Abstract Bias in machine learning models remains a critical challenge, particularly in datasets with numeric features where discrimination may be subtle and hard to detect. Existing fairness frameworks rely on expert knowledge of marginalized groups, such as specific racial groups, and categorical features defining them. Furthermore, most frameworks evaluate bias in models rather than datasets, despite the fact that model bias can often be traced back to dataset shortcomings. Our research aims to remedy this gap by capturing dataset flaws in a set of meta-features at the dataset level, and to warn practitioners of bias risk when using such datasets for model training. We neither restrict the feature type nor expect domain knowledge. To this end, we develop methods to synthesize biased datasets and extend current fairness metrics to continuous features in order to quantify dataset-level discrimination risks. Our approach constructs a meta-database of diverse datasets, from which we derive transferable meta-features that capture dataset properties indicative of bias risk. Our findings demonstrate that dataset-level characteristics can serve as cost-effective indicators of bias risk, providing a novel method for data auditing that does not rely on expert knowledge. This work lays the foundation for early-warning systems, moving beyond model-focused assessments toward a data-centric approach.

Abstract In Natural Language Processing (NLP), pre-trained language models (LLMs) are widely employed and refined for various tasks. These models have shown considerable social and geographic biases creating skewed or even unfair representations of certain groups. Research focuses on biases toward L2 (English as a second language) regions but neglects bias within L1 (first language) regions. In this work, we ask if there is regional bias within L1 regions already inherent in pre-trained LLMs and, if so, what the consequences are in terms of downstream model performance. We contribute an investigation framework specifically tailored for low-resource regions, offering a method to identify bias without imposing strict requirements for labeled datasets. Our research reveals subtle geographic variations in the word embeddings of BERT, even in cultures traditionally perceived as similar. These nuanced features, once captured, have the potential to significantly impact downstream tasks. Generally, models exhibit comparable performance on datasets that share similarities, and conversely, performance may diverge when datasets differ in their nuanced features embedded within the language. It is crucial to note that estimating model performance solely based on standard benchmark datasets may not necessarily apply to the datasets with distinct features from the benchmark datasets. Our proposed framework plays a pivotal role in identifying and addressing biases detected in word embeddings, particularly evident in low-resource regions such as New Zealand.

Abstract Active learning allows machine learning models to be trained using fewer labels while retaining similar performance to traditional supervised learning. An active learner selects the most informative data points, requests their labels, and retrains itself. While this approach is promising, it raises the question of how to determine when the model is ‘good enough’ without the additional labels required for traditional evaluation. Previously, different stopping criteria have been proposed aiming to identify the optimal stopping point. Yet, optimality can only be expressed as a domain-dependent trade-off between accuracy and the number of labels, and no criterion is superior in all applications. As a further complication, a comparison of criteria for a particular real-world application would require practitioners to collect additional labelled data they are aiming to avoid by using active learning in the first place. This work enables practitioners to employ active learning by providing actionable recommendations for which stopping criteria are best for a given real-world scenario. We contribute the first large-scale comparison of stopping criteria for pool-based active learning, using a cost measure to quantify the accuracy/label trade-off, public implementations of all stopping criteria we evaluate, and an open-source framework for evaluating stopping criteria. Our research enables practitioners to substantially reduce labelling costs by utilizing the stopping criterion which best suits their domain.

Machine Learning typically assumes that training and test sets are independently drawn from the same distribution, but this assumption is often violated in practice which creates a bias. Many attempts to identify and mitigate this bias have been proposed, but they usually rely on ground-truth information. But what if the researcher is not even aware of the bias? In contrast to prior work, this paper introduces a new method, Imitate, to identify and mitigate Selection Bias in the case that we may not know if (and where) a bias is present, and hence no ground-truth information is available. Imitate investigates the dataset's probability density, then adds generated points in order to smooth out the density and have it resemble a Gaussian, the most common density occurring in real-world applications. If the artificial points focus on certain areas and are not widespread, this could indicate a Selection Bias where these areas are underrepresented in the sample. We demonstrate the effectiveness of the proposed method in both, synthetic and real-world datasets. We also point out limitations and future research directions.

Spreadsheets are one of the most widely used data management tools. Although intuitive and easy to use, they suffer from a number of issues. Numerous publications indicate that most spreadsheets contain errors and that spreadsheet-based data shadow systems lead to problems such as the “spreadmart dilemma” creating inconsistent views on organizational data. In this paper, we describe “ The Semantic Spreadsheet ”, a new data model for a semantically accurate spreadsheet system. Unlike the existing data models, the described model is a presentation independent data model based on the Resource Description Framework (RDF) that avoids the spreadmart dilemma by providing a semantically sound data structure. We describe the set-theoretic and relational algebraic operations on the data model and show how it can improve data integrity.