Bias in speech emotion recognition (SER) systems often stems from spurious correlations between speaker characteristics and emotional labels, leading to unfair predictions across demographic groups. Many existing debiasing methods require model-specific changes or demographic annotations, limiting their practical use. We present CO-VADA, a Confidence-Oriented Voice Augmentation Debiasing Approach that mitigates bias without modifying model architecture or relying on demographic information. CO-VADA identifies training samples that reflect bias patterns present in the training data and then applies voice conversion to alter irrelevant attributes and generate samples. These augmented samples introduce speaker variations that differ from dominant patterns in the data, guiding the model to focus more on emotion-relevant features. Our framework is compatible with various SER models and voice conversion tools, making it a scalable and practical solution for improving fairness in SER systems.

Abstract--Federated learning (FL) enables privacy-preserving collaborative model training but remains vulnerable to adversarial behaviors that compromise model utility or fairness across sensitive groups. While extensive studies have examined attacks targeting either objective, strategies that simultaneously degrade both utility and fairness remain largely unexplored. T o bridge this gap, we introduce the Dual-Facet Attack (DF A), a novel threat model that concurrently undermines predictive accuracy and group fairness. Two variants, Synchronous DF A (S-DF A) and Split DF A (Sp-DF A), are further proposed to capture distinct real-world collusion scenarios. Experimental results show that existing robust FL defenses, including hybrid aggregation schemes, fail to resist DF As effectively. T o counter these threats, we propose GuardFed, a self-adaptive defense framework that maintains a fairness-aware reference model using a small amount of clean server data augmented with synthetic samples. In each training round, GuardFed computes a dual-perspective trust score for every client by jointly evaluating its utility deviation and fairness degradation, thereby enabling selective aggregation of trustworthy updates. Extensive experiments on real-world datasets demonstrate that GuardFed consistently preserves both accuracy and fairness under diverse non-IID and adversarial conditions, achieving state-of-the-art performance compared with existing robust FL methods. The rapid advancement of deep learning (DL) has greatly accelerated the deployment of intelligent automation systems [1], providing smart services across diverse application domains. Alongside this evolution, there is an increasing emphasis on human-centered values such as privacy, fairness, and security, which extend beyond traditional performance-oriented objectives. Y anli Li is with the School of Artificial Intelligence and Computer Science, Nantong University, Nantong, 226019, China, and also with the School of Electrical and Computer Engineering, The University of Sydney, Sydney, 2006, Australia (e-mail: yanli.li@sydney.edu.au).

The impact of human label variation (HLV) on model fairness is an unexplored topic. This paper examines the interplay by comparing training on majority-vote labels with a range of HLV methods. Our experiments show that without explicit debiasing, HLV training methods have a positive impact on fairness.

Differential privacy (DP) is a prominent method for protecting information about individuals during data analysis. Training neural networks with differentially private stochastic gradient descent (DPSGD) influences the model's learning dynamics and, consequently, its output. This can affect the model's performance and fairness. While the majority of studies on the topic report a negative impact on fairness, it has recently been suggested that fairness levels comparable to non-private models can be achieved by optimizing hyperparameters for performance directly on differentially private models (rather than re-using hyperparameters from non-private models, as is common practice). In this work, we analyze the generalizabil-ity of this claim by 1) comparing the disparate impact of DPSGD on different performance metrics, and 2) analyzing it over a wide range of hyperparameter settings. We highlight that a disparate impact on one metric does not necessarily imply a disparate impact on another. Most importantly, we show that while optimizing hyperparameters directly on differentially private models does not mitigate the disparate impact of DPSGD reliably, it can still lead to improved utility-fairness trade-offs compared to re-using hyperparameters from non-private models. We stress, however, that any form of hyperparameter tuning entails additional privacy leakage, calling for careful considerations of how to balance privacy, utility and fairness. Finally, we extend our analyses to DPSGD-Global-Adapt, a variant of DPSGD designed to mitigate the disparate impact on accuracy, and conclude that this alternative may not be a robust solution with respect to hyperparameter choice.

Decision trees have been studied extensively in the context of fairness, aiming to maximize prediction performance while ensuring non-discrimination against different groups. Techniques in this space usually focus on imposing constraints at training time, constraining the search space so that solutions which display unacceptable values of relevant metrics are not considered, discarded, or discouraged. If we assume one target variable y and one sensitive attribute s, the design space of tree learning algorithms can be spanned as follows: (i) One can have one tree T that is built using an objective function that is a function of y, s, and T. For instance, one can build a tree based on the weighted information gain regarding y (maximizing) and s (minimizing). (ii) The second option is to have one tree model T that uses an objective function in y and T and a constraint on s and T. Here, s is no longer part of the objective, but part of a constraint. This can be achieved greedily by aborting a further split as soon as the condition that optimizes the objective in y fails to satisfy the constraint on s. A simple way to explore other splits is to backtrack during tree construction once a fairness constraint is violated. (iii) The third option is to have two trees T_y and T_s, one for y and one for s, such that the tree structure for y and s does not have to be shared. In this way, information regarding y and regarding s can be used independently, without having to constrain the choices in tree construction by the mutual information between the two variables. Quite surprisingly, of the three options, only the first one and the greedy variant of the second have been studied in the literature so far. In this paper, we introduce the above two additional options from that design space and characterize them experimentally on multiple datasets.

Large Language Models (LLMs) are widely used for downstream tasks such as tabular classification, where ensuring fairness in their outputs is critical for inclusivity, equal representation, and responsible AI deployment. This study introduces a novel approach to enhancing LLM performance and fairness through the concept of a dynamic validation set, which evolves alongside the test set, replacing the traditional static validation approach. We also propose an iterative algorithm, SMITE, to select optimal in-context examples, with each example set validated against its corresponding dynamic validation set. The in-context set with the lowest total error is used as the final demonstration set. Our experiments across four different LLMs show that our proposed techniques significantly improve both predictive accuracy and fairness compared to baseline methods. To our knowledge, this is the first study to apply dynamic validation in the context of in-context learning for LLMs.

-- Fairness remains a critical concern in healthcare, where unequal access to services and treatment outcomes can adversely affect patient health. While Federated Learning (FL) presents a collaborative and privacy-preserving approach to model training, ensuring fairness is challenging due to heterogeneous data across institutions, and current research primarily addresses non-medical applications. To fill this gap, we establish the first experimental benchmark for fairness in medical FL, evaluating six representative FL methods across diverse demographic attributes and imaging modalities. We introduce FairFedMed, the first medical FL dataset specifically designed to study group fairness (i.e., demographics). It comprises two parts: FairFedMed-Oph, featuring 2D fundus and 3D OCT ophthalmology samples with six demographic attributes; and FairFedMed-Chest, which simulates real cross-institutional FL using subsets of CheXpert and MIMIC-CXR. Together, they support both simulated and real-world FL across diverse medical modalities and demographic groups. Existing FL models often underperform on medical images and overlook fairness across demographic groups. To address this, we propose FairLoRA, a fairness-aware FL framework based on SVD-based low-rank approximation. Experimental results on the FairFedMed dataset demonstrate that FairLoRA not only achieves state-of-the-art performance in medical image classification but also significantly improves fairness across diverse populations. Our code and dataset can be accessible via the link: https: //wang.hms.harvard.edu/fairfedmed/. Minghan Li, Y an Luo, and Mengyu Wang are with Harvard AI and Robotics Lab and Harvard Ophthalmology AI lab, Harvard University, Boston, Massachusetts, USA (e-mail: mili4@meei.harvard.edu;

Electronic Health Record (EHR) data encompass diverse modalities -- text, images, and medical codes -- that are vital for clinical decision-making. To process these complex data, multimodal AI (MAI) has emerged as a powerful approach for fusing such information. However, most existing MAI models optimize for better prediction performance, potentially reinforcing biases across patient subgroups. Although bias-reduction techniques for multimodal models have been proposed, the individual strengths of each modality and their interplay in both reducing bias and optimizing performance remain underexplored. In this work, we introduce FAME (Fairness-Aware Multimodal Embeddings), a framework that explicitly weights each modality according to its fairness contribution. FAME optimizes both performance and fairness by incorporating a combined loss function. We leverage the Error Distribution Disparity Index (EDDI) to measure fairness across subgroups and propose a sign-agnostic aggregation method to balance fairness across subgroups, ensuring equitable model outcomes. We evaluate FAME with BEHRT and BioClinicalBERT, combining structured and unstructured EHR data, and demonstrate its effectiveness in terms of performance and fairness compared with other baselines across multiple EHR prediction tasks.

Multimodal learning, which integrates diverse data sources such as images, text, and structured data, has proven superior to unimodal counterparts in high-stakes decision-making. However, while performance gains remain the gold standard for evaluating multimodal systems, concerns around bias and robustness are frequently overlooked. In this context, this paper explores two key research questions (RQs): (i) RQ1 examines whether adding a modality con-sistently enhances performance and investigates its role in shaping fairness measures, assessing whether it mitigates or amplifies bias in multimodal models; (ii) RQ2 investigates the impact of missing modalities at inference time, analyzing how multimodal models generalize in terms of both performance and fairness. Our analysis reveals that incorporating new modalities during training consistently enhances the performance of multimodal models, while fairness trends exhibit variability across different evaluation measures and datasets. Additionally, the absence of modalities at inference degrades performance and fairness, raising concerns about its robustness in real-world deployment. We conduct extensive experiments using multimodal healthcare datasets containing images, time series, and structured information to validate our findings.

Recent methods leverage a hypernet to handle the performance-fairness trade-offs in federated learning. This hypernet maps the clients' preferences between model performance and fairness to preference-specifc models on the trade-off curve, known as local Pareto front. However, existing methods typically adopt a uniform preference sampling distribution to train the hypernet across clients, neglecting the inherent heterogeneity of their local Pareto fronts. Meanwhile, from the perspective of generalization, they do not consider the gap between local and global Pareto fronts on the global dataset. To address these limitations, we propose HetPFL to effectively learn both local and global Pareto fronts. HetPFL comprises Preference Sampling Adaptation (PSA) and Preference-aware Hypernet Fusion (PHF). PSA adaptively determines the optimal preference sampling distribution for each client to accommodate heterogeneous local Pareto fronts. While PHF performs preference-aware fusion of clients' hypernets to ensure the performance of the global Pareto front. We prove that HetPFL converges linearly with respect to the number of rounds, under weaker assumptions than existing methods. Extensive experiments on four datasets show that HetPFL significantly outperforms seven baselines in terms of the quality of learned local and global Pareto fronts.