Research Papers

A clinical artificial intelligence (AI) system is often validated on a held-out set of data which it has not been exposed to before (e.g., data from a different hospital with a distinct electronic health record system). This evaluation process is meant to mimic the deployment of an AI system on data in the wild; those which are currently unseen by the system yet are expected to be encountered in a clinical setting. However, when data in the wild differ from the held-out set of data, a phenomenon referred to as distribution shift, and lack ground-truth annotations, it becomes unclear the extent to which AI-based findings can be trusted on data in the wild. Here, we introduce SUDO, a framework for evaluating AI systems without ground-truth annotations. SUDO assigns temporary labels to data points in the wild and directly uses them to train distinct models, with the highest performing model indicative of the most likely label. Through experiments with AI systems developed for dermatology images, histopathology patches, and clinical reports, we show that SUDO can be a reliable proxy for model performance and thus identify unreliable predictions. We also demonstrate that SUDO informs the selection of models and allows for the previously out-of-reach assessment of algorithmic bias for data in the wild without ground-truth annotations. The ability to triage unreliable predictions for further inspection and assess the algorithmic bias of AI systems can improve the integrity of research findings and contribute to the deployment of ethical AI systems in medicine.

The evaluation of the fidelity of eXplainable Artificial Intelligence (XAI) methods to their underlying models is a challenging task, primarily due to the absence of a ground truth for explanations. However, assessing fidelity is a necessary step for ensuring a correct XAI methodology. In this study, we conduct a fair and objective comparison of the current state-of-the-art XAI methods by introducing three novel image datasets with reliable ground truth for explanations. The primary objective of this comparison is to identify methods with low fidelity and eliminate them from further research, thereby promoting the development of more trustworthy and effective XAI techniques. Our results demonstrate that XAI methods based on the backpropagation of output information to input yield higher accuracy and reliability compared to methods relying on sensitivity analysis or Class Activation Maps (CAM). However, the backpropagation method tends to generate more noisy saliency maps. These findings have significant implications for the advancement of XAI methods, enabling the elimination of erroneous explanations and fostering the development of more robust and reliable XAI.

Monitoring surface cracks in infrastructure is crucial for structural health monitoring. Automatic visual inspection offers an effective solution, especially in hard-to-reach areas. Machine learning approaches have proven their effectiveness but typically require large annotated datasets for supervised training. Once a crack is detected, monitoring its severity often demands precise segmentation of the damage. However, pixel-level annotation of images for segmentation is labor-intensive. To mitigate this cost, one can leverage explainable artificial intelligence (XAI) to derive segmentations from the explanations of a classifier, requiring only weak image-level supervision. This paper proposes applying this methodology to segment and monitor surface cracks. We evaluate the performance of various XAI methods and examine how this approach facilitates severity quantification and growth monitoring. Results reveal that while the resulting segmentation masks may exhibit lower quality than those produced by supervised methods, they remain meaningful and enable severity monitoring, thus reducing substantial labeling costs.

EgoBlur: Responsible Innovation in Aria

Nikhil Raina, Guruprasad Somasundaram, Kang Zheng +12

Project Aria pushes the frontiers of Egocentric AI with large-scale real-world data collection using purposely designed glasses with privacy first approach. To protect the privacy of bystanders being recorded by the glasses, our research protocols are designed to ensure recorded video is processed by an AI anonymization model that removes bystander faces and vehicle license plates. Detected face and license plate regions are processed with a Gaussian blur such that these personal identification information (PII) regions are obscured. This process helps to ensure that anonymized versions of the video is retained for research purposes. In Project Aria, we have developed a state-of-the-art anonymization system EgoBlur. In this paper, we present extensive analysis of EgoBlur on challenging datasets comparing its performance with other state-of-the-art systems from industry and academia including extensive Responsible AI analysis on recently released Casual Conversations V2 dataset.

Bound by the Bounty: Collaboratively Shaping Evaluation Processes for Queer AI Harms

Organizers of QueerInAI, Nathan Dennler, Anaelia Ovalle +11

Bias evaluation benchmarks and dataset and model documentation have emerged as central processes for assessing the biases and harms of artificial intelligence (AI) systems. However, these auditing processes have been criticized for their failure to integrate the knowledge of marginalized communities and consider the power dynamics between auditors and the communities. Consequently, modes of bias evaluation have been proposed that engage impacted communities in identifying and assessing the harms of AI systems (e.g., bias bounties). Even so, asking what marginalized communities want from such auditing processes has been neglected. In this paper, we ask queer communities for their positions on, and desires from, auditing processes. To this end, we organized a participatory workshop to critique and redesign bias bounties from queer perspectives. We found that when given space, the scope of feedback from workshop participants goes far beyond what bias bounties afford, with participants questioning the ownership, incentives, and efficacy of bounties. We conclude by advocating for community ownership of bounties and complementing bounties with participatory processes (e.g., co-creation).

Evaluating the Impact of Social Determinants on Health Prediction in the Intensive Care Unit

Ming Ying Yang, Gloria Hyunjung Kwak, Tom Pollard +2

Social determinants of health (SDOH) -- the conditions in which people live, grow, and age -- play a crucial role in a person's health and well-being. There is a large, compelling body of evidence in population health studies showing that a wide range of SDOH is strongly correlated with health outcomes. Yet, a majority of the risk prediction models based on electronic health records (EHR) do not incorporate a comprehensive set of SDOH features as they are often noisy or simply unavailable. Our work links a publicly available EHR database, MIMIC-IV, to well-documented SDOH features. We investigate the impact of such features on common EHR prediction tasks across different patient populations. We find that community-level SDOH features do not improve model performance for a general patient population, but can improve data-limited model fairness for specific subpopulations. We also demonstrate that SDOH features are vital for conducting thorough audits of algorithmic biases beyond protective attributes. We hope the new integrated EHR-SDOH database will enable studies on the relationship between community health and individual outcomes and provide new benchmarks to study algorithmic biases beyond race, gender, and age.

Reveal to Revise: An Explainable AI Life Cycle for Iterative Bias Correction of Deep Models

Frederik Pahde, Maximilian Dreyer, Wojciech Samek +1

State-of-the-art machine learning models often learn spurious correlations embedded in the training data. This poses risks when deploying these models for high-stake decision-making, such as in medical applications like skin cancer detection. To tackle this problem, we propose Reveal to Revise (R2R), a framework entailing the entire eXplainable Artificial Intelligence (XAI) life cycle, enabling practitioners to iteratively identify, mitigate, and (re-)evaluate spurious model behavior with a minimal amount of human interaction. In the first step (1), R2R reveals model weaknesses by finding outliers in attributions or through inspection of latent concepts learned by the model. Secondly (2), the responsible artifacts are detected and spatially localized in the input data, which is then leveraged to (3) revise the model behavior. Concretely, we apply the methods of RRR, CDEP and ClArC for model correction, and (4) (re-)evaluate the model's performance and remaining sensitivity towards the artifact. Using two medical benchmark datasets for Melanoma detection and bone age estimation, we apply our R2R framework to VGG, ResNet and EfficientNet architectures and thereby reveal and correct real dataset-intrinsic artifacts, as well as synthetic variants in a controlled setting. Completing the XAI life cycle, we demonstrate multiple R2R iterations to mitigate different biases. Code is available on https://github.com/maxdreyer/Reveal2Revise.

Although AI has significant potential to transform society, there are serious concerns about its ability to behave and make decisions responsibly. Many ethical regulations, principles, and guidelines for responsible AI have been issued recently. However, these principles are high-level and difficult to put into practice. In the meantime much effort has been put into responsible AI from the algorithm perspective, but they are limited to a small subset of ethical principles amenable to mathematical analysis. Responsible AI issues go beyond data and algorithms and are often at the system-level crosscutting many system components and the entire software engineering lifecycle. Based on the result of a systematic literature review, this paper identifies one missing element as the system-level guidance - how to design the architecture of responsible AI systems. We present a summary of design patterns that can be embedded into the AI systems as product features to contribute to responsible-AI-by-design.

Information-Theoretic Bias Reduction via Causal View of Spurious Correlation

Seonguk Seo, Joon-Young Lee, Bohyung Han

We propose an information-theoretic bias measurement technique through a causal interpretation of spurious correlation, which is effective to identify the feature-level algorithmic bias by taking advantage of conditional mutual information. Although several bias measurement methods have been proposed and widely investigated to achieve algorithmic fairness in various tasks such as face recognition, their accuracy- or logit-based metrics are susceptible to leading to trivial prediction score adjustment rather than fundamental bias reduction. Hence, we design a novel debiasing framework against the algorithmic bias, which incorporates a bias regularization loss derived by the proposed information-theoretic bias measurement approach. In addition, we present a simple yet effective unsupervised debiasing technique based on stochastic label noise, which does not require the explicit supervision of bias information. The proposed bias measurement and debiasing approaches are validated in diverse realistic scenarios through extensive experiments on multiple standard benchmarks.

Recent AI-related scandals have shed a spotlight on accountability in AI, with increasing public interest and concern. This paper draws on literature from public policy and governance to make two contributions. First, we propose an AI accountability ecosystem as a useful lens on the system, with different stakeholders requiring and contributing to specific accountability mechanisms. We argue that the present ecosystem is unbalanced, with a need for improved transparency via AI explainability and adequate documentation and process formalisation to support internal audit, leading up eventually to external accreditation processes. Second, we use a case study in the gambling sector to illustrate in a subset of the overall ecosystem the need for industry-specific accountability principles and processes. We define and evaluate critically the implementation of key accountability principles in the gambling industry, namely addressing algorithmic bias and model explainability, before concluding and discussing directions for future work based on our findings. Keywords: Accountability, Explainable AI, Algorithmic Bias, Regulation.

Explaining the predictions of opaque machine learning algorithms is an important and challenging task, especially as complex models are increasingly used to assist in high-stakes decisions such as those arising in healthcare and finance. Most popular tools for post-hoc explainable artificial intelligence (XAI) are either insensitive to context (e.g., feature attributions) or difficult to summarize (e.g., counterfactuals). In this paper, I introduce $\textit{rational Shapley values}$, a novel XAI method that synthesizes and extends these seemingly incompatible approaches in a rigorous, flexible manner. I leverage tools from decision theory and causal modeling to formalize and implement a pragmatic approach that resolves a number of known challenges in XAI. By pairing the distribution of random variables with the appropriate reference class for a given explanation task, I illustrate through theory and experiments how user goals and knowledge can inform and constrain the solution set in an iterative fashion. The method compares favorably to state of the art XAI tools in a range of quantitative and qualitative comparisons.

Understanding and Avoiding AI Failures: A Practical Guide

Heather M. Williams, Roman V. Yampolskiy

As AI technologies increase in capability and ubiquity, AI accidents are becoming more common. Based on normal accident theory, high reliability theory, and open systems theory, we create a framework for understanding the risks associated with AI applications. In addition, we also use AI safety principles to quantify the unique risks of increased intelligence and human-like qualities in AI. Together, these two fields give a more complete picture of the risks of contemporary AI. By focusing on system properties near accidents instead of seeking a root cause of accidents, we identify where attention should be paid to safety for current generation AI systems.