Freezing of gait (FOG) is a debilitating feature of Parkinson's disease (PD), which is a cause of injurious falls among PD patients. Recent advances in wearable-based human activity recognition (HAR) technology have enabled the detection of FOG subtypes across benchmark datasets. Since FOG manifestation is heterogeneous, developing models that quantify FOG consistently across patients with varying demographics, FOG types, and PD conditions is important. Bias and fairness in FOG models remain understudied in HAR, with research focused mainly on FOG detection using single benchmark datasets. We evaluated the bias and fairness of HAR models for wearable-based FOG detection across demographics and PD conditions using multiple datasets and the effectiveness of transfer learning as a potential bias mitigation approach. Our evaluation using demographic parity ratio (DPR) and equalized odds ratio (EOR) showed model bias (DPR & EOR < 0.8) for all stratified demographic variables, including age, sex, and disease duration. Our experiments demonstrated that transfer learning from multi-site datasets and generic human activity representations significantly improved fairness (average change in DPR +0.027, +0.039, respectively) and performance (average change in F1-score +0.026, +0.018, respectively) across attributes, supporting the hypothesis that generic human activity representations learn fairer representations applicable to health analytics.

The rapid advancement of Language Model technologies has opened new opportunities, but also introduced new challenges related to bias and fairness. This paper explores the uncharted territory of potential biases in state-of-the-art universal text embedding models towards specific document and query writing styles within Information Retrieval (IR) systems. Our investigation reveals that different embedding models exhibit different preferences of document writing style, while more informal and emotive styles are less favored by most embedding models. In terms of query writing styles, many embedding models tend to match the style of the query with the style of the retrieved documents, but some show a consistent preference for specific styles. Text embedding models fine-tuned on synthetic data generated by LLMs display a consistent preference for certain style of generated data. These biases in text embedding based IR systems can inadvertently silence or marginalize certain communication styles, thereby posing a significant threat to fairness in information retrieval. Finally, we also compare the answer styles of Retrieval Augmented Generation (RAG) systems based on different LLMs and find out that most text embedding models are biased towards LLM's answer styles when used as evaluation metrics for answer correctness. This study sheds light on the critical issue of writing style based bias in IR systems, offering valuable insights for the development of more fair and robust models.

Efforts to mitigate bias and enhance fairness in the artificial intelligence (AI) community have predominantly focused on technical solutions. While numerous reviews have addressed bias in AI, this review uniquely focuses on the practical limitations of technical solutions in healthcare settings, providing a structured analysis across five key dimensions affecting their real-world implementation: who defines bias and fairness; which mitigation strategy to use and prioritize among dozens that are inconsistent and incompatible; when in the AI development stages the solutions are most effective; for which populations; and the context in which the solutions are designed. We illustrate each limitation with empirical studies focusing on healthcare and biomedical applications. Moreover, we discuss how value-sensitive AI, a framework derived from technology design, can engage stakeholders and ensure that their values are embodied in bias and fairness mitigation solutions. Finally, we discuss areas that require further investigation and provide practical recommendations to address the limitations covered in the study.

Large language models (LLMs) can exhibit bias in a variety of ways. Such biases can create or exacerbate unfair outcomes for certain groups within a protected attribute, including, but not limited to sex, race, sexual orientation, or age. This paper aims to provide a technical guide for practitioners to assess bias and fairness risks in LLM use cases. The main contribution of this work is a decision framework that allows practitioners to determine which metrics to use for a specific LLM use case. To achieve this, this study categorizes LLM bias and fairness risks, maps those risks to a taxonomy of LLM use cases, and then formally defines various metrics to assess each type of risk. As part of this work, several new bias and fairness metrics are introduced, including innovative counterfactual metrics as well as metrics based on stereotype classifiers. Instead of focusing solely on the model itself, the sensitivity of both prompt-risk and model-risk are taken into account by defining evaluations at the level of an LLM use case, characterized by a model and a population of prompts. Furthermore, because all of the evaluation metrics are calculated solely using the LLM output, the proposed framework is highly practical and easily actionable for practitioners.

In recent years, the affective computing (AC) and human-robot interaction (HRI) research communities have put fairness at the centre of their research agenda. However, none of the existing work has addressed the problem of machine learning (ML) bias in HRI settings. In addition, many of the current datasets for AC and HRI are "small", making ML bias and debias analysis challenging. This paper presents the first work to explore ML bias analysis and mitigation of three small multimodal datasets collected within both a human-human and robot-human wellbeing coaching settings. The contributions of this work includes: i) being the first to explore the problem of ML bias and fairness within HRI settings; and ii) providing a multimodal analysis evaluated via modelling performance and fairness metrics across both high and low-level features and proposing a simple and effective data augmentation strategy (MixFeat) to debias the small datasets presented within this paper; and iii) conducting extensive experimentation and analyses to reveal ML fairness insights unique to AC and HRI research in order to distill a set of recommendations to aid AC and HRI researchers to be more engaged with fairness-aware ML-based research.

Chatbots have been studied for more than half a century. With the rapid development of natural language processing (NLP) technologies in recent years, chatbots using large language models (LLMs) have received much attention nowadays. Compared with traditional ones, modern chatbots are more powerful and have been used in real-world applications. There are however, bias and fairness concerns in modern chatbot design. Due to the huge amounts of training data, extremely large model sizes, and lack of interpretability, bias mitigation and fairness preservation of modern chatbots are challenging. Thus, a comprehensive overview on bias and fairness in chatbot systems is given in this paper. The history of chatbots and their categories are first reviewed. Then, bias sources and potential harms in applications are analyzed. Considerations in designing fair and unbiased chatbot systems are examined. Finally, future research directions are discussed.

Pretrained machine learning models are known to perpetuate and even amplify existing biases in data, which can result in unfair outcomes that ultimately impact user experience. Therefore, it is crucial to understand the mechanisms behind those prejudicial biases to ensure that model performance does not result in discriminatory behaviour toward certain groups or populations. In this work, we define gender bias as our case study. We quantify bias amplification in pretraining and after fine-tuning on three families of vision-and-language models. We investigate the connection, if any, between the two learning stages, and evaluate how bias amplification reflects on model performance. Overall, we find that bias amplification in pretraining and after fine-tuning are independent. We then examine the effect of continued pretraining on gender-neutral data, finding that this reduces group disparities, i.e., promotes fairness, on VQAv2 and retrieval tasks without significantly compromising task performance.

Rapid advancements of large language models (LLMs) have enabled the processing, understanding, and generation of human-like text, with increasing integration into systems that touch our social sphere. Despite this success, these models can learn, perpetuate, and amplify harmful social biases. In this paper, we present a comprehensive survey of bias evaluation and mitigation techniques for LLMs. We first consolidate, formalize, and expand notions of social bias and fairness in natural language processing, defining distinct facets of harm and introducing several desiderata to operationalize fairness for LLMs. We then unify the literature by proposing three intuitive taxonomies, two for bias evaluation, namely metrics and datasets, and one for mitigation. Our first taxonomy of metrics for bias evaluation disambiguates the relationship between metrics and evaluation datasets, and organizes metrics by the different levels at which they operate in a model: embeddings, probabilities, and generated text. Our second taxonomy of datasets for bias evaluation categorizes datasets by their structure as counterfactual inputs or prompts, and identifies the targeted harms and social groups; we also release a consolidation of publicly-available datasets for improved access. Our third taxonomy of techniques for bias mitigation classifies methods by their intervention during pre-processing, in-training, intra-processing, and post-processing, with granular subcategories that elucidate research trends. Finally, we identify open problems and challenges for future work. Synthesizing a wide range of recent research, we aim to provide a clear guide of the existing literature that empowers researchers and practitioners to better understand and prevent the propagation of bias in LLMs.

In this paper, we trace the biases in current natural language processing (NLP) models back to their origins in racism, sexism, and homophobia over the last 500 years. We review literature from critical race theory, gender studies, data ethics, and digital humanities studies, and summarize the origins of bias in NLP models from these social science perspective. We show how the causes of the biases in the NLP pipeline are rooted in social issues. Finally, we argue that the only way to fix the bias and unfairness in NLP is by addressing the social problems that caused them in the first place and by incorporating social sciences and social scientists in efforts to mitigate bias in NLP models. We provide actionable recommendations for the NLP research community to do so.

We provide a psychometric-grounded exposition of bias and fairness as applied to a typical machine learning pipeline for affective computing. We expand on an interpersonal communication framework to elucidate how to identify sources of bias that may arise in the process of inferring human emotions and other psychological constructs from observed behavior. Various methods and metrics for measuring fairness and bias are discussed along with pertinent implications within the United States legal context. We illustrate how to measure some types of bias and fairness in a case study involving automatic personality and hireability inference from multimodal data collected in video interviews for mock job applications. We encourage affective computing researchers and practitioners to encapsulate bias and fairness in their research processes and products and to consider their role, agency, and responsibility in promoting equitable and just systems. Personal use of this material is permitted. The tools used in affective computing (AC), which enable machines to identify people's behaviors and mental states, are being increasingly utilized in education, healthcare, and the workplace. One application is to aid in the allocation of limited resources (e.g., counseling, mental health care, in-person interviews) via automated screening [1-3]. In these types of high-stakes scenarios, the assessments provided by AC systems can directly affect the decision processes which influence the amount of attention, care, and opportunities afforded to individuals. As such, it is important that these processes are accurate, unbiased, and fair because any deficiencies or errors present in these systems stemming from the data they were trained on, the types of algorithms used, or the decision processes themselves, may disproportionately impact different groups of people and lead to ethical and legal concerns, not to mention pain and suffering for the vulnerable groups impacted. Simply put, AC systems must deter, not propagate, extant systems of inequity and injustice. Fortunately, we have decades of guidance on how to construct fair and unbiased measurement systems.