Research Papers

Many instances of algorithmic bias are caused by subpopulation shifts. For example, ML models often perform worse on demographic groups that are underrepresented in the training data. In this paper, we study whether enforcing algorithmic fairness during training improves the performance of the trained model in the \emph{target domain}. On one hand, we conceive scenarios in which enforcing fairness does not improve performance in the target domain. In fact, it may even harm performance. On the other hand, we derive necessary and sufficient conditions under which enforcing algorithmic fairness leads to the Bayes model in the target domain. We also illustrate the practical implications of our theoretical results in simulations and on real data.

Algorithms that "Don't See Color": Comparing Biases in Lookalike and Special Ad Audiences

Piotr Sapiezynski, Avijit Ghosh, Levi Kaplan +2

Researchers and journalists have repeatedly shown that algorithms commonly used in domains such as credit, employment, healthcare, or criminal justice can have discriminatory effects. Some organizations have tried to mitigate these effects by simply removing sensitive features from an algorithm's inputs. In this paper, we explore the limits of this approach using a unique opportunity. In 2019, Facebook agreed to settle a lawsuit by removing certain sensitive features from inputs of an algorithm that identifies users similar to those provided by an advertiser for ad targeting, making both the modified and unmodified versions of the algorithm available to advertisers. We develop methodologies to measure biases along the lines of gender, age, and race in the audiences created by this modified algorithm, relative to the unmodified one. Our results provide experimental proof that merely removing demographic features from a real-world algorithmic system's inputs can fail to prevent biased outputs. As a result, organizations using algorithms to help mediate access to important life opportunities should consider other approaches to mitigating discriminatory effects.

Artificial Intelligence (AI) has been used extensively in automatic decision making in a broad variety of scenarios, ranging from credit ratings for loans to recommendations of movies. Traditional design guidelines for AI models focus essentially on accuracy maximization, but recent work has shown that economically irrational and socially unacceptable scenarios of discrimination and unfairness are likely to arise unless these issues are explicitly addressed. This undesirable behavior has several possible sources, such as biased datasets used for training that may not be detected in black-box models. After pointing out connections between such bias of AI and the problem of induction, we focus on Popper's contributions after Hume's, which offer a logical theory of preferences. An AI model can be preferred over others on purely rational grounds after one or more attempts at refutation based on accuracy and fairness. Inspired by such epistemological principles, this paper proposes a structured approach to mitigate discrimination and unfairness caused by bias in AI systems. In the proposed computational framework, models are selected and enhanced after attempts at refutation. To illustrate our discussion, we focus on hiring decision scenarios where an AI system filters in which job applicants should go to the interview phase.

Artificial intelligence (AI) models trained using medical images for clinical tasks often exhibit bias in the form of disparities in performance between subgroups. Since not all sources of biases in real-world medical imaging data are easily identifiable, it is challenging to comprehensively assess how those biases are encoded in models, and how capable bias mitigation methods are at ameliorating performance disparities. In this article, we introduce a novel analysis framework for systematically and objectively investigating the impact of biases in medical images on AI models. We developed and tested this framework for conducting controlled in silico trials to assess bias in medical imaging AI using a tool for generating synthetic magnetic resonance images with known disease effects and sources of bias. The feasibility is showcased by using three counterfactual bias scenarios to measure the impact of simulated bias effects on a convolutional neural network (CNN) classifier and the efficacy of three bias mitigation strategies. The analysis revealed that the simulated biases resulted in expected subgroup performance disparities when the CNN was trained on the synthetic datasets. Moreover, reweighing was identified as the most successful bias mitigation strategy for this setup, and we demonstrated how explainable AI methods can aid in investigating the manifestation of bias in the model using this framework. Developing fair AI models is a considerable challenge given that many and often unknown sources of biases can be present in medical imaging datasets. In this work, we present a novel methodology to objectively study the impact of biases and mitigation strategies on deep learning pipelines, which can support the development of clinical AI that is robust and responsible.

Load shapes derived from smart meter data are frequently employed to analyze daily energy consumption patterns, particularly in the context of applications like Demand Response (DR). Nevertheless, one of the most important challenges to this endeavor lies in identifying the most suitable consumer clusters with similar consumption behaviors. In this paper, we present a novel machine learning based framework in order to achieve optimal load profiling through a real case study, utilizing data from almost 5000 households in London. Four widely used clustering algorithms are applied specifically K-means, K-medoids, Hierarchical Agglomerative Clustering and Density-based Spatial Clustering. An empirical analysis as well as multiple evaluation metrics are leveraged to assess those algorithms. Following that, we redefine the problem as a probabilistic classification one, with the classifier emulating the behavior of a clustering algorithm,leveraging Explainable AI (xAI) to enhance the interpretability of our solution. According to the clustering algorithm analysis the optimal number of clusters for this case is seven. Despite that, our methodology shows that two of the clusters, almost 10\% of the dataset, exhibit significant internal dissimilarity and thus it splits them even further to create nine clusters in total. The scalability and versatility of our solution makes it an ideal choice for power utility companies aiming to segment their users for creating more targeted Demand Response programs.

Machine learning offers great potential for automated prediction of post-stroke symptoms and their response to rehabilitation. Major challenges for this endeavour include the very high dimensionality of neuroimaging data, the relatively small size of the datasets available for learning, and how to effectively combine neuroimaging and tabular data (e.g. demographic information and clinical characteristics). This paper evaluates several solutions based on two strategies. The first is to use 2D images that summarise MRI scans. The second is to select key features that improve classification accuracy. Additionally, we introduce the novel approach of training a convolutional neural network (CNN) on images that combine regions-of-interest extracted from MRIs, with symbolic representations of tabular data. We evaluate a series of CNN architectures (both 2D and a 3D) that are trained on different representations of MRI and tabular data, to predict whether a composite measure of post-stroke spoken picture description ability is in the aphasic or non-aphasic range. MRI and tabular data were acquired from 758 English speaking stroke survivors who participated in the PLORAS study. The classification accuracy for a baseline logistic regression was 0.678 for lesion size alone, rising to 0.757 and 0.813 when initial symptom severity and recovery time were successively added. The highest classification accuracy 0.854 was observed when 8 regions-of-interest was extracted from each MRI scan and combined with lesion size, initial severity and recovery time in a 2D Residual Neural Network.Our findings demonstrate how imaging and tabular data can be combined for high post-stroke classification accuracy, even when the dataset is small in machine learning terms. We conclude by proposing how the current models could be improved to achieve even higher levels of accuracy using images from hospital scanners.

Generative multimodal models based on diffusion models have seen tremendous growth and advances in recent years. Models such as DALL-E and Stable Diffusion have become increasingly popular and successful at creating images from texts, often combining abstract ideas. However, like other deep learning models, they also reflect social biases they inherit from their training data, which is often crawled from the internet. Manually auditing models for biases can be very time and resource consuming and is further complicated by the unbounded and unconstrained nature of inputs these models can take. Research into bias measurement and quantification has generally focused on small single-stage models working on a single modality. Thus the emergence of multistage multimodal models requires a different approach. In this paper, we propose Multimodal Composite Association Score (MCAS) as a new method of measuring gender bias in multimodal generative models. Evaluating both DALL-E 2 and Stable Diffusion using this approach uncovered the presence of gendered associations of concepts embedded within the models. We propose MCAS as an accessible and scalable method of quantifying potential bias for models with different modalities and a range of potential biases.

Data quality is the key factor for the development of trustworthy AI in healthcare. A large volume of curated datasets with controlled confounding factors can help improve the accuracy, robustness and privacy of downstream AI algorithms. However, access to good quality datasets is limited by the technical difficulty of data acquisition and large-scale sharing of healthcare data is hindered by strict ethical restrictions. Data synthesis algorithms, which generate data with a similar distribution as real clinical data, can serve as a potential solution to address the scarcity of good quality data during the development of trustworthy AI. However, state-of-the-art data synthesis algorithms, especially deep learning algorithms, focus more on imaging data while neglecting the synthesis of non-imaging healthcare data, including clinical measurements, medical signals and waveforms, and electronic healthcare records (EHRs). Thus, in this paper, we will review the synthesis algorithms, particularly for non-imaging medical data, with the aim of providing trustworthy AI in this domain. This tutorial-styled review paper will provide comprehensive descriptions of non-imaging medical data synthesis on aspects including algorithms, evaluations, limitations and future research directions.

Learning Occupational Task-Shares Dynamics for the Future of Work

Subhro Das, Sebastian Steffen, Wyatt Clarke +3

The recent wave of AI and automation has been argued to differ from previous General Purpose Technologies (GPTs), in that it may lead to rapid change in occupations' underlying task requirements and persistent technological unemployment. In this paper, we apply a novel methodology of dynamic task shares to a large dataset of online job postings to explore how exactly occupational task demands have changed over the past decade of AI innovation, especially across high, mid and low wage occupations. Notably, big data and AI have risen significantly among high wage occupations since 2012 and 2016, respectively. We built an ARIMA model to predict future occupational task demands and showcase several relevant examples in Healthcare, Administration, and IT. Such task demands predictions across occupations will play a pivotal role in retraining the workforce of the future.

Fair Allocation through Selective Information Acquisition

William Cai, Johann Gaebler, Nikhil Garg +1

Public and private institutions must often allocate scare resources under uncertainty. Banks, for example, extend credit to loan applicants based in part on their estimated likelihood of repaying a loan. But when the quality of information differs across candidates (e.g., if some applicants lack traditional credit histories), common lending strategies can lead to disparities across groups. Here we consider a setting in which decision makers -- before allocating resources -- can choose to spend some of their limited budget further screening select individuals. We present a computationally efficient algorithm for deciding whom to screen that maximizes a standard measure of social welfare. Intuitively, decision makers should screen candidates on the margin, for whom the additional information could plausibly alter the allocation. We formalize this idea by showing the problem can be reduced to solving a series of linear programs. Both on synthetic and real-world datasets, this strategy improves utility, illustrating the value of targeted information acquisition in such decisions. Further, when there is social value for distributing resources to groups for whom we have a priori poor information -- like those without credit scores -- our approach can substantially improve the allocation of limited assets.

Explainable machine learning (ML) enables human learning from ML, human appeal of automated model decisions, regulatory compliance, and security audits of ML models. Explainable ML (i.e. explainable artificial intelligence or XAI) has been implemented in numerous open source and commercial packages and explainable ML is also an important, mandatory, or embedded aspect of commercial predictive modeling in industries like financial services. However, like many technologies, explainable ML can be misused, particularly as a faulty safeguard for harmful black-boxes, e.g. fairwashing or scaffolding, and for other malevolent purposes like stealing models and sensitive training data. To promote best-practice discussions for this already in-flight technology, this short text presents internal definitions and a few examples before covering the proposed guidelines. This text concludes with a seemingly natural argument for the use of interpretable models and explanatory, debugging, and disparate impact testing methods in life- or mission-critical ML systems.

Large Language Models (LLMs) are being adopted across a wide range of tasks, including decision-making processes in industries where bias in AI systems is a significant concern. Recent research indicates that LLMs can harbor implicit biases even when they pass explicit bias evaluations. Building upon the frameworks of the LLM Implicit Association Test (IAT) Bias and LLM Decision Bias, this study highlights that newer or larger language models do not automatically exhibit reduced bias; in some cases, they displayed higher bias scores than their predecessors, such as in Meta's Llama series and OpenAI's GPT models. This suggests that increasing model complexity without deliberate bias mitigation strategies can unintentionally amplify existing biases. The variability in bias scores within and across providers underscores the need for standardized evaluation metrics and benchmarks for bias assessment. The lack of consistency indicates that bias mitigation is not yet a universally prioritized goal in model development, which can lead to unfair or discriminatory outcomes. By broadening the detection of implicit bias, this research provides a more comprehensive understanding of the biases present in advanced models and underscores the critical importance of addressing these issues to ensure the development of fair and responsible AI systems.