Research Papers

As AI systems, particularly generative models, increasingly influence decision-making, ensuring that they are able to fairly represent diverse human preferences becomes crucial. This paper introduces a novel framework for evaluating epistemic fairness in preference learning models inspired by economic theories of inequality and Rawlsian justice. We propose metrics adapted from the Gini Coefficient, Atkinson Index, and Kuznets Ratio to quantify fairness in these models. We validate our approach using two datasets: a custom visual preference dataset (AI-EDI-Space) and the Jester Jokes dataset. Our analysis reveals variations in model performance across users, highlighting potential epistemic injustices. We explore pre-processing and in-processing techniques to mitigate these inequalities, demonstrating a complex relationship between model efficiency and fairness. This work contributes to AI ethics by providing a framework for evaluating and improving epistemic fairness in preference learning models, offering insights for developing more inclusive AI systems in contexts where diverse human preferences are crucial.

As Artificial Intelligence applications expand, the evaluation of models faces heightened scrutiny. Ensuring public readiness requires evaluation datasets, which differ from training data by being disjoint and ethically sourced in compliance with privacy regulations. The performance and fairness of face recognition systems depend significantly on the quality and representativeness of these evaluation datasets. This data is sometimes scraped from the internet without user's consent, causing ethical concerns that can prohibit its use without proper releases. In rare cases, data is collected in a controlled environment with consent, however, this process is time-consuming, expensive, and logistically difficult to execute. This creates a barrier for those unable to conjure the immense resources required to gather ethically sourced evaluation datasets. To address these challenges, we introduce the Synthetic Identity Generation pipeline, or SIG, that allows for the targeted creation of ethical, balanced datasets for face recognition evaluation. Our proposed and demonstrated pipeline generates high-quality images of synthetic identities with controllable pose, facial features, and demographic attributes, such as race, gender, and age. We also release an open-source evaluation dataset named ControlFace10k, consisting of 10,008 face images of 3,336 unique synthetic identities balanced across race, gender, and age, generated using the proposed SIG pipeline. We analyze ControlFace10k along with a non-synthetic BUPT dataset using state-of-the-art face recognition algorithms to demonstrate its effectiveness as an evaluation tool. This analysis highlights the dataset's characteristics and its utility in assessing algorithmic bias across different demographic groups.

Federated Learning (FL) in Deep Learning (DL)-automated medical image segmentation helps preserving privacy by enabling collaborative model training without sharing patient data. However, FL faces challenges with data heterogeneity among institutions, leading to suboptimal global models. Integrating Disentangled Representation Learning (DRL) in FL can enhance robustness by separating data into distinct representations. Existing DRL methods assume heterogeneity lies solely in style features, overlooking content-based variability like lesion size and shape. We propose FedGS, a novel FL aggregation method, to improve segmentation performance on small, under-represented targets while maintaining overall efficacy. FedGS demonstrates superior performance over FedAvg, particularly for small lesions, across PolypGen and LiTS datasets. The code and pre-trained checkpoints are available at the following link: https://github.com/Trustworthy-AI-UU-NKI/Federated-Learning-Disentanglement

In Explainable AI (XAI), counterfactual explanations (CEs) are a well-studied method to communicate feature relevance through contrastive reasoning of "what if" to explain AI models' predictions. However, they only focus on important (i.e., relevant) features and largely disregard less important (i.e., irrelevant) ones. Such irrelevant features can be crucial in many applications, especially when users need to ensure that an AI model's decisions are not affected or biased against specific attributes such as gender, race, religion, or political affiliation. To address this gap, the concept of alterfactual explanations (AEs) has been proposed. AEs explore an alternative reality of "no matter what", where irrelevant features are substituted with alternative features (e.g., "republicans" -> "democrats") within the same attribute (e.g., "politics") while maintaining a similar prediction output. This serves to validate whether AI model predictions are influenced by the specified attributes. Despite the promise of AEs, there is a lack of computational approaches to systematically generate them, particularly in the text domain, where creating AEs for AI text classifiers presents unique challenges. This paper addresses this challenge by formulating AE generation as an optimization problem and introducing MoMatterXAI, a novel algorithm that generates AEs for text classification tasks. Our approach achieves high fidelity of up to 95% while preserving context similarity of over 90% across multiple models and datasets. A human study further validates the effectiveness of AEs in explaining AI text classifiers to end users. All codes will be publicly available.

Recently, the research community of computerized medical imaging has started to discuss and address potential fairness issues that may emerge when developing and deploying AI systems for medical image analysis. This chapter covers some of the pressing challenges encountered when doing research in this area, and it is intended to raise questions and provide food for thought for those aiming to enter this research field. The chapter first discusses various sources of bias, including data collection, model training, and clinical deployment, and their impact on the fairness of machine learning algorithms in medical image computing. We then turn to discussing open challenges that we believe require attention from researchers and practitioners, as well as potential pitfalls of naive application of common methods in the field. We cover a variety of topics including the impact of biased metrics when auditing for fairness, the leveling down effect, task difficulty variations among subgroups, discovering biases in unseen populations, and explaining biases beyond standard demographic attributes.

In education data mining (EDM) communities, machine learning has achieved remarkable success in discovering patterns and structures to tackle educational challenges. Notably, fairness and algorithmic bias have gained attention in learning analytics of EDM. With the increasing demand for the right to be forgotten, there is a growing need for machine learning models to forget sensitive data and its impact, particularly within the realm of EDM. The paradigm of selective forgetting, also known as machine unlearning, has been extensively studied to address this need by eliminating the influence of specific data from a pre-trained model without complete retraining. However, existing research assumes that interactive data removal operations are conducted in secure and reliable environments, neglecting potential malicious unlearning requests to undermine the fairness of machine learning systems. In this paper, we introduce a novel class of selective forgetting attacks designed to compromise the fairness of learning models while maintaining their predictive accuracy, thereby preventing the model owner from detecting the degradation in model performance. Additionally, we propose an innovative optimization framework for selective forgetting attacks, capable of generating malicious unlearning requests across various attack scenarios. We validate the effectiveness of our proposed selective forgetting attacks on fairness through extensive experiments using diverse EDM datasets.

Evaluating Algorithmic Bias in Models for Predicting Academic Performance of Filipino Students

Valdemar Švábenský, Mélina Verger, Maria Mercedes T. Rodrigo +5

Algorithmic bias is a major issue in machine learning models in educational contexts. However, it has not yet been studied thoroughly in Asian learning contexts, and only limited work has considered algorithmic bias based on regional (sub-national) background. As a step towards addressing this gap, this paper examines the population of 5,986 students at a large university in the Philippines, investigating algorithmic bias based on students' regional background. The university used the Canvas learning management system (LMS) in its online courses across a broad range of domains. Over the period of three semesters, we collected 48.7 million log records of the students' activity in Canvas. We used these logs to train binary classification models that predict student grades from the LMS activity. The best-performing model reached AUC of 0.75 and weighted F1-score of 0.79. Subsequently, we examined the data for bias based on students' region. Evaluation using three metrics: AUC, weighted F1-score, and MADD showed consistent results across all demographic groups. Thus, no unfairness was observed against a particular student group in the grade predictions.

The worldwide adoption of machine learning (ML) and deep learning models, particularly in critical sectors, such as healthcare and finance, presents substantial challenges in maintaining individual privacy and fairness. These two elements are vital to a trustworthy environment for learning systems. While numerous studies have concentrated on protecting individual privacy through differential privacy (DP) mechanisms, emerging research indicates that differential privacy in machine learning models can unequally impact separate demographic subgroups regarding prediction accuracy. This leads to a fairness concern, and manifests as biased performance. Although the prevailing view is that enhancing privacy intensifies fairness disparities, a smaller, yet significant, subset of research suggests the opposite view. In this article, with extensive evaluation results, we demonstrate that the impact of differential privacy on fairness is not monotonous. Instead, we observe that the accuracy disparity initially grows as more DP noise (enhanced privacy) is added to the ML process, but subsequently diminishes at higher privacy levels with even more noise. Moreover, implementing gradient clipping in the differentially private stochastic gradient descent ML method can mitigate the negative impact of DP noise on fairness. This mitigation is achieved by moderating the disparity growth through a lower clipping threshold.

Recently, watermarking schemes for large language models (LLMs) have been proposed to distinguish text generated by machines and by humans. The present paper explores philosophical, political, and ethical ramifications of implementing and using watermarking schemes. A definition of authorship that includes both machines (LLMs) and humans is proposed to serve as a backdrop. It is argued that private watermarks may provide private companies with sweeping rights to determine authorship, which is incompatible with traditional standards of authorship determination. Then, possible ramifications of the so-called entropy dependence of watermarking mechanisms are explored. It is argued that entropy may vary for different, socially salient groups. This could lead to group dependent rates at which machine generated text is detected. Specifically, groups more interested in low entropy text may face the challenge that it is harder to detect machine generated text that is of interest to them.

In the current machine translation (MT) landscape, the Transformer architecture stands out as the gold standard, especially for high-resource language pairs. This research delves into its efficacy for low-resource language pairs including both the English$\leftrightarrow$Irish and English$\leftrightarrow$Marathi language pairs. Notably, the study identifies the optimal hyperparameters and subword model type to significantly improve the translation quality of Transformer models for low-resource language pairs. The scarcity of parallel datasets for low-resource languages can hinder MT development. To address this, gaHealth was developed, the first bilingual corpus of health data for the Irish language. Focusing on the health domain, models developed using this in-domain dataset exhibited very significant improvements in BLEU score when compared with models from the LoResMT2021 Shared Task. A subsequent human evaluation using the multidimensional quality metrics error taxonomy showcased the superior performance of the Transformer system in reducing both accuracy and fluency errors compared to an RNN-based counterpart. Furthermore, this thesis introduces adaptNMT and adaptMLLM, two open-source applications streamlined for the development, fine-tuning, and deployment of neural machine translation models. These tools considerably simplify the setup and evaluation process, making MT more accessible to both developers and translators. Notably, adaptNMT, grounded in the OpenNMT ecosystem, promotes eco-friendly natural language processing research by highlighting the environmental footprint of model development. Fine-tuning of MLLMs by adaptMLLM demonstrated advancements in translation performance for two low-resource language pairs: English$\leftrightarrow$Irish and English$\leftrightarrow$Marathi, compared to baselines from the LoResMT2021 Shared Task.

Generative artificial intelligence (AI) is poised to reshape the way individuals communicate and interact. While this form of AI has the potential to efficiently make numerous human decisions, there is limited understanding of how individuals respond to its use in social interaction. In particular, it remains unclear how individuals engage with algorithms when the interaction entails consequences for other people. Here, we report the results of a large-scale pre-registered online experiment (N = 3,552) indicating diminished fairness, trust, trustworthiness, cooperation, and coordination by human players in economic twoplayer games, when the decision of the interaction partner is taken over by ChatGPT. On the contrary, we observe no adverse welfare effects when individuals are uncertain about whether they are interacting with a human or generative AI. Therefore, the promotion of AI transparency, often suggested as a solution to mitigate the negative impacts of generative AI on society, shows a detrimental effect on welfare in our study. Concurrently, participants frequently delegate decisions to ChatGPT, particularly when the AI's involvement is undisclosed, and individuals struggle to discern between AI and human decisions.

The rising use of Artificial Intelligence (AI) in human detection on Edge camera systems has led to accurate but complex models, challenging to interpret and debug. Our research presents a diagnostic method using Explainable AI (XAI) for model debugging, with expert-driven problem identification and solution creation. Validated on the Bytetrack model in a real-world office Edge network, we found the training dataset as the main bias source and suggested model augmentation as a solution. Our approach helps identify model biases, essential for achieving fair and trustworthy models.