Fair graph clustering seeks partitions that respect network structure while maintaining proportional representation across sensitive groups, with applications spanning community detection, team formation, resource allocation, and social network analysis. Many existing approaches enforce rigid constraints or rely on multi-stage pipelines (e.g., spectral embedding followed by $k$-means), limiting trade-off control, interpretability, and scalability. We introduce \emph{DFNMF}, an end-to-end deep nonnegative tri-factorization tailored to graphs that directly optimizes cluster assignments with a soft statistical-parity regularizer. A single parameter $λ$ tunes the fairness--utility balance, while nonnegativity yields parts-based factors and transparent soft memberships. The optimization uses sparse-friendly alternating updates and scales near-linearly with the number of edges. Across synthetic and real networks, DFNMF achieves substantially higher group balance at comparable modularity, often dominating state-of-the-art baselines on the Pareto front. The code is available at https://github.com/SiamakGhodsi/DFNMF.git.

Fairness-aware classification requires balancing performance and fairness, often intensified by intersectional biases. Conflicting fairness definitions further complicate the task, making it difficult to identify universally fair solutions. Despite growing regulatory and societal demands for equitable AI, popular toolkits offer limited support for exploring multi-dimensional fairness and related trade-offs. To address this, we present mmm-fair, an open-source toolkit leveraging boosting-based ensemble approaches that dynamically optimizes model weights to jointly minimize classification errors and diverse fairness violations, enabling flexible multi-objective optimization. The system empowers users to deploy models that align with their context-specific needs while reliably uncovering intersectional biases often missed by state-of-the-art methods. In a nutshell, mmm-fair uniquely combines in-depth multi-attribute fairness, multi-objective optimization, a no-code, chat-based interface, LLM-powered explanations, interactive Pareto exploration for model selection, custom fairness constraint definition, and deployment-ready models in a single open-source toolkit, a combination rarely found in existing fairness tools. Demo walkthrough available at: https://youtu.be/_rcpjlXFqkw.

Ensuring fairness in machine learning remains a significant challenge, as models often inherit biases from their training data. Generative models have recently emerged as a promising approach to mitigate bias at the data level while preserving utility. However, many rely on deep architectures, despite evidence that simpler models can be highly effective for tabular data. In this work, we introduce TabFairGDT, a novel method for generating fair synthetic tabular data using autoregressive decision trees. To enforce fairness, we propose a soft leaf resampling technique that adjusts decision tree outputs to reduce bias while preserving predictive performance. Our approach is non-parametric, effectively capturing complex relationships between mixed feature types, without relying on assumptions about the underlying data distributions. We evaluate TabFairGDT on benchmark fairness datasets and demonstrate that it outperforms state-of-the-art (SOTA) deep generative models, achieving better fairness-utility trade-off for downstream tasks, as well as higher synthetic data quality. Moreover, our method is lightweight, highly efficient, and CPU-compatible, requiring no data preprocessing. Remarkably, TabFairGDT achieves a 72% average speedup over the fastest SOTA baseline across various dataset sizes, and can generate fair synthetic data for medium-sized datasets (10 features, 10K samples) in just one second on a standard CPU, making it an ideal solution for realworld fairness-sensitive applications.

Recent literature highlights the critical role of neighborhood construction in deriving model-agnostic explanations, with a growing trend toward deploying generative models to improve synthetic instance quality, especially for explaining text classifiers. These approaches overcome the challenges in neighborhood construction posed by the unstructured nature of texts, thereby improving the quality of explanations. However, the deployed generators are usually implemented via neural networks and lack inherent explainability, sparking arguments over the transparency of the explanation process itself. To address this limitation while preserving neighborhood quality, this paper introduces a probability-based editing method as an alternative to black-box text generators. This approach generates neighboring texts by implementing manipulations based on in-text contexts. Substituting the generator-based construction process with recur-sive probability-based editing, the resultant explanation method, XPROB (explainer with probability-based editing), exhibits com-petitive performance according to the evaluation conducted on two real-world datasets. Additionally, XPROB's fully transparent and more controllable construction process leads to superior stability compared to the generator-based explainers.

Abstract Computer vision systems are employed in a variety of high-impact applications. However, making them trustworthy requires methods for the detection of potential biases in their training data, before models learn to harm already disadvantaged groups in downstream applications. Image data are typically represented via extracted features, which can be hand-crafted or pre-trained neural network embeddings. In this work, we introduce a framework for bias discovery given such features that is based on optimal transport theory; it uses the (quadratic) Wasserstein distance to quantify disparity between the feature distributions of two demographic groups (e.g., women vs men). In this context, we show that the Kantorovich potentials of the images, which are a byproduct of computing the Wasserstein distance and act as “transportation prices", can serve as bias scores by indicating which images might exhibit distinct biased characteristics. We thus introduce a visual dataset exploration pipeline that helps auditors identify common characteristics across high- or low-scored images as potential sources of bias. We conduct a case study to identify prospective gender biases and demonstrate theoretically-derived properties with experiments on the CelebA and Biased MNIST datasets.

In this article the statement after Equation 1 had an error in the published version. Please refer the correction as follows: “where ν = T#µ and T# is the push-forward of µ along the function T : X → Y” was incorrectly written as “where T# is the push-forward of µ along the function T : X → Y. Furthermore, Equation 1 itself was incorrectly formulated. Namely, the integral should have been over X and not over X × Y. The original article has been corrected.

Abstract Group work is a prevalent activity in educational settings, where students are often divided into topic-specific groups based on their preferences. The grouping should reflect students’ aspirations as much as possible. Usually, the resulting groups should also be balanced in terms of protected attributes like gender, as studies suggest that students may learn better in mixed-gender groups. Moreover, to allow a fair workload across the groups, the cardinalities of the different groups should be balanced. In this paper, we introduce a multi-fair capacitated (MFC) grouping problem that fairly partitions students into non-overlapping groups while ensuring balanced group cardinalities (with a lower and an upper bound), and maximizing the diversity of members regarding the protected attribute. To obtain the MFC grouping, we propose three approaches: a greedy heuristic approach, a knapsack-based approach using vanilla maximal knapsack formulation, and an MFC knapsack approach based on group fairness knapsack formulation. Experimental results on a real dataset and a semi-synthetic dataset show that our proposed methods can satisfy students’ preferences and deliver balanced and diverse groups regarding cardinality and the protected attribute, respectively.

Abstract Class imbalance poses a major challenge for machine learning as most supervised learning models might exhibit bias towards the majority class and under-perform in the minority class. Cost-sensitive learning tackles this problem by treating the classes differently, formulated typically via a user-defined fixed misclassification cost matrix provided as input to the learner. Such parameter tuning is a challenging task that requires domain knowledge and moreover, wrong adjustments might lead to overall predictive performance deterioration. In this work, we propose a novel cost-sensitive boosting approach for imbalanced data that dynamically adjusts the misclassification costs over the boosting rounds in response to model’s performance instead of using a fixed misclassification cost matrix. Our method, called AdaCC, is parameter-free as it relies on the cumulative behavior of the boosting model in order to adjust the misclassification costs for the next boosting round and comes with theoretical guarantees regarding the training error. Experiments on 27 real-world datasets from different domains with high class imbalance demonstrate the superiority of our method over 12 state-of-the-art cost-sensitive boosting approaches exhibiting consistent improvements in different measures, for instance, in the range of [0.3–28.56%] for AUC, [3.4–21.4%] for balanced accuracy, [4.8–45%] for gmean and [7.4–85.5%] for recall.

Hate speech detection is a common downstream application of natural language processing (NLP) in the real world. In spite of the increasing accuracy, current data-driven approaches could easily learn biases from the imbalanced data distributions originating from humans. The deployment of biased models could further enhance the existing social biases. But unlike handling tabular data, defining and mitigating biases in text classifiers, which deal with unstructured data, are more challenging. A popular solution for improving machine learning fairness in NLP is to conduct the debiasing process with a list of potentially discriminated words given by human annotators. In addition to suffering from the risks of overlooking the biased terms, exhaustively identifying bias with human annotators are unsustainable since discrimination is variable among different datasets and may evolve over time. To this end, we propose an automatic misuse detector (MiD) relying on an explanation method for detecting potential bias. And built upon that, an end-to-end debiasing framework with the proposed staged correction is designed for text classifiers without any external resources required.

The importance of the neighborhood for training a local surrogate model to approximate the local decision boundary of a black box classifier has been already highlighted in the literature. Several attempts have been made to construct a better neighborhood for high dimensional data, like texts, by using generative autoencoders. However, existing approaches mainly generate neighbors by selecting purely at random from the latent space and struggle under the curse of dimensionality to learn a good local decision boundary. To overcome this problem, we propose a progressive approximation of the neighborhood using counterfactual instances as initial landmarks and a careful 2-stage sampling approach to refine counterfactuals and generate factuals in the neighborhood of the input instance to be explained. Our work focuses on textual data and our explanations consist of both word-level explanations from the original instance (intrinsic) and the neighborhood (extrinsic) and factual- and counterfactual-instances discovered during the neighborhood generation process that further reveal the effect of altering certain parts in the input text. Our experiments on real-world datasets demonstrate that our method outperforms the competitors in terms of usefulness and stability (for the qualitative part) and completeness, compactness and correctness (for the quantitative part).

Online class imbalance learning deals with data streams that are affected by both concept drift and class imbalance. Online learning tries to find a trade-off between exploiting previously learned information and incorporating new information into the model. This requires both the incremental update of the model and the ability to unlearn outdated information. The improper use of unlearning, however, can lead to the retroactive interference problem, a phenomenon that occurs when newly learned information interferes with the old information and impedes the recall of previously learned information. The problem becomes more severe when the classes are not equally represented, resulting in the removal of minority information from the model. In this work, we propose the Drift-Aware Multi-Memory Model (DAM3), which addresses the class imbalance problem in online learning for memory-based models. DAM3 mitigates class imbalance by incorporating an imbalance-sensitive drift detector, preserving a balanced representation of classes in the model, and resolving retroactive interference using a working memory that prevents the forgetting of old information. We show through experiments on real-world and synthetic datasets that the proposed method mitigates class imbalance and outperforms the state-of-the-art methods.

Automated decision making based on big data and machine learning (ML) algorithms can result in discriminatory decisions against certain protected groups defined upon personal data like gender, race, sexual orientation etc. Such algorithms designed to discover patterns in big data might not only pick up any encoded societal biases in the training data, but even worse, they might reinforce such biases resulting in more severe discrimination. The majority of thus far proposed fairness-aware machine learning approaches focus solely on the pre-, in- or post-processing steps of the machine learning process, that is, input data, learning algorithms or derived models, respectively. However, the fairness problem cannot be isolated to a single step of the ML process. Rather, discrimination is often a result of complex interactions between big data and algorithms, and therefore, a more holistic approach is required. The proposed FAE (Fairness-Aware Ensemble) framework combines fairness-related interventions at both pre-and post-processing steps of the data analysis process. In the pre-processing step, we tackle the problems of under-representation of the protected group (group imbalance) and of class-imbalance by generating balanced training samples. In the post-processing step, we tackle the problem of class overlapping by shifting the decision boundary in the direction of fairness.

The widespread use of ML-based decision making in domains with high societal impact such as recidivism, job hiring and loan credit has raised a lot of concerns regarding potential discrimination. In particular, in certain cases it has been observed that ML algorithms can provide different decisions based on sensitive attributes such as gender or race and therefore can lead to discrimination. Although, several fairness-aware ML approaches have been proposed, their focus has been largely on preserving the overall classification accuracy while improving fairness in predictions for both protected and non-protected groups (defined based on the sensitive attribute(s)). The overall accuracy however is not a good indicator of performance in case of class imbalance, as it is biased towards the majority class. As we will see in our experiments, many of the fairness-related datasets suffer from class imbalance and therefore, tackling fairness requires also tackling the imbalance problem. To this end, we propose AdaFair, a fairness-aware classifier based on AdaBoost that further updates the weights of the instances in each boosting round taking into account a cumulative notion of fairness based upon all current ensemble members, while explicitly tackling class-imbalance by optimizing the number of ensemble members for balanced classification error. Our experiments show that our approach can achieve parity in true positive and true negative rates for both protected and non-protected groups, while it significantly outperforms existing fairness-aware methods up to 25% in terms of balanced error.

Huge amounts of textual streams are generated nowadays, especially in social networks like Twitter and Facebook. As the discussion topics and user opinions on those topics change drastically with time, those streams undergo changes in data distribution, leading to changes in the concept to be learned, a phenomenon called concept drift. One particular type of drift, that has not yet attracted a lot of attention is feature drift, i.e., changes in the features that are relevant for the learning task at hand. In this work, we propose an approach for handling feature drifts in textual streams. Our approach integrates i) an ensemble-based mechanism to accurately predict the feature/word values for the next time-point by taking into account the different features might be subject to different temporal trends and ii) a sketch-based feature space maintenance mechanism that allows for a memory-bounded maintenance of the feature space over the stream. Experiments with textual streams from the sentiment analysis, email preference and spam detection demonstrate that our approach achieves significantly better or competitive performance compared to baselines.

Stream classification algorithms traditionally treat arriving observations as independent. However, in many applications the arriving examples may depend on the "entity" that generated them, e.g. in product reviewing or in the interactions of users with an application server. In this study, we investigate the potential of this dependency by partitioning the original stream of observations into entity-centric substreams and by incorporating entity-specific information into the learning model. We propose a k Nearest Neighbour inspired stream classification approach (kNN), in which the label of an arriving observation is predicted by exploiting knowledge on the observations belonging to this entity and to entities similar to it. For the computation of entity similarity, we consider knowledge about the observations and knowledge about the entity, potentially transferred from another domain. To distinguish between cases where this kind of knowledge transfer is beneficial for stream classification and cases where the knowledge on the entities does not contribute to classifying the observations, we also propose a heuristic approach based on random sampling of substreams using k Random Entities (kRE). Our learning scenario is not fully supervised: after acquiring labels for the initial few observations of each entity, we assume that no additional labels arrive, and attempt to predict the labels of near-future and far-future observations from that initial seed. We report on our findings from three datasets.

Sentiment analysis is an important task in order to gain insights over the huge amounts of opinions that are generated in the social media on a daily basis. Although there is a lot of work on sentiment analysis, there are no many datasets available which one can use for developing new methods and for evaluation. To the best of our knowledge, the largest dataset for sentiment analysis is TSentiment [8], a 1.6 millions machine-annotated tweets dataset covering a period of about 3 months in 2009. This dataset however is too short and therefore insufficient to study heterogeneous, fast evolving streams. Therefore, we annotated the Twitter dataset of 2015 (228 million tweets without retweets and 275 million with retweets) and we make it publicly available for research. For the annotation we leverage the power of unlabeled data, together with labeled data using semi-supervised learning and in particular, Self-Learning and Co-Training. Our main contribution is the provision of the TSentiment15 dataset together with insights from the analysis, which includes a batch and a stream-processing of the data. In the former, all labeled and unlabeled data are available to the algorithms from the beginning, whereas in the later, they are revealed gradually based on their arrival time in the stream.

A collection of datasets crawled from Amazon, “Amazon reviews”, is popular in the evaluation of recommendation systems. These datasets, however, contain redundancies (duplicated recommendations for variants of certain items). These redundancies went unnoticed in earlier use of these datasets and thus incurred to a certain extent wrong conclusions in the evaluation of algorithms tested on these datasets. We analyze the nature and amount of these redundancies and their impact on the evaluation of recommendation methods. While the general and obvious conclusion is that redundancies should be avoided and datasets should be carefully preprocessed, we observe more specifically that their impact depends on the complexity of the methods. With this work, we also want to raise the awareness of the importance of data quality, model understanding, and appropriate evaluation.

Top-k is a well-studied problem in the literature, due to its wide spectrum of applications, like information retrieval, database querying, Web search and data mining. In the big data era, the volume of the data and their velocity, call for efficient parallel solutions that overcome the restricted resources of a single machine. Our motivating application is recommenders, which typically deal with big numbers of users and items, but other applications might benefit as well, like keyword search. In this paper, we propose a parallel top-k MapReduce algorithm that, unlike existing MapReduce solutions, manages to handle cases in which the k results do not fit in memory.

Isotopic mapping has become an indispensable tool for the assessment of mobility and trade of the past. However, modeling and understanding spatio-temporal isotopic variation is complicated by the small number of available samples, potential mobility of the investigated samples, sample preservation quality, uncertainty of measurements, and so forth. In this work, we use data mining techniques to build an isotopic map (descriptive modeling) and to determine the spatial origin of new samples (predictive modeling). In particular, we propose a clustering-based isotope ratio model and a scoring function for the origin prediction of new samples. Our data was extracted from real animal finds from an Alpine passage that spans three countries (Germany, Austria, and Italy) and comprises a high variety of isotopes and geological characteristics. Our results and evaluation by domain experts show that it is possible to derive a model of the area for both descriptive and predictive purposes.

Although living in the information age for decades, paperwork is still a tedious part of everybody's life. Assistance systems that implement techniques of digitization and document understanding may offer considerable reductions in time and effort for the users. A large portion of paper documents like invoices, delivery receipts or admonitions are based on a fixed company specific template and therefore exhibit a high degree of similarity. In this work, we propose a template extraction method over a stream of incoming documents and a template allocation method for assigning new instances from the stream to the most suitable templates. Our method employs text augmented by layout information to represent the digital image of the paper document. Document similarity is assessed with respect to both textual and layout parts of the document; the matching terms contribute accordingly to their distance to the query terms. To be more robust against distortions on the documents due to the digitization process, the templates are not static, rather they are maintained in an online fashion based on their new assigned documents. Real data experiments show that the combination of textual and layout information and the continuous template adaptation through online update, improves the template identification quality of earlier proposed methods.

Nowadays, WWW brings overwhelming variety of choices to consumers. Recommendation systems facilitate the selection by issuing recommendations to them. Recommendations for users, or groups, are determined by considering users similar to the users in question. Scanning the whole database for locating similar users, though, is expensive. Existing approaches build cluster models by employing full-dimensional clustering to find sets of similar users. As the datasets we deal with are high-dimensional and incomplete, full-dimensional clustering is not the best option. To this end, we explore the fault-tolerant subspace clustering approach. We extend the concept of fault tolerance to density-based subspace clustering, and to speed up our algorithms, we introduce the significance threshold for considering only promising dimensions for subspace extension. Moreover, as we potentially receive a multitude of users from subspace clustering, we propose a weighted ranking approach to refine the set of like-minded users. Our experiments on real movie datasets show that the diversification of the similar users that the subspace clustering approaches offer results in better recommendations compared to traditional collaborative filtering and full-dimensional clustering approaches.

Typically, recommendations are computed by considering users similar to the user in question. However, scanning the whole database of users for locating similar users is expensive. Existing approaches build user profiles by employing full-dimensional clustering to find sets of similar users. As the datasets we deal with are high-dimensional and incomplete, full-dimensional clustering is not the best option. To this end, we explore the fault tolerance subspace clustering approach that detects clusters of similar users in subspaces of the original feature space and also allows for missing values. Our experiments on real movie datasets show that the diversification of the similar users through subspace clustering results in better recommendations comparing to traditional collaborative filtering and full dimensional clustering approaches.

Nowadays social media are widely used for the broadcasting of different types of information, such as events, activities and opinions. Analyzing this vast amount of data for extracting models that describe individual users or groups of users has gained a lot of attention lately. In this work we analyze individual users and monitor changes in their published content over time. We propose a topic-based user profiling and monitoring approach for change detection and monitoring of profile evolution. Our method is capable of detecting persistent topics representing long term interests of the user as well as short term topics that refer to everyday events or reactions to the news. We evaluate our approach on real data from Twitter.

Clustering of high dimensional data streams is an important problem in many application domains, a prominent example being network monitoring. Several approaches have been lately proposed for solving independently the different aspects of the problem. There exist methods for clustering over full dimensional streams and methods for finding clusters in subspaces of high dimensional static data. Yet only a few approaches have been proposed so far which tackle both the stream and the high dimensionality aspects of the problem simultaneously. In this work, we propose a new density-based projected clustering algorithm, HDDSTREAM, for high dimensional data streams. Our algorithm summarizes both the data points and the dimensions where these points are grouped together and maintains these summaries online, as new points arrive over time and old points expire due to ageing. Our experimental results illustrate the effectiveness and the efficiency of HDDSTREAM and also demonstrate that it could serve as a trigger for detecting drastic changes in the underlying stream population, like bursts of network attacks.

This demo presents Pattern-Miner, an integrated environment for pattern management and mining that deals with the whole lifecycle of patterns, from their generation (using data mining techniques) to their storage and querying, putting also emphasis on the comparison between patterns and meta-mining operations over the extracted patterns. Pattern comparison (comparing results of the data mining process) and meta-mining are high level pattern operations that can be applied in a variety of applications, from database change management to image comparison and retrieval.

Decision trees are among the most popular pattern types in data mining due to their intuitive representation. However, little attention has been given on the definition of measures of semantic similarity between decision trees. In this work, we present a general framework for similarity estimation that includes as special cases the estimation of semantic similarity between decision trees, as well as various forms of similarity estimation on classification datasets with respect to different probability distributions defined over the attribute-class space of the datasets. The similarity estimation is based on the partitions induced by the decision trees on the attribute space of the datasets. We use the proposed framework in order to estimate the semantic similarity of decision trees induced from different subsamples of classification datasets; we evaluate its performance with respect to the empirical semantic similarity, which we estimate on the basis of independent hold-out test sets. The availability of similarity measures on decision trees opens a wide range of possibilities for meta-analysis and meta-mining of the data mining results.

There is much recent work on detecting and tracking change in clusters, often based on the study of the spatiotemporal properties of a cluster. For the many applications where cluster change is relevant, among them customer relationship management, fraud detection and marketing, it is also necessary to provide insights about the nature of cluster change: Is a cluster corresponding to a group of customers simply disappearing or are its members migrating to other clusters? Is a new emerging cluster reflecting a new target group of customers or does it rather consist of existing customers whose preferences shift? To answer such questions, we propose the framework MONIC for modeling and tracking of cluster transitions. Our cluster transition model encompasses changes that involve more than one cluster, thus allowing for insights on cluster change in the whole clustering. Our transition tracking mechanism is not based on the topological properties of clusters, which are only available for some types of clustering, but on the contents of the underlying data stream. We present our first results on monitoring cluster transitions over the ACM digital library.