Publications
2026
- ConferenceDetoxAI: A Python Toolkit for Debiasing Deep Learning Models in Computer VisionIn Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track and Demo Track, 2026
While machine learning fairness has made significant progress in recent years, most existing solutions focus on tabular data and are poorly suited for vision-based classification tasks, which rely heavily on deep learning. To bridge this gap, we introduce DetoxAI, an open-source Python library for improving fairness in deep learning vision classifiers through post-hoc debiasing. DetoxAI implements state-of-the-art debiasing algorithms, fairness metrics, and visualization tools. It supports debiasing via interventions in internal representations and includes attribution-based visualization tools and quantitative algorithmic fairness metrics to show how bias is mitigated. This paper presents the motivation, design, and use cases of DetoxAI, demonstrating its tangible value to engineers and researchers.",
@inproceedings{detoxai2025, title = {DetoxAI: A Python Toolkit for Debiasing Deep Learning Models in Computer Vision}, year = {2026}, author = {St\k{e}pka, Ignacy and Sztukiewicz, Lukasz and Wili\'{n}ski, Micha\l{} and Stefanowski, Jerzy}, address = {Cham}, publisher = {Springer Nature Switzerland}, booktitle = {Machine Learning and Knowledge Discovery in Databases. Applied Data Science Track and Demo Track}, pages = {502--505}, isbn = {978-3-032-06129-4}, } - PreprintMICA: Multivariate Infini Compressive Attention for Time Series ForecastingWilla Potosnak, Nina Żukowska, Michał Wiliński, Dan Howarth, Ignacy Stępka, Mononito Goswami, and Artur DubrawskiarXiv preprint, 2026
Multivariate forecasting with Transformers faces a core scalability challenge: modeling cross-channel dependencies via attention compounds attention’s quadratic sequence complexity with quadratic channel scaling, making full cross-channel attention impractical for high-dimensional time series. We propose Multivariate Infini Compressive Attention (MICA), an architectural design to extend channel-independent Transformers to channel-dependent forecasting. By adapting efficient attention techniques from the sequence dimension to the channel dimension, MICA adds a cross-channel attention mechanism to channel-independent backbones that scales linearly with channel count and context length. We evaluate channel-independent Transformer architectures with and without MICA across multiple forecasting benchmarks. MICA reduces forecast error over its channel-independent counterparts by 5.4% on average and up to 25.4% on individual datasets, highlighting the importance of explicit cross-channel modeling. Moreover, models with MICA rank first among deep multivariate Transformer and MLP baselines. MICA models also scale more efficiently with respect to both channel count and context length than Transformer baselines that compute attention across both the temporal and channel dimensions, establishing compressive attention as a practical solution for scalable multivariate forecasting.
@article{potosnak2026mica, title = {{MICA}: Multivariate Infini Compressive Attention for Time Series Forecasting}, author = {Potosnak, Willa and {\.{Z}}ukowska, Nina and Wili\'{n}ski, Micha\l{} and Howarth, Dan and St\k{e}pka, Ignacy and Goswami, Mononito and Dubrawski, Artur}, year = {2026}, journal = {arXiv preprint}, } - PreprintLost in Reconstruction: Aligning Action Representations with Language in Vision-Language-Action ModelsLi Wenjie, Yash Jangir, Ignacy Stępka, Yash Agarwal, Marion Kipsang, and Yonatan BiskarXiv preprint, 2026
Action verbs describe not only the physical outcomes of actions, but also how those actions are performed. Yet action representations in vision-language-action models (VLAs) are typically optimized for reconstruction under L1/L2 losses in raw action space, where numerical proximity need not reflect linguistically meaningful distinctions. On BridgeV2, we show that action trajectories contain verb-grounding information beyond visual state changes, and that reconstruction-only discrete tokenization systematically erodes this information. To address this problem, we introduce SALT, a Semantically ALigned action Tokenizer that augments a VQ-VAE-style tokenizer with an auxiliary objective requiring a frozen vision-language model to recover the episode instruction from quantized action latents. Policies trained with SALT achieve 71.9% average success in SimplerEnv, compared with 42.7% for a reconstruction-only VQ-VAE tokenizer and 31.2% for FAST. SALT also develops verb-specialized codes while maintaining reconstruction fidelity. These results show that robot action trajectories provide a source of language grounding and that preserving this structure in action representations can substantially improve language-conditioned control.
@article{li2026salt, title = {Lost in Reconstruction: Aligning Action Representations with Language in Vision-Language-Action Models}, author = {Wenjie, Li and Jangir, Yash and St\k{e}pka, Ignacy and Agarwal, Yash and Kipsang, Marion and Bisk, Yonatan}, year = {2026}, journal = {arXiv preprint}, }
2025
- ConferenceCounterfactual Explanations with Probabilistic Guarantees on their Robustness to Model ChangeIgnacy Stępka, Mateusz Lango, and Jerzy StefanowskiIn 31st SIGKDD Conference on Knowledge Discovery and Data Mining - Research Track, 2025
Counterfactual explanations (CFEs) guide users on how to adjust inputs to machine learning models to achieve desired outputs. While existing research primarily addresses static scenarios, real-world applications often involve data or model changes, potentially invalidating previously generated CFEs and rendering user-induced input changes ineffective. Current methods addressing this issue often support only specific models or change types, require extensive hyperparameter tuning, or fail to provide probabilistic guarantees on CFE robustness to model changes. This paper proposes a novel approach for generating CFEs that provides probabilistic guarantees for any model and change type, while offering interpretable and easy-to-select hyperparameters. We establish a theoretical framework for probabilistically defining robustness to model change and demonstrate how our BetaRCE method directly stems from it. BetaRCE is a post-hoc method applied alongside a chosen base CFE generation method to enhance the quality of the explanation beyond robustness. It facilitates a transition from the base explanation to a more robust one with user-adjusted probability bounds. Through experimental comparisons with baselines, we show that BetaRCE yields robust, most plausible, and closest to baseline counterfactual explanations.
@inproceedings{stepka2024cfeprob, title = {Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Change}, author = {St\k{e}pka, Ignacy and Lango, Mateusz and Stefanowski, Jerzy}, booktitle = {31st SIGKDD Conference on Knowledge Discovery and Data Mining - Research Track}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, year = {2025}, doi = {10.1145/3690624.3709300} } - ConferenceTowards Fairness for the Right Reasons: Using Saliency Maps to Evaluate Bias Removal in Neural NetworksIn The 3rd World Conference on eXplainable Artificial Intelligence, XAI-2025, Istanbul, Türkiye, 2025
The widespread adoption of machine learning systems has raised critical concerns about fairness and bias, making mitigating harmful biases essential for AI development. In this paper, we investigate the relationship between debiasing and removing artifacts in neural networks for computer vision tasks. First, we introduce a set of novel XAI-based metrics that analyze saliency maps to assess shifts in a model’s decision-making process. Then, we demonstrate that successful debiasing methods systematically redirect model focus away from protected attributes. Finally, we show that techniques originally developed for artifact removal can be effectively repurposed for improving fairness. These findings provide evidence for the existence of a bidirectional connection between ensuring fairness and removing artifacts corresponding to protected attributes.
@inproceedings{sztukiewicz2025, title = {Towards Fairness for the Right Reasons: Using Saliency Maps to Evaluate Bias Removal in Neural Networks}, booktitle = {The 3rd World Conference on eXplainable Artificial Intelligence, XAI-2025}, author = {Sztukiewicz, Lukasz and St\k{e}pka, Ignacy and Wili\'{n}ski, Micha\l{} and Stefanowski, Jerzy}, year = {2025}, location = {Istanbul, T\"urkiye}, } - WorkshopMitigating Persistent Client Dropout in Asynchronous Decentralized Federated LearningIn FedKDD Workshop at the 31st SIGKDD Conference on Knowledge Discovery and Data Mining, Toronto, Canada, 2025
We consider the problem of persistent client dropout in asynchronous Decentralized Federated Learning (DFL). Asynchronicity and decentralization obfuscate information about model updates among federation peers, making recovery from a client dropout difficult. Access to the number of learning epochs, data distributions, and all the information necessary to precisely reconstruct the missing neighbor’s loss functions is limited. We show that obvious mitigations do not adequately address the problem and introduce adaptive strategies based on client reconstruction. We show that these strategies can effectively recover some performance loss caused by dropout. Our work focuses on asynchronous DFL with local regularization and differs substantially from that in the existing literature. We evaluate the proposed methods on tabular and image datasets, involve three DFL algorithms, and three data heterogeneity scenarios (iid, non-iid, class-focused non-iid). Our experiments show that the proposed adaptive strategies can be effective in maintaining robustness of federated learning, even if they do not reconstruct the missing client’s data precisely. We also discuss the limitations and identify future avenues for tackling the problem of client dropout.
@inproceedings{stepka2025mitigating, title = {Mitigating Persistent Client Dropout in Asynchronous Decentralized Federated Learning}, booktitle = {FedKDD Workshop at the 31st SIGKDD Conference on Knowledge Discovery and Data Mining}, author = {St\k{e}pka, Ignacy and Gisolfi, Nicholas and Tr\k{e}bacz, Kacper and Dubrawski, Artur}, year = {2025}, location = {Toronto, Canada}, } - WorkshopExplaining Concept Drift through the Evolution of Group CounterfactualsIgnacy Stępka, and Jerzy StefanowskiIn 2nd TempXAI Workshop for Explainable AI in Time Series and Data Streams at Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 2025
Machine learning models in dynamic environments often suffer from concept drift, where changes in the data distribution degrade performance. While detecting this drift is a well-studied topic, explaining how and why the model’s decision-making logic changes still remains a significant challenge. In this paper, we introduce a novel methodology to explain concept drift by analyzing the temporal evolution of group-based counterfactual explanations (GCEs). Our approach tracks shifts in the GCEs’ cluster centroids and their associated counterfactual action vectors before and after a drift. These evolving GCEs act as an interpretable proxy, revealing structural changes in the model’s decision boundary and its underlying rationale. We operationalize this analysis within a three-layer framework that synergistically combines insights from the data layer (distributional shifts), the model layer (prediction disagreement), and our proposed explanation layer. We show that such holistic view allows for a more comprehensive diagnosis of drift, making it possible to distinguish between different root causes, such as a spatial data shift versus a re-labeling of concepts.
@inproceedings{stepka2025gce, title = {Explaining Concept Drift through the Evolution of Group Counterfactuals}, author = {St\k{e}pka, Ignacy and Stefanowski, Jerzy}, year = {2025}, booktitle = {2nd TempXAI Workshop for Explainable AI in Time Series and Data Streams at Joint European Conference on Machine Learning and Knowledge Discovery in Databases}, }
2024
- JournalA Multi–Criteria Approach for Selecting an Explanation from the Set of Counterfactuals Produced by an Ensemble of ExplainersIgnacy Stępka, Mateusz Lango, and Jerzy StefanowskiInternational Journal of Applied Mathematics and Computer Science, 2024
Counterfactuals are widely used to explain ML model predictions by providing alternative scenarios for obtaining the more desired predictions. They can be generated by a variety of methods that optimize different, sometimes conflicting, quality measures and produce quite different solutions. However, choosing the most appropriate explanation method and one of the generated counterfactuals is not an easy task. Instead of forcing the user to test many different explanation methods and analysing conflicting solutions, in this paper, we propose to use a multi-stage ensemble approach that will select single counterfactual based on multiple-critera analysis, in order to offer a compromise solution that scores well on varied quality measures. This approach exploits the dominance relation and the ideal point decision aid method, which selects one counterfactual from the Pareto front. The conducted experiments demonstrated that the proposed approach generates fully actionable counterfactuals with attractive compromise values of the considered quality measures.
@article{stepka2024multi, title = {A Multi--Criteria Approach for Selecting an Explanation from the Set of Counterfactuals Produced by an Ensemble of Explainers}, author = {St\k{e}pka, Ignacy and Lango, Mateusz and Stefanowski, Jerzy}, journal = {International Journal of Applied Mathematics and Computer Science}, volume = {34}, number = {1}, pages = {119--133}, year = {2024}, doi = {10.61822/amcs-2024-0009}, } - WorkshopA SAT-based approach to rigorous verification of Bayesian networksIgnacy Stępka, Nicholas Gisolfi, and Artur DubrawskiIn Workshop on Explainable and Robust AI for Industry 4.0 & 5.0 (X-RAI) at Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 2024
Recent advancements in machine learning have accelerated its widespread adoption across various real-world applications. However, in safety-critical domains, the deployment of machine learning models is riddled with challenges due to their complexity, lack of interpretability, and absence of formal guarantees regarding their behavior. In this paper, we introduce a verification framework tailored for Bayesian networks, designed to address these drawbacks. Our framework comprises two key components: (1) a two-step compilation and encoding scheme that translates Bayesian networks into Boolean logic literals, and (2) formal verification queries that leverage these literals to verify various properties encoded as constraints. Specifically, we introduce two verification queries: if-then rules (ITR) and feature monotonicity (FMO). We benchmark the efficiency of our verification scheme and demonstrate its practical utility in real-world scenarios.
@inproceedings{stepka2024sat, title = {A SAT-based approach to rigorous verification of Bayesian networks}, author = {St\k{e}pka, Ignacy and Gisolfi, Nicholas and Dubrawski, Artur}, year = {2024}, booktitle = {Workshop on Explainable and Robust AI for Industry 4.0 & 5.0 (X-RAI) at Joint European Conference on Machine Learning and Knowledge Discovery in Databases}, }
2023
- ConferenceOn usefulness of dominance relation for selecting counterfactuals from the ensemble of explainersIgnacy Stępka, Mateusz Lango, and Jerzy StefanowskiIn Proceedings of the 4rd Polish Conference on Artificial Intelligence, PP-RAI 2023, 2023
Counterfactual explanations are widely used to explain ML model predictions by providing alternative scenarios. However, choosing the most appropriate explanation method and one of generated counterfactuals is not an easy task. In this paper, we propose an approach that filters out a large set of counterfactuals generated by a set of diverse algorithms through a multi-criteria subset selection problem solved using the dominance relation. Experiments show that exploiting the dominance relation results in a concise set of counterfactual explanations.
@inproceedings{stepka2023usefulness, title = {On usefulness of dominance relation for selecting counterfactuals from the ensemble of explainers}, author = {St\k{e}pka, Ignacy and Lango, Mateusz and Stefanowski, Jerzy}, booktitle = {Proceedings of the 4rd Polish Conference on Artificial Intelligence, PP-RAI 2023}, year = {2023}, publisher = {Wydawnictwo Politechniki Łódzkiej}, doi = {10.34658/9788366741928}, pages = {125-130}, }