explanation failures. Leveraging this typology, we discuss some research directions to help practitioners better understand the limitations of XAI systems and enhance the quality of ML explanations.

Instead of building deep learning models from scratch, developers are more and more relying on adapting pre-trained models to their customized tasks. However, powerful pre-trained models may be misused for unethical or illegal tasks, e.g., privacy inference and unsafe content generation. In this paper, we introduce a pioneering learning paradigm, non-fine-tunable learning, which prevents the pre-trained model from being fine-tuned to indecent tasks while preserving its performance on the original task. To fulfill this goal, we propose SOPHON, a protection framework that reinforces a given pre-trained model to be resistant to being fine-tuned in pre-defined restricted domains. Nonetheless, this is challenging due to a diversity of complicated fine-tuning strategies that may be adopted by adversaries. Inspired by model-agnostic meta-learning, we overcome this difficulty by designing sophisticated fine-tuning simulation and fine-tuning evaluation algorithms. In addition, we carefully design the optimization process to entrap the pre-trained model within a hard-to-escape local optimum regarding restricted domains. We have conducted extensive experiments on two deep learning modes (classification and generation), seven restricted domains, and six model architectures to verify the effectiveness of SOPHON. Experiment results verify that fine-tuning SOPHON-protected models incurs an overhead comparable to or even greater than training from scratch. Furthermore, we confirm the robustness of SOPHON to three fine-tuning methods, five optimizers, various learning rates and batch sizes. SOPHON may help boost further investigations into safe and responsible AI.

As the adoption of explainable AI (XAI) continues to expand, the urgency to address its privacy implications intensifies. Despite a growing corpus of research in AI privacy and explainability, there is little attention on privacy-preserving model explanations. This article presents the first thorough survey about privacy attacks on model explanations and their countermeasures. Our contribution to this field comprises a thorough analysis of research papers with a connected taxonomy that facilitates the categorisation of privacy attacks and countermeasures based on the targeted explanations. This work also includes an initial investigation into the causes of privacy leaks. Finally, we discuss unresolved issues and prospective research directions uncovered in our analysis. This survey aims to be a valuable resource for the research community and offers clear insights for those new to this domain. To support ongoing research, we have established an online resource repository, which will be continuously updated with new and relevant findings. Interested readers are encouraged to access our repository at https://github.com/tamlhp/awesome-privex.

In this paper, we delve into the rapidly evolving challenge of misinformation detection, with a specific focus on the nuanced manipulation of narrative frames - an under-explored area within the AI community. The potential for Generative AI models to generate misleading narratives underscores the urgency of this problem. Drawing from communication and framing theories, we posit that the presentation or 'framing' of accurate information can dramatically alter its interpretation, potentially leading to misinformation. We highlight this issue through real-world examples, demonstrating how shifts in narrative frames can transmute fact-based information into misinformation. To tackle this challenge, we propose an innovative approach leveraging the power of pre-trained Large Language Models and deep neural networks to detect misinformation originating from accurate facts portrayed under different frames. These advanced AI techniques offer unprecedented capabilities in identifying complex patterns within unstructured data critical for examining the subtleties of narrative frames. The objective of this paper is to bridge a significant research gap in the AI domain, providing valuable insights and methodologies for tackling framing-induced misinformation, thus contributing to the advancement of responsible and trustworthy AI technologies. Several experiments are intensively conducted and experimental results explicitly demonstrate the various impact of elements of framing theory proving the rationale of applying framing theory to increase the performance in misinformation detection.

AI systems cannot exist without data. Now that AI models (data science and AI) have matured and are readily available to apply in practice, most organizations struggle with the data infrastructure to do so. There is a growing need for data engineers that know how to prepare data for AI systems or that can setup enterprise-wide data architectures for analytical projects. But until now, the data engineering part of AI engineering has not been getting much attention, in favor of discussing the modeling part. In this paper we aim to change this by perform a mapping study on data engineering for AI systems, i.e., AI data engineering. We found 25 relevant papers between January 2019 and June 2023, explaining AI data engineering activities. We identify which life cycle phases are covered, which technical solutions or architectures are proposed and which lessons learned are presented. We end by an overall discussion of the papers with implications for practitioners and researchers. This paper creates an overview of the body of knowledge on data engineering for AI. This overview is useful for practitioners to identify solutions and best practices as well as for researchers to identify gaps.

Deepfakes represent one of the toughest challenges in the world of Cybersecurity and Digital Forensics, especially considering the high-quality results obtained with recent generative AI-based solutions. Almost all generative models leave unique traces in synthetic data that, if analyzed and identified in detail, can be exploited to improve the generalization limitations of existing deepfake detectors. In this paper we analyzed deepfake images in the frequency domain generated by both GAN and Diffusion Model engines, examining in detail the underlying statistical distribution of Discrete Cosine Transform (DCT) coefficients. Recognizing that not all coefficients contribute equally to image detection, we hypothesize the existence of a unique ``discriminative fingerprint", embedded in specific combinations of coefficients. To identify them, Machine Learning classifiers were trained on various combinations of coefficients. In addition, the Explainable AI (XAI) LIME algorithm was used to search for intrinsic discriminative combinations of coefficients. Finally, we performed a robustness test to analyze the persistence of traces by applying JPEG compression. The experimental results reveal the existence of traces left by the generative models that are more discriminative and persistent at JPEG attacks. Code and dataset are available at https://github.com/opontorno/dcts_analysis_deepfakes.

Understanding Teacher Perspectives and Experiences after Deployment of AI Literacy Curriculum in Middle-school Classrooms

Prerna Ravi, Annalisa Broski, Glenda Stump +3

Artificial Intelligence (AI) and its associated applications are ubiquitous in today's world, making it imperative that students and their teachers understand how it works and the ramifications arising from its usage. In this study, we investigate the experiences of seven teachers following their implementation of modules from the MIT RAICA (Responsible AI for Computational Action) curriculum. Through semi-structured interviews, we investigated their instructional strategies as they engaged with the AI curriculum in their classroom, how their teaching and learning beliefs about AI evolved with the curriculum as well as how those beliefs impacted their implementation of the curriculum. Our analysis suggests that the AI modules not only expanded our teachers' knowledge in the field, but also prompted them to recognize its daily applications and their ethical and societal implications, so that they could better engage with the content they deliver to students. Teachers were able to leverage their own interdisciplinary backgrounds to creatively introduce foundational AI topics to students to maximize engagement and playful learning. Our teachers advocated their need for better external support when navigating technological resources, additional time for preparation given the novelty of the curriculum, more flexibility within curriculum timelines, and additional accommodations for students of determination. Our findings provide valuable insights for enhancing future iterations of AI literacy curricula and teacher professional development (PD) resources.

Causal machine learning tools are beginning to see use in real-world policy evaluation tasks to flexibly estimate treatment effects. One issue with these methods is that the machine learning models used are generally black boxes, i.e., there is no globally interpretable way to understand how a model makes estimates. This is a clear problem in policy evaluation applications, particularly in government, because it is difficult to understand whether such models are functioning in ways that are fair, based on the correct interpretation of evidence and transparent enough to allow for accountability if things go wrong. However, there has been little discussion of transparency problems in the causal machine learning literature and how these might be overcome. This paper explores why transparency issues are a problem for causal machine learning in public policy evaluation applications and considers ways these problems might be addressed through explainable AI tools and by simplifying models in line with interpretable AI principles. It then applies these ideas to a case-study using a causal forest model to estimate conditional average treatment effects for a hypothetical change in the school leaving age in Australia. It shows that existing tools for understanding black-box predictive models are poorly suited to causal machine learning and that simplifying the model to make it interpretable leads to an unacceptable increase in error (in this application). It concludes that new tools are needed to properly understand causal machine learning models and the algorithms that fit them.

Deep Learning has already been successfully applied to analyze industrial sensor data in a variety of relevant use cases. However, the opaque nature of many well-performing methods poses a major obstacle for real-world deployment. Explainable AI (XAI) and especially feature attribution techniques promise to enable insights about how such models form their decision. But the plain application of such methods often fails to provide truly informative and problem-specific insights to domain experts. In this work, we focus on the specific task of detecting faults in rolling element bearings from vibration signals. We propose a novel and domain-specific feature attribution framework that allows us to evaluate how well the underlying logic of a model corresponds with expert reasoning. Utilizing the framework we are able to validate the trustworthiness and to successfully anticipate the generalization ability of different well-performing deep learning models. Our methodology demonstrates how signal processing tools can effectively be used to enhance Explainable AI techniques and acts as a template for similar problems.

This research investigates road traffic accident severity in the UK, using a combination of machine learning, econometric, and statistical methods on historical data. We employed various techniques, including correlation analysis, regression models, GMM for error term issues, and time-series forecasting with VAR and ARIMA models. Our approach outperforms naive forecasting with an MASE of 0.800 and ME of -73.80. We also built a random forest classifier with 73% precision, 78% recall, and a 73% F1-score. Optimizing with H2O AutoML led to an XGBoost model with an RMSE of 0.176 and MAE of 0.087. Factor Analysis identified key variables, and we used SHAP for Explainable AI, highlighting influential factors like Driver_Home_Area_Type and Road_Type. Our study enhances understanding of accident severity and offers insights for evidence-based road safety policies.

Learning by Self-Explaining

Wolfgang Stammer, Felix Friedrich, David Steinmann +3

Much of explainable AI research treats explanations as a means for model inspection. Yet, this neglects findings from human psychology that describe the benefit of self-explanations in an agent's learning process. Motivated by this, we introduce a novel workflow in the context of image classification, termed Learning by Self-Explaining (LSX). LSX utilizes aspects of self-refining AI and human-guided explanatory machine learning. The underlying idea is that a learner model, in addition to optimizing for the original predictive task, is further optimized based on explanatory feedback from an internal critic model. Intuitively, a learner's explanations are considered "useful" if the internal critic can perform the same task given these explanations. We provide an overview of important components of LSX and, based on this, perform extensive experimental evaluations via three different example instantiations. Our results indicate improvements via Learning by Self-Explaining on several levels: in terms of model generalization, reducing the influence of confounding factors, and providing more task-relevant and faithful model explanations. Overall, our work provides evidence for the potential of self-explaining within the learning phase of an AI model.

A Deep Dive into Perturbations as Evaluation Technique for Time Series XAI

Udo Schlegel, Daniel A. Keim

Explainable Artificial Intelligence (XAI) has gained significant attention recently as the demand for transparency and interpretability of machine learning models has increased. In particular, XAI for time series data has become increasingly important in finance, healthcare, and climate science. However, evaluating the quality of explanations, such as attributions provided by XAI techniques, remains challenging. This paper provides an in-depth analysis of using perturbations to evaluate attributions extracted from time series models. A perturbation analysis involves systematically modifying the input data and evaluating the impact on the attributions generated by the XAI method. We apply this approach to several state-of-the-art XAI techniques and evaluate their performance on three time series classification datasets. Our results demonstrate that the perturbation analysis approach can effectively evaluate the quality of attributions and provide insights into the strengths and limitations of XAI techniques. Such an approach can guide the selection of XAI methods for time series data, e.g., focusing on return time rather than precision, and facilitate the development of more reliable and interpretable machine learning models for time series analysis.