Research Papers

Biased datasets are ubiquitous and present a challenge for machine learning. For a number of categories on a dataset that are equally important but some are sparse and others are common, the learning algorithms will favor the ones with more presence. The problem of biased datasets is especially sensitive when dealing with minority people groups. How can we, from biased data, generate algorithms that treat every person equally? This work explores one way to mitigate bias using a debiasing variational autoencoder with experiments on facial expression recognition.

Deep neural networks have been widely adopted in numerous domains due to their high performance and accessibility to developers and application-specific end-users. Fundamental to image-based applications is the development of Convolutional Neural Networks (CNNs), which possess the ability to automatically extract features from data. However, comprehending these complex models and their learned representations, which typically comprise millions of parameters and numerous layers, remains a challenge for both developers and end-users. This challenge arises due to the absence of interpretable and transparent tools to make sense of black-box models. There exists a growing body of Explainable Artificial Intelligence (XAI) literature, including a collection of methods denoted Class Activation Maps (CAMs), that seek to demystify what representations the model learns from the data, how it informs a given prediction, and why it, at times, performs poorly in certain tasks. We propose a novel XAI visualization method denoted CAManim that seeks to simultaneously broaden and focus end-user understanding of CNN predictions by animating the CAM-based network activation maps through all layers, effectively depicting from end-to-end how a model progressively arrives at the final layer activation. Herein, we demonstrate that CAManim works with any CAM-based method and various CNN architectures. Beyond qualitative model assessments, we additionally propose a novel quantitative assessment that expands upon the Remove and Debias (ROAD) metric, pairing the qualitative end-to-end network visual explanations assessment with our novel quantitative "yellow brick ROAD" assessment (ybROAD). This builds upon prior research to address the increasing demand for interpretable, robust, and transparent model assessment methodology, ultimately improving an end-user's trust in a given model's predictions.

As AI/ML models, including Large Language Models, continue to scale with massive datasets, so does their consumption of undeniably limited natural resources, and impact on society. In this collaboration between AI, Sustainability, HCI and legal researchers, we aim to enable a transition to sustainable AI development by enabling stakeholders across the AI value chain to assess and quantitfy the environmental and societal impact of AI. We present the ESG Digital and Green Index (DGI), which offers a dashboard for assessing a company's performance in achieving sustainability targets. This includes monitoring the efficiency and sustainable use of limited natural resources related to AI technologies (water, electricity, etc). It also addresses the societal and governance challenges related to AI. The DGI creates incentives for companies to align their pathway with the Sustainable Development Goals (SDGs). The value, challenges and limitations of our methodology and findings are discussed in the paper.

Artificial intelligence (AI) recently had its 'iPhone moment' and adoption has drastically accelerated. Quantum computing appears poised to follow suit over the next years. However, while there has been discourse about how to use AI responsibly, there is still little appreciation and awareness among executives, managers, and practitioners about the broader ethical questions and implications raised by the intersection of these emerging technologies. In this article, it is highlighted why quantum computing and AI ethics must be taken seriously by businesspersons and how these technologies affect strategic decisions; moreover, recommendations and action areas are formulated.

The integration of AI into radiology introduces opportunities for improved clinical care provision and efficiency but it demands a meticulous approach to mitigate potential risks as with any other new technology. Beginning with rigorous pre-deployment evaluation and validation, the focus should be on ensuring models meet the highest standards of safety, effectiveness and efficacy for their intended applications. Input and output guardrails implemented during production usage act as an additional layer of protection, identifying and addressing individual failures as they occur. Continuous post-deployment monitoring allows for tracking population-level performance (data drift), fairness, and value delivery over time. Scheduling reviews of post-deployment model performance and educating radiologists about new algorithmic-driven findings is critical for AI to be effective in clinical practice. Recognizing that no single AI solution can provide absolute assurance even when limited to its intended use, the synergistic application of quality assurance at multiple levels - regulatory, clinical, technical, and ethical - is emphasized. Collaborative efforts between stakeholders spanning healthcare systems, industry, academia, and government are imperative to address the multifaceted challenges involved. Trust in AI is an earned privilege, contingent on a broad set of goals, among them transparently demonstrating that the AI adheres to the same rigorous safety, effectiveness and efficacy standards as other established medical technologies. By doing so, developers can instil confidence among providers and patients alike, enabling the responsible scaling of AI and the realization of its potential benefits. The roadmap presented herein aims to expedite the achievement of deployable, reliable, and safe AI in radiology.

Work in AI ethics and fairness has made much progress in regulating LLMs to reflect certain values, such as fairness, truth, and diversity. However, it has taken the problem of how LLMs might 'mean' anything at all for granted. Without addressing this, it is not clear what imbuing LLMs with such values even means. In response, we provide a general theory of meaning that extends beyond humans. We use this theory to explicate the precise nature of LLMs as meaning-agents. We suggest that the LLM, by virtue of its position as a meaning-agent, already grasps the constructions of human society (e.g. morality, gender, and race) in concept. Consequently, under certain ethical frameworks, currently popular methods for model alignment are limited at best and counterproductive at worst. Moreover, unaligned models may help us better develop our moral and social philosophy.

On the stability, correctness and plausibility of visual explanation methods based on feature importance

Romain Xu-Darme, Jenny Benois-Pineau, Romain Giot +4

In the field of Explainable AI, multiples evaluation metrics have been proposed in order to assess the quality of explanation methods w.r.t. a set of desired properties. In this work, we study the articulation between the stability, correctness and plausibility of explanations based on feature importance for image classifiers. We show that the existing metrics for evaluating these properties do not always agree, raising the issue of what constitutes a good evaluation metric for explanations. Finally, in the particular case of stability and correctness, we show the possible limitations of some evaluation metrics and propose new ones that take into account the local behaviour of the model under test.

Explainable AI (XAI) methods have mostly been built to investigate and shed light on single machine learning models and are not designed to capture and explain differences between multiple models effectively. This paper addresses the challenge of understanding and explaining differences between machine learning models, which is crucial for model selection, monitoring and lifecycle management in real-world applications. We propose DeltaXplainer, a model-agnostic method for generating rule-based explanations describing the differences between two binary classifiers. To assess the effectiveness of DeltaXplainer, we conduct experiments on synthetic and real-world datasets, covering various model comparison scenarios involving different types of concept drift.

With ubiquitous exposure of AI systems today, we believe AI development requires crucial considerations to be deemed trustworthy. While the potential of AI systems is bountiful, though, is still unknown-as are their risks. In this work, we offer a brief, high-level overview of societal impacts of AI systems. To do so, we highlight the requirement of multi-disciplinary governance and convergence throughout its lifecycle via critical systemic examinations (e.g., energy consumption), and later discuss induced effects on the environment (i.e., carbon footprint) and its users (i.e., social development). In particular, we consider these impacts from a multi-disciplinary perspective: computer science, sociology, environmental science, and so on to discuss its inter-connected societal risks and inability to simultaneously satisfy aspects of well-being. Therefore, we accentuate the necessity of holistically addressing pressing concerns of AI systems from a socioethical impact assessment perspective to explicate its harmful societal effects to truly enable humanity-centered Trustworthy AI.

There are now many explainable AI methods for understanding the decisions of a machine learning model. Among these are those based on counterfactual reasoning, which involve simulating features changes and observing the impact on the prediction. This article proposes to view this simulation process as a source of creating a certain amount of knowledge that can be stored to be used, later, in different ways. This process is illustrated in the additive model and, more specifically, in the case of the naive Bayes classifier, whose interesting properties for this purpose are shown.

Causal machine learning methods which flexibly generate heterogeneous treatment effect estimates could be very useful tools for governments trying to make and implement policy. However, as the critical artificial intelligence literature has shown, governments must be very careful of unintended consequences when using machine learning models. One way to try and protect against unintended bad outcomes is with AI Fairness methods which seek to create machine learning models where sensitive variables like race or gender do not influence outcomes. In this paper we argue that standard AI Fairness approaches developed for predictive machine learning are not suitable for all causal machine learning applications because causal machine learning generally (at least so far) uses modelling to inform a human who is the ultimate decision-maker while AI Fairness approaches assume a model that is making decisions directly. We define these scenarios as indirect and direct decision-making respectively and suggest that policy-making is best seen as a joint decision where the causal machine learning model usually only has indirect power. We lay out a definition of fairness for this scenario - a model that provides the information a decision-maker needs to accurately make a value judgement about just policy outcomes - and argue that the complexity of causal machine learning models can make this difficult to achieve. The solution here is not traditional AI Fairness adjustments, but careful modelling and awareness of some of the decision-making biases that these methods might encourage which we describe.

This report first takes stock of XAI-related requirements appearing in various EU directives, regulations, guidelines, and CJEU case law. This analysis of existing requirements will permit us to have a clearer vision of the purposes, the ``why'', of XAI, which we separate into five categories: contestability, empowerment/redressing information asymmetries, control over system performance, evaluation of algorithmic decisions, and public administration transparency. The analysis of legal requirements also permits us to create four categories of recipients for explainability: data science teams; human operators of the system; persons affected by algorithmic decisions, and regulators/judges/auditors. Lastly, we identify four main operational contexts for explainability: XAI for the upstream design and testing phase; XAI for human-on-the-loop control; XAI for human-in-the-loop control; and XAI for ex-post challenges and investigations.Second, we will present user-centered design methodology, which takes the purposes, the recipients and the operational context into account in order to develop optimal XAI solutions.Third, we will suggest a methodology to permit suppliers and users of high-risk AI applications to propose local XAI solutions that are effective in the sense of being ``meaningful'', for example, useful in light of the operational, safety and fundamental rights contexts. The process used to develop these ``meaningful'' XAI solutions will be based on user-centric design principles examined in the second part.Fourth, we will suggest that the European Commission issue guidelines to provide a harmonised approach to defining ``meaningful'' explanations based on the purposes, audiences and operational contexts of AI systems. These guidelines would apply to the AI Act, but also to the other EU texts requiring explanations for algorithmic systems and results.