Przemysław Biecek

Mateusz Krzyziński, Mikołaj Spytek, Hubert Baniecki et al.

Machine and deep learning survival models demonstrate similar or even improved time-to-event prediction capabilities compared to classical statistical learning methods yet are too complex to be interpreted by humans. Several model-agnostic explanations are available to overcome this issue; however, none directly explain the survival function prediction. In this paper, we introduce SurvSHAP(t), the first time-dependent explanation that allows for interpreting survival black-box models. It is based on SHapley Additive exPlanations with solid theoretical foundations and a broad adoption among machine learning practitioners. The proposed methods aim to enhance precision diagnostics and support domain experts in making decisions. Experiments on synthetic and medical data confirm that SurvSHAP(t) can detect variables with a time-dependent effect, and its aggregation is a better determinant of the importance of variables for a prediction than SurvLIME. SurvSHAP(t) is model-agnostic and can be applied to all models with functional output. We provide an accessible implementation of time-dependent explanations in Python at http://github.com/MI2DataLab/survshap.

Adrian Stando, Mustafa Cavus, Przemysław Biecek

Imbalanced data poses a significant challenge in classification as model performance is affected by insufficient learning from minority classes. Balancing methods are often used to address this problem. However, such techniques can lead to problems such as overfitting or loss of information. This study addresses a more challenging aspect of balancing methods - their impact on model behavior. To capture these changes, Explainable Artificial Intelligence tools are used to compare models trained on datasets before and after balancing. In addition to the variable importance method, this study uses the partial dependence profile and accumulated local effects techniques. Real and simulated datasets are tested, and an open-source Python package edgaro is developed to facilitate this analysis. The results obtained show significant changes in model behavior due to balancing methods, which can lead to biased models toward a balanced distribution. These findings confirm that balancing analysis should go beyond model performance comparisons to achieve higher reliability of machine learning models. Therefore, we propose a new method performance gain plot for informed data balancing strategy to make an optimal selection of balancing method by analyzing the measure of change in model behavior versus performance gain.

Hubert Baniecki, Przemyslaw Biecek

Explainable artificial intelligence (XAI) methods are portrayed as a remedy for debugging and trusting statistical and deep learning models, as well as interpreting their predictions. However, recent advances in adversarial machine learning (AdvML) highlight the limitations and vulnerabilities of state-of-the-art explanation methods, putting their security and trustworthiness into question. The possibility of manipulating, fooling or fairwashing evidence of the model's reasoning has detrimental consequences when applied in high-stakes decision-making and knowledge discovery. This survey provides a comprehensive overview of research concerning adversarial attacks on explanations of machine learning models, as well as fairness metrics. We introduce a unified notation and taxonomy of methods facilitating a common ground for researchers and practitioners from the intersecting research fields of AdvML and XAI. We discuss how to defend against attacks and design robust interpretation methods. We contribute a list of existing insecurities in XAI and outline the emerging research directions in adversarial XAI (AdvXAI). Future work should address improving explanation methods and evaluation protocols to take into account the reported safety issues.

Szymon Płotka, Maciej Chrabaszcz, Przemyslaw Biecek

Recent advances in Vision Transformers (ViTs) have significantly enhanced medical image segmentation by facilitating the learning of global relationships. However, these methods face a notable challenge in capturing diverse local and global long-range sequential feature representations, particularly evident in whole-body CT (WBCT) scans. To overcome this limitation, we introduce Swin Soft Mixture Transformer (Swin SMT), a novel architecture based on Swin UNETR. This model incorporates a Soft Mixture-of-Experts (Soft MoE) to effectively handle complex and diverse long-range dependencies. The use of Soft MoE allows for scaling up model parameters maintaining a balance between computational complexity and segmentation performance in both training and inference modes. We evaluate Swin SMT on the publicly available TotalSegmentator-V2 dataset, which includes 117 major anatomical structures in WBCT images. Comprehensive experimental results demonstrate that Swin SMT outperforms several state-of-the-art methods in 3D anatomical structure segmentation, achieving an average Dice Similarity Coefficient of 85.09%. The code and pre-trained weights of Swin SMT are publicly available at https://github.com/MI2DataLab/SwinSMT.

Piotr Komorowski, Hubert Baniecki, Przemysław Biecek

As deep learning models increasingly find applications in critical domains such as medical imaging, the need for transparent and trustworthy decision-making becomes paramount. Many explainability methods provide insights into how these models make predictions by attributing importance to input features. As Vision Transformer (ViT) becomes a promising alternative to convolutional neural networks for image classification, its interpretability remains an open research question. This paper investigates the performance of various interpretation methods on a ViT applied to classify chest X-ray images. We introduce the notion of evaluating faithfulness, sensitivity, and complexity of ViT explanations. The obtained results indicate that Layerwise relevance propagation for transformers outperforms Local interpretable model-agnostic explanations and Attention visualization, providing a more accurate and reliable representation of what a ViT has actually learned. Our findings provide insights into the applicability of ViT explanations in medical imaging and highlight the importance of using appropriate evaluation criteria for comparing them.

Mikołaj Spytek, Mateusz Krzyziński, Sophie Hanna Langbein et al.

Due to their flexibility and superior performance, machine learning models frequently complement and outperform traditional statistical survival models. However, their widespread adoption is hindered by a lack of user-friendly tools to explain their internal operations and prediction rationales. To tackle this issue, we introduce the survex R package, which provides a cohesive framework for explaining any survival model by applying explainable artificial intelligence techniques. The capabilities of the proposed software encompass understanding and diagnosing survival models, which can lead to their improvement. By revealing insights into the decision-making process, such as variable effects and importances, survex enables the assessment of model reliability and the detection of biases. Thus, transparency and responsibility may be promoted in sensitive areas, such as biomedical research and healthcare applications.

Mustafa Cavus, Przemysław Biecek

The expected goal provides a more representative measure of the team and player performance which also suit the low-scoring nature of football instead of score in modern football. The score of a match involves randomness and often may not represent the performance of the teams and players, therefore it has been popular to use the alternative statistics in recent years such as shots on target, ball possessions, and drills. To measure the probability of a shot being a goal by the expected goal, several features are used to train an expected goal model which is based on the event and tracking football data. The selection of these features, the size and date of the data, and the model which are used as the parameters that may affect the performance of the model. Using black-box machine learning models for increasing the predictive performance of the model decreases its interpretability that causes the loss of information that can be gathered from the model. This paper proposes an accurate expected goal model trained consisting of 315,430 shots from seven seasons between 2014-15 and 2020-21 of the top-five European football leagues. Moreover, this model is explained by using explainable artificial intelligence tool to obtain an explainable expected goal model for evaluating a team or player performance. To the best of our knowledge, this is the first paper that demonstrates a practical application of an explainable artificial intelligence tool aggregated profiles to explain a group of observations on an accurate expected goal model for monitoring the team and player performance. Moreover, these methods can be generalized to other sports branches.

Hubert Baniecki, Bartlomiej Sobieski, Patryk Szatkowski et al.

Time-to-event prediction, e.g. cancer survival analysis or hospital length of stay, is a highly prominent machine learning task in medical and healthcare applications. However, only a few interpretable machine learning methods comply with its challenges. To facilitate a comprehensive explanatory analysis of survival models, we formally introduce time-dependent feature effects and global feature importance explanations. We show how post-hoc interpretation methods allow for finding biases in AI systems predicting length of stay using a novel multi-modal dataset created from 1235 X-ray images with textual radiology reports annotated by human experts. Moreover, we evaluate cancer survival models beyond predictive performance to include the importance of multi-omics feature groups based on a large-scale benchmark comprising 11 datasets from The Cancer Genome Atlas (TCGA). Model developers can use the proposed methods to debug and improve machine learning algorithms, while physicians can discover disease biomarkers and assess their significance. We hope the contributed open data and code resources facilitate future work in the emerging research direction of explainable survival analysis.

Przemyslaw Biecek, Hubert Baniecki, Mateusz Krzyzinski et al.

The usual goal of supervised learning is to find the best model, the one that optimizes a particular performance measure. However, what if the explanation provided by this model is completely different from another model and different again from another model despite all having similarly good fit statistics? Is it possible that the equally effective models put the spotlight on different relationships in the data? Inspired by Anscombe's quartet, this paper introduces a Rashomon Quartet, i.e. a set of four models built on a synthetic dataset which have practically identical predictive performance. However, the visual exploration reveals distinct explanations of the relations in the data. This illustrative example aims to encourage the use of methods for model visualization to compare predictive models beyond their performance.

Katarzyna Kobylińska, Mateusz Krzyziński, Rafał Machowicz et al.

The machine learning modeling process conventionally culminates in selecting a single model that maximizes a selected performance metric. However, this approach leads to abandoning a more profound analysis of slightly inferior models. Particularly in medical and healthcare studies, where the objective extends beyond predictions to valuable insight generation, relying solely on a single model can result in misleading or incomplete conclusions. This problem is particularly pertinent when dealing with a set of models known as $\textit{Rashomon set}$, with performance close to maximum one. Such a set can be numerous and may contain models describing the data in a different way, which calls for comprehensive analysis. This paper introduces a novel process to explore models in the Rashomon set, extending the conventional modeling approach. We propose the $\texttt{Rashomon_DETECT}$ algorithm to detect models with different behavior. It is based on recent developments in the eXplainable Artificial Intelligence (XAI) field. To quantify differences in variable effects among models, we introduce the Profile Disparity Index (PDI) based on measures from functional data analysis. To illustrate the effectiveness of our approach, we showcase its application in predicting survival among hemophagocytic lymphohistiocytosis (HLH) patients - a foundational case study. Additionally, we benchmark our approach on other medical data sets, demonstrating its versatility and utility in various contexts. If differently behaving models are detected in the Rashomon set, their combined analysis leads to more trustworthy conclusions, which is of vital importance for high-stakes applications such as medical applications.

Katarzyna Woźnica, Piotr Wilczyński, Przemysław Biecek

Machine learning applications cover a wide range of predictive tasks in which tabular datasets play a significant role. However, although they often address similar problems, tabular datasets are typically treated as standalone tasks. The possibilities of using previously solved problems are limited due to the lack of structured contextual information about their features and the lack of understanding of the relations between them. To overcome this limitation, we propose a new approach called Semantic Feature Net (SeFNet), capturing the semantic meaning of the analyzed tabular features. By leveraging existing ontologies and domain knowledge, SeFNet opens up new opportunities for sharing insights between diverse predictive tasks. One such opportunity is the Dataset Ontology-based Semantic Similarity (DOSS) measure, which quantifies the similarity between datasets using relations across their features. In this paper, we present an example of SeFNet prepared for a collection of predictive tasks in healthcare, with the features' relations derived from the SNOMED-CT ontology. The proposed SeFNet framework and the accompanying DOSS measure address the issue of limited contextual information in tabular datasets. By incorporating domain knowledge and establishing semantic relations between features, we enhance the potential for meta-learning and enable valuable insights to be shared across different predictive tasks.

Mustafa Cavus, Adrian Stando, Przemyslaw Biecek

The expected goal models have gained popularity, but their interpretability is often limited, especially when trained using black-box methods. Explainable artificial intelligence tools have emerged to enhance model transparency and extract descriptive knowledge for a single observation or for all observations. However, explaining black-box models for a specific group of observations may be more useful in some domains. This paper introduces the glocal explanations (between local and global levels) of the expected goal models to enable performance analysis at the team and player levels by proposing the use of aggregated versions of the SHAP values and partial dependence profiles. This allows knowledge to be extracted from the expected goal model for a player or team rather than just a single shot. In addition, we conducted real-data applications to illustrate the usefulness of aggregated SHAP and aggregated profiles. The paper concludes with remarks on the potential of these explanations for performance analysis in soccer analytics.

Bastian Pfeifer, Mateusz Krzyzinski, Hubert Baniecki et al.

Explainable AI (XAI) is an increasingly important area of machine learning research, which aims to make black-box models transparent and interpretable. In this paper, we propose a novel approach to XAI that uses the so-called counterfactual paths generated by conditional permutations of features. The algorithm measures feature importance by identifying sequential permutations of features that most influence changes in model predictions. It is particularly suitable for generating explanations based on counterfactual paths in knowledge graphs incorporating domain knowledge. Counterfactual paths introduce an additional graph dimension to current XAI methods in both explaining and visualizing black-box models. Experiments with synthetic and medical data demonstrate the practical applicability of our approach.

Artur Żółkowski, Mateusz Krzyziński, Piotr Wilczyński et al.

The number of standardized policy documents regarding climate policy and their publication frequency is significantly increasing. The documents are long and tedious for manual analysis, especially for policy experts, lawmakers, and citizens who lack access or domain expertise to utilize data analytics tools. Potential consequences of such a situation include reduced citizen governance and involvement in climate policies and an overall surge in analytics costs, rendering less accessibility for the public. In this work, we use a Latent Dirichlet Allocation-based pipeline for the automatic summarization and analysis of 10-years of national energy and climate plans (NECPs) for the period from 2021 to 2030, established by 27 Member States of the European Union. We focus on analyzing policy framing, the language used to describe specific issues, to detect essential nuances in the way governments frame their climate policies and achieve climate goals. The methods leverage topic modeling and clustering for the comparative analysis of policy documents across different countries. It allows for easier integration in potential user-friendly applications for the development of theories and processes of climate policy. This would further lead to better citizen governance and engagement over climate policies and public policy research.

Hubert Baniecki, Maciej Chrabaszcz, Andreas Holzinger et al.

Evaluating explanations of image classifiers regarding ground truth, e.g. segmentation masks defined by human perception, primarily evaluates the quality of the models under consideration rather than the explanation methods themselves. Driven by this observation, we propose a framework for $\textit{jointly}$ evaluating the robustness of safety-critical systems that $\textit{combine}$ a deep neural network with an explanation method. These are increasingly used in real-world applications like medical image analysis or robotics. We introduce a fine-tuning procedure to (mis)align model$\unicode{x2013}$explanation pipelines with ground truth and use it to quantify the potential discrepancy between worst and best-case scenarios of human alignment. Experiments across various model architectures and post-hoc local interpretation methods provide insights into the robustness of vision transformers and the overall vulnerability of such AI systems to potential adversarial attacks.

Weronika Hryniewska-Guzik, Maria Kędzierska, Przemysław Biecek

Lung cancer and covid-19 have one of the highest morbidity and mortality rates in the world. For physicians, the identification of lesions is difficult in the early stages of the disease and time-consuming. Therefore, multi-task learning is an approach to extracting important features, such as lesions, from small amounts of medical data because it learns to generalize better. We propose a novel multi-task framework for classification, segmentation, reconstruction, and detection. To the best of our knowledge, we are the first ones who added detection to the multi-task solution. Additionally, we checked the possibility of using two different backbones and different loss functions in the segmentation task.

Mikolaj Spytek, Weronika Hryniewska-Guzik, Jaroslaw Zygierewicz et al.

The prediction of age is a challenging task with various practical applications in high-impact fields like the healthcare domain or criminology. Despite the growing number of models and their increasing performance, we still know little about how these models work. Numerous examples of failures of AI systems show that performance alone is insufficient, thus, new methods are needed to explore and explain the reasons for the model's predictions. In this paper, we investigate the use of Explainable Artificial Intelligence (XAI) for age prediction focusing on two specific modalities, EEG signal and lung X-rays. We share predictive models for age to facilitate further research on new techniques to explain models for these modalities.

Piotr Wilczyński, Artur Żółkowski, Mateusz Krzyziński et al.

This paper introduces HADES, a novel tool for automatic comparative documents with similar structures. HADES is designed to streamline the work of professionals dealing with large volumes of documents, such as policy documents, legal acts, and scientific papers. The tool employs a multi-step pipeline that begins with processing PDF documents using topic modeling, summarization, and analysis of the most important words for each topic. The process concludes with an interactive web app with visualizations that facilitate the comparison of the documents. HADES has the potential to significantly improve the productivity of professionals dealing with high volumes of documents, reducing the time and effort required to complete tasks related to comparative document analysis. Our package is publically available on GitHub.

Bartlomiej Sobieski, Matthew Tivnan, Dawid Płudowski et al.

Diffusion models are prone to generating structural hallucinations - samples that match the statistical properties of the training data yet defy underlying structural rules, resulting in anomalies like hands with more than five fingers. Recent research studied this failure mode from several viewpoints, offering partial explanations to their occurrence, such as mode interpolation. In this work, we propose a complementary perspective that treats hallucinations as instabilities on the model-induced manifold. We begin by showing that a hallucination filter based on such instabilities matches or exceeds the performance of the recently proposed temporal one. By tracing the source of these instabilities, we identify local intrinsic dimension (LID) as their primary driver and propose Intrinsic Quenching (IQ), a direct corrective mechanism that deflates it to alleviate hallucinations. IQ consistently outperforms standard hallucination reduction baselines across a wide array of benchmarks and offers a highly promising solution for enforcing anatomical consistency in downstream medical imaging tasks.

Przemyslaw Biecek

The data revolution has generated a huge demand for data-driven solutions. This demand propels a growing number of easy-to-use tools and training for aspiring data scientists that enable the rapid building of predictive models. Today, weapons of math destruction can be easily built and deployed without detailed planning and validation. This rapidly extends the list of AI failures, i.e. deployments that lead to financial losses or even violate democratic values such as equality, freedom and justice. The lack of planning, rules and standards around the model development leads to the ,,anarchisation of AI". This problem is reported under different names such as validation debt, reproducibility crisis, and lack of explainability. Post-mortem analysis of AI failures often reveals mistakes made in the early phase of model development or data acquisition. Thus, instead of curing the consequences of deploying harmful models, we shall prevent them as early as possible by putting more attention to the initial planning stage. In this paper, we propose a quick and simple framework to support planning of AI solutions. The POCA framework is based on four pillars: Performance, Opaqueness, Consequences, and Assumptions. It helps to set the expectations and plan the constraints for the AI solution before any model is built and any data is collected. With the help of the POCA method, preliminary requirements can be defined for the model-building process, so that costly model misspecification errors can be identified as soon as possible or even avoided. AI researchers, product owners and business analysts can use this framework in the initial stages of building AI solutions.

Vladimir Zaigrajew, Hubert Baniecki, Przemyslaw Biecek

Sparse autoencoders (SAEs) are useful for detecting and steering interpretable features in neural networks, with particular potential for understanding complex multimodal representations. Given their ability to uncover interpretable features, SAEs are particularly valuable for analyzing large-scale vision-language models (e.g., CLIP and SigLIP), which are fundamental building blocks in modern systems yet remain challenging to interpret and control. However, current SAE methods are limited by optimizing both reconstruction quality and sparsity simultaneously, as they rely on either activation suppression or rigid sparsity constraints. To this end, we introduce Matryoshka SAE (MSAE), a new architecture that learns hierarchical representations at multiple granularities simultaneously, enabling a direct optimization of both metrics without compromise. MSAE establishes a new state-of-the-art Pareto frontier between reconstruction quality and sparsity for CLIP, achieving 0.99 cosine similarity and less than 0.1 fraction of variance unexplained while maintaining ~80% sparsity. Finally, we demonstrate the utility of MSAE as a tool for interpreting and controlling CLIP by extracting over 120 semantic concepts from its representation to perform concept-based similarity search and bias analysis in downstream tasks like CelebA. We make the codebase available at https://github.com/WolodjaZ/MSAE.

Weronika Hryniewska, Adrianna Grudzień, Przemysław Biecek

The increased interest in deep learning applications, and their hard-to-detect biases result in the need to validate and explain complex models. However, current explanation methods are limited as far as both the explanation of the reasoning process and prediction results are concerned. They usually only show the location in the image that was important for model prediction. The lack of possibility to interact with explanations makes it difficult to verify and understand exactly how the model works. This creates a significant risk when using the model. The risk is compounded by the fact that explanations do not take into account the semantic meaning of the explained objects. To escape from the trap of static and meaningless explanations, we propose a tool and a process called LIMEcraft. LIMEcraft enhances the process of explanation by allowing a user to interactively select semantically consistent areas and thoroughly examine the prediction for the image instance in case of many image features. Experiments on several models show that our tool improves model safety by inspecting model fairness for image pieces that may indicate model bias. The code is available at: http://github.com/MI2DataLab/LIMEcraft

Hubert Baniecki, Dariusz Parzych, Przemyslaw Biecek

The growing need for in-depth analysis of predictive models leads to a series of new methods for explaining their local and global properties. Which of these methods is the best? It turns out that this is an ill-posed question. One cannot sufficiently explain a black-box machine learning model using a single method that gives only one perspective. Isolated explanations are prone to misunderstanding, leading to wrong or simplistic reasoning. This problem is known as the Rashomon effect and refers to diverse, even contradictory, interpretations of the same phenomenon. Surprisingly, most methods developed for explainable and responsible machine learning focus on a single-aspect of the model behavior. In contrast, we showcase the problem of explainability as an interactive and sequential analysis of a model. This paper proposes how different Explanatory Model Analysis (EMA) methods complement each other and discusses why it is essential to juxtapose them. The introduced process of Interactive EMA (IEMA) derives from the algorithmic side of explainable machine learning and aims to embrace ideas developed in cognitive sciences. We formalize the grammar of IEMA to describe potential human-model dialogues. It is implemented in a widely used human-centered open-source software framework that adopts interactivity, customizability and automation as its main traits. We conduct a user study to evaluate the usefulness of IEMA, which indicates that an interactive sequential analysis of a model increases the performance and confidence of human decision making.

Przemyslaw Biecek, Wojciech Samek

Explainable Artificial Intelligence (XAI) is a young but very promising field of research. Unfortunately, the progress in this field is currently slowed down by divergent and incompatible goals. We separate various threads tangled within the area of XAI into two complementary cultures of human/value-oriented explanations (BLUE XAI) and model/validation-oriented explanations (RED XAI). This position paper argues that the area of RED XAI is currently under-explored, i.e., more methods for explainability are desperately needed to question models (e.g., extract knowledge from well-performing models as well as spotting and fixing bugs in faulty models), and the area of RED XAI hides great opportunities and potential for important research necessary to ensure the safety of AI systems. We conclude this paper by presenting promising challenges in this area.

Mustafa Cavus, Przemysław Biecek

Predictive models may generate biased predictions when classifying imbalanced datasets. This happens when the model favors the majority class, leading to low performance in accurately predicting the minority class. To address this issue, balancing or resampling methods are critical data-centric AI approaches in the modeling process to improve prediction performance. However, there have been debates and questions about the functionality of these methods in recent years. In particular, many candidate models may exhibit very similar predictive performance, called the Rashomon effect, in model selection, and they may even produce different predictions for the same observations. Selecting one of these models without considering the predictive multiplicity -- which is the case of yielding conflicting models' predictions for any sample -- can result in blind selection. In this paper, the impact of balancing methods on predictive multiplicity is examined using the Rashomon effect. It is crucial because the blind model selection in data-centric AI is risky from a set of approximately equally accurate models. This may lead to severe problems in model selection, validation, and explanation. To tackle this matter, we conducted real dataset experiments to observe the impact of balancing methods on predictive multiplicity through the Rashomon effect by using a newly proposed metric obscurity in addition to the existing ones: ambiguity and discrepancy. Our findings showed that balancing methods inflate the predictive multiplicity and yield varying results. To monitor the trade-off between the prediction performance and predictive multiplicity for conducting the modeling process responsibly, we proposed using the extended version of the performance-gain plot when balancing the training data.

Bartlomiej Sobieski, Przemysław Biecek

Despite increasing progress in development of methods for generating visual counterfactual explanations, especially with the recent rise of Denoising Diffusion Probabilistic Models, previous works consider them as an entirely local technique. In this work, we take the first step at globalizing them. Specifically, we discover that the latent space of Diffusion Autoencoders encodes the inference process of a given classifier in the form of global directions. We propose a novel proxy-based approach that discovers two types of these directions with the use of only single image in an entirely black-box manner. Precisely, g-directions allow for flipping the decision of a given classifier on an entire dataset of images, while h-directions further increase the diversity of explanations. We refer to them in general as Global Counterfactual Directions (GCDs). Moreover, we show that GCDs can be naturally combined with Latent Integrated Gradients resulting in a new black-box attribution method, while simultaneously enhancing the understanding of counterfactual explanations. We validate our approach on existing benchmarks and show that it generalizes to real-world use-cases.

Bartlomiej Sobieski, Jakub Grzywaczewski, Bartlomiej Sadlej et al.

Visual counterfactual explanations (VCEs) have recently gained immense popularity as a tool for clarifying the decision-making process of image classifiers. This trend is largely motivated by what these explanations promise to deliver -- indicate semantically meaningful factors that change the classifier's decision. However, we argue that current state-of-the-art approaches lack a crucial component -- the region constraint -- whose absence prevents from drawing explicit conclusions, and may even lead to faulty reasoning due to phenomenons like confirmation bias. To address the issue of previous methods, which modify images in a very entangled and widely dispersed manner, we propose region-constrained VCEs (RVCEs), which assume that only a predefined image region can be modified to influence the model's prediction. To effectively sample from this subclass of VCEs, we propose Region-Constrained Counterfactual Schrödinger Bridges (RCSB), an adaptation of a tractable subclass of Schrödinger Bridges to the problem of conditional inpainting, where the conditioning signal originates from the classifier of interest. In addition to setting a new state-of-the-art by a large margin, we extend RCSB to allow for exact counterfactual reasoning, where the predefined region contains only the factor of interest, and incorporating the user to actively interact with the RVCE by predefining the regions manually.

Wiktoria Mieleszczenko-Kowszewicz, Dawid Płudowski, Filip Kołodziejczyk et al.

This study explores how the Large Language Models (LLMs) adjust linguistic features to create personalized persuasive outputs. While research showed that LLMs personalize outputs, a gap remains in understanding the linguistic features of their persuasive capabilities. We identified 13 linguistic features crucial for influencing personalities across different levels of the Big Five model of personality. We analyzed how prompts with personality trait information influenced the output of 19 LLMs across five model families. The findings show that models use more anxiety-related words for neuroticism, increase achievement-related words for conscientiousness, and employ fewer cognitive processes words for openness to experience. Some model families excel at adapting language for openness to experience, others for conscientiousness, while only one model adapts language for neuroticism. Our findings show how LLMs tailor responses based on personality cues in prompts, indicating their potential to create persuasive content affecting the mind and well-being of the recipients.

Sophie Hanna Langbein, Mateusz Krzyziński, Mikołaj Spytek et al.

With the spread and rapid advancement of black box machine learning models, the field of interpretable machine learning (IML) or explainable artificial intelligence (XAI) has become increasingly important over the last decade. This is particularly relevant for survival analysis, where the adoption of IML techniques promotes transparency, accountability and fairness in sensitive areas, such as clinical decision making processes, the development of targeted therapies, interventions or in other medical or healthcare related contexts. More specifically, explainability can uncover a survival model's potential biases and limitations and provide more mathematically sound ways to understand how and which features are influential for prediction or constitute risk factors. However, the lack of readily available IML methods may have deterred medical practitioners and policy makers in public health from leveraging the full potential of machine learning for predicting time-to-event data. We present a comprehensive review of the limited existing amount of work on IML methods for survival analysis within the context of the general IML taxonomy. In addition, we formally detail how commonly used IML methods, such as such as individual conditional expectation (ICE), partial dependence plots (PDP), accumulated local effects (ALE), different feature importance measures or Friedman's H-interaction statistics can be adapted to survival outcomes. An application of several IML methods to real data on data on under-5 year mortality of Ghanaian children from the Demographic and Health Surveys (DHS) Program serves as a tutorial or guide for researchers, on how to utilize the techniques in practice to facilitate understanding of model decisions or predictions.

Mustafa Cavus, Przemyslaw Biecek

The Rashomon effect presents a significant challenge in model selection. It occurs when multiple models achieve similar performance on a dataset but produce different predictions, resulting in predictive multiplicity. This is especially problematic in high-stakes environments, where arbitrary model outcomes can have serious consequences. Traditional model selection methods prioritize accuracy and fail to address this issue. Factors such as class imbalance and irrelevant variables further complicate the situation, making it harder for models to provide trustworthy predictions. Data-centric AI approaches can mitigate these problems by prioritizing data optimization, particularly through preprocessing techniques. However, recent studies suggest preprocessing methods may inadvertently inflate predictive multiplicity. This paper investigates how data preprocessing techniques like balancing and filtering methods impact predictive multiplicity and model stability, considering the complexity of the data. We conduct the experiments on 21 real-world datasets, applying various balancing and filtering techniques, and assess the level of predictive multiplicity introduced by these methods by leveraging the Rashomon effect. Additionally, we examine how filtering techniques reduce redundancy and enhance model generalization. The findings provide insights into the relationship between balancing methods, data complexity, and predictive multiplicity, demonstrating how data-centric AI strategies can improve model performance.

Weronika Hryniewska-Guzik, Jakub Bilski, Bartosz Chrostowski et al.

Robust and highly accurate lung segmentation in X-rays is crucial in medical imaging. This study evaluates deep learning solutions for this task, ranking existing methods and analyzing their performance under diverse image modifications. Out of 61 analyzed papers, only nine offered implementation or pre-trained models, enabling assessment of three prominent methods: Lung VAE, TransResUNet, and CE-Net. The analysis revealed that CE-Net performs best, demonstrating the highest values in dice similarity coefficient and intersection over union metric.

Krzysztof Jankowski, Bartlomiej Sobieski, Mateusz Kwiatkowski et al.

Foundation models have emerged as pivotal tools, tackling many complex tasks through pre-training on vast datasets and subsequent fine-tuning for specific applications. The Segment Anything Model is one of the first and most well-known foundation models for computer vision segmentation tasks. This work presents a multi-faceted red-teaming analysis that tests the Segment Anything Model against challenging tasks: (1) We analyze the impact of style transfer on segmentation masks, demonstrating that applying adverse weather conditions and raindrops to dashboard images of city roads significantly distorts generated masks. (2) We focus on assessing whether the model can be used for attacks on privacy, such as recognizing celebrities' faces, and show that the model possesses some undesired knowledge in this task. (3) Finally, we check how robust the model is to adversarial attacks on segmentation masks under text prompts. We not only show the effectiveness of popular white-box attacks and resistance to black-box attacks but also introduce a novel approach - Focused Iterative Gradient Attack (FIGA) that combines white-box approaches to construct an efficient attack resulting in a smaller number of modified pixels. All of our testing methods and analyses indicate a need for enhanced safety measures in foundation models for image segmentation.

Stanisław Giziński, Paulina Kaczyńska, Hubert Ruczyński et al.

There exists a growing discourse around the domination of Big Tech on the landscape of artificial intelligence (AI) research, yet our comprehension of this phenomenon remains cursory. This paper aims to broaden and deepen our understanding of Big Tech's reach and power within AI research. It highlights the dominance not merely in terms of sheer publication volume but rather in the propagation of new ideas or memes. Current studies often oversimplify the concept of influence to the share of affiliations in academic papers, typically sourced from limited databases such as arXiv or specific academic conferences. The main goal of this paper is to unravel the specific nuances of such influence, determining which AI ideas are predominantly driven by Big Tech entities. By employing network and memetic analysis on AI-oriented paper abstracts and their citation network, we are able to grasp a deeper insight into this phenomenon. By utilizing two databases: OpenAlex and S2ORC, we are able to perform such analysis on a much bigger scale than previous attempts. Our findings suggest that while Big Tech-affiliated papers are disproportionately more cited in some areas, the most cited papers are those affiliated with both Big Tech and Academia. Focusing on the most contagious memes, their attribution to specific affiliation groups (Big Tech, Academia, mixed affiliation) seems equally distributed between those three groups. This suggests that the notion of Big Tech domination over AI research is oversimplified in the discourse.

Vladimir Zaigrajew, Hubert Baniecki, Lukasz Tulczyjew et al.

Remote sensing (RS) applications in the space domain demand machine learning (ML) models that are reliable, robust, and quality-assured, making red teaming a vital approach for identifying and exposing potential flaws and biases. Since both fields advance independently, there is a notable gap in integrating red teaming strategies into RS. This paper introduces a methodology for examining ML models operating on hyperspectral images within the HYPERVIEW challenge, focusing on soil parameters' estimation. We use post-hoc explanation methods from the Explainable AI (XAI) domain to critically assess the best performing model that won the HYPERVIEW challenge and served as an inspiration for the model deployed on board the INTUITION-1 hyperspectral mission. Our approach effectively red teams the model by pinpointing and validating key shortcomings, constructing a model that achieves comparable performance using just 1% of the input features and a mere up to 5% performance loss. Additionally, we propose a novel way of visualizing explanations that integrate domain-specific information about hyperspectral bands (wavelengths) and data transformations to better suit interpreting models for hyperspectral image analysis.

Dawid Płudowski, Francesco Spinnato, Piotr Wilczyński et al.

Counterfactual explanations provide an intuitive way to understand model decisions by identifying minimal changes required to alter an outcome. However, applying counterfactual methods to time series models remains challenging due to temporal dependencies, high dimensionality, and the lack of an intuitive human-interpretable representation. We introduce MASCOTS, a method that leverages the Bag-of-Receptive-Fields representation alongside symbolic transformations inspired by Symbolic Aggregate Approximation. By operating in a symbolic feature space, it enhances interpretability while preserving fidelity to the original data and model. Unlike existing approaches that either depend on model structure or autoencoder-based sampling, MASCOTS directly generates meaningful and diverse counterfactual observations in a model-agnostic manner, operating on both univariate and multivariate data. We evaluate MASCOTS on univariate and multivariate benchmark datasets, demonstrating comparable validity, proximity, and plausibility to state-of-the-art methods, while significantly improving interpretability and sparsity. Its symbolic nature allows for explanations that can be expressed visually, in natural language, or through semantic representations, making counterfactual reasoning more accessible and actionable.

Weronika Hryniewska-Guzik, Bartosz Sawicki, Przemysław Biecek

This paper presents a comprehensive comparative analysis of explainable artificial intelligence (XAI) ensembling methods. Our research brings three significant contributions. Firstly, we introduce a novel ensembling method, NormEnsembleXAI, that leverages minimum, maximum, and average functions in conjunction with normalization techniques to enhance interpretability. Secondly, we offer insights into the strengths and weaknesses of XAI ensemble methods. Lastly, we provide a library, facilitating the practical implementation of XAI ensembling, thus promoting the adoption of transparent and interpretable deep learning models.

Weronika Hryniewska-Guzik, Luca Longo, Przemysław Biecek

Explainable Artificial Intelligence has gained significant attention due to the widespread use of complex deep learning models in high-stake domains such as medicine, finance, and autonomous cars. However, different explanations often present different aspects of the model's behavior. In this research manuscript, we explore the potential of ensembling explanations generated by deep classification models using convolutional model. Through experimentation and analysis, we aim to investigate the implications of combining explanations to uncover a more coherent and reliable patterns of the model's behavior, leading to the possibility of evaluating the representation learned by the model. With our method, we can uncover problems of under-representation of images in a certain class. Moreover, we discuss other side benefits like features' reduction by replacing the original image with its explanations resulting in the removal of some sensitive information. Through the use of carefully selected evaluation metrics from the Quantus library, we demonstrated the method's superior performance in terms of Localisation and Faithfulness, compared to individual explanations.

Paulina Tomaszewska, Przemysław Biecek

Does the stethoscope in the picture make the adjacent person a doctor or a patient? This, of course, depends on the contextual relationship of the two objects. If it's obvious, why don't explanation methods for vision models use contextual information? In this paper, we (1) review the most popular methods of explaining computer vision models by pointing out that they do not take into account context information, (2) show examples of failures of popular XAI methods, (3) provide examples of real-world use cases where spatial context plays a significant role, (4) propose new research directions that may lead to better use of context information in explaining computer vision models, (5) argue that a change in approach to explanations is needed from 'where' to 'how'.

Hubert Baniecki, Przemyslaw Biecek

A common belief is that intrinsically interpretable deep learning models ensure a correct, intuitive understanding of their behavior and offer greater robustness against accidental errors or intentional manipulation. However, these beliefs have not been comprehensively verified, and growing evidence casts doubt on them. In this paper, we highlight the risks related to overreliance and susceptibility to adversarial manipulation of these so-called "intrinsically (aka inherently) interpretable" models by design. We introduce two strategies for adversarial analysis with prototype manipulation and backdoor attacks against prototype-based networks, and discuss how concept bottleneck models defend against these attacks. Fooling the model's reasoning by exploiting its use of latent prototypes manifests the inherent uninterpretability of deep neural networks, leading to a false sense of security reinforced by a visual confirmation bias. The reported limitations of part-prototype networks put their trustworthiness and applicability into question, motivating further work on the robustness and alignment of (deep) interpretable models.

Wiktoria Mieleszczenko-Kowszewicz, Beata Bajcar, Jolanta Babiak et al.

Be careful what you ask for, you just might get it. This saying fits with the way large language models (LLMs) are trained, which, instead of being rewarded for correctness, are increasingly rewarded for pleasing the recipient. So, they are increasingly effective at persuading us that their answers are valuable. But what tricks do they use in this persuasion? In this study, we examine what are the psycholinguistic features of the responses used by twelve different language models. By grouping response content according to rational or emotional prompts and exploring social influence principles employed by LLMs, we ask whether and how we can mitigate the risks of LLM-driven mass misinformation. We position this study within the broader discourse on human-centred AI, emphasizing the need for interdisciplinary approaches to mitigate cognitive and societal risks posed by persuasive AI responses.

Mustafa Cavus, Jan N. van Rijn, Przemysław Biecek

Automated machine learning systems efficiently streamline model selection but often focus on a single best-performing model, overlooking explanation uncertainty, an essential concern in human centered explainable AI. To address this, we propose a novel framework that incorporates model multiplicity into explanation generation by aggregating partial dependence profiles (PDP) from a set of near optimal models, known as the Rashomon set. The resulting Rashomon PDP captures interpretive variability and highlights areas of disagreement, providing users with a richer, uncertainty aware view of feature effects. To evaluate its usefulness, we introduce two quantitative metrics, the coverage rate and the mean width of confidence intervals, to evaluate the consistency between the standard PDP and the proposed Rashomon PDP. Experiments on 35 regression datasets from the OpenML CTR23 benchmark suite show that in most cases, the Rashomon PDP covers less than 70% of the best model's PDP, underscoring the limitations of single model explanations. Our findings suggest that Rashomon PDP improves the reliability and trustworthiness of model interpretations by adding additional information that would otherwise be neglected. This is particularly useful in high stakes domains where transparency and confidence are critical.

Agata Kaczmarek, Dawid Płudowski, Piotr Wilczyński et al.

The "Fake or Real" competition hosted on Kaggle (https://www.kaggle.com/competitions/fake-or-real-the-impostor-hunt ) is the second part of a series of follow-up competitions and hackathons related to the "Assurance for Space Domain AI Applications" project funded by the European Space Agency (https://assurance-ai.space-codev.org/ ). The competition idea is based on two real-life AI security threats identified within the project -- data poisoning and overreliance in Large Language Models. The task is to distinguish between the proper output from LLM and the output generated under malicious modification of the LLM. As this problem was not extensively researched, participants are required to develop new techniques to address this issue or adjust already existing ones to this problem's statement.

Weronika Hryniewska-Guzik, Przemyslaw Biecek

The success of machine learning algorithms is inherently related to the extraction of meaningful features, as they play a pivotal role in the performance of these algorithms. Central to this challenge is the quality of data representation. However, the ability to generalize and extract these features effectively from unseen datasets is also crucial. In light of this, we introduce a novel concept: the polyrepresentation. Polyrepresentation integrates multiple representations of the same modality extracted from distinct sources, for example, vector embeddings from the Siamese Network, self-supervised models, and interpretable radiomic features. This approach yields better performance metrics compared to relying on a single representation. Additionally, in the context of X-ray images, we demonstrate the transferability of the created polyrepresentation to a smaller dataset, underscoring its potential as a pragmatic and resource-efficient approach in various image-related solutions. It is worth noting that the concept of polyprepresentation on the example of medical data can also be applied to other domains, showcasing its versatility and broad potential impact.

Krzysztof Kotowski, Ramez Shendy, Jakub Nalepa et al.

This competition hosted on Kaggle (https://www.kaggle.com/competitions/trojan-horse-hunt-in-space) is the first part of a series of follow-up competitions and hackathons related to the "Assurance for Space Domain AI Applications" project funded by the European Space Agency (https://assurance-ai.space-codev.org/). The competition idea is based on one of the real-life AI security threats identified within the project -- the adversarial poisoning of continuously fine-tuned satellite telemetry forecasting models. The task is to develop methods for finding and reconstructing triggers (trojans) in advanced models for satellite telemetry forecasting used in safety-critical space operations. Participants are provided with 1) a large public dataset of real-life multivariate satellite telemetry (without triggers), 2) a reference model trained on the clean data, 3) a set of poisoned neural hierarchical interpolation (N-HiTS) models for time series forecasting trained on the dataset with injected triggers, and 4) Jupyter notebook with the training pipeline and baseline algorithm (the latter will be published in the last month of the competition). The main task of the competition is to reconstruct a set of 45 triggers (i.e., short multivariate time series segments) injected into the training data of the corresponding set of 45 poisoned models. The exact characteristics (i.e., shape, amplitude, and duration) of these triggers must be identified by participants. The popular Neural Cleanse method is adopted as a baseline, but it is not designed for time series analysis and new approaches are necessary for the task. The impact of the competition is not limited to the space domain, but also to many other safety-critical applications of advanced time series analysis where model poisoning may lead to serious consequences.

Hubert Baniecki, Giuseppe Casalicchio, Bernd Bischl et al.

We discover a theoretical connection between explanation estimation and distribution compression that significantly improves the approximation of feature attributions, importance, and effects. While the exact computation of various machine learning explanations requires numerous model inferences and becomes impractical, the computational cost of approximation increases with an ever-increasing size of data and model parameters. We show that the standard i.i.d. sampling used in a broad spectrum of algorithms for post-hoc explanation leads to an approximation error worthy of improvement. To this end, we introduce Compress Then Explain (CTE), a new paradigm of sample-efficient explainability. It relies on distribution compression through kernel thinning to obtain a data sample that best approximates its marginal distribution. CTE significantly improves the accuracy and stability of explanation estimation with negligible computational overhead. It often achieves an on-par explanation approximation error 2-3x faster by using fewer samples, i.e. requiring 2-3x fewer model evaluations. CTE is a simple, yet powerful, plug-in for any explanation method that now relies on i.i.d. sampling.

Hubert Baniecki, Giuseppe Casalicchio, Bernd Bischl et al.

We study the robustness of global post-hoc explanations for predictive models trained on tabular data. Effects of predictor features in black-box supervised learning are an essential diagnostic tool for model debugging and scientific discovery in applied sciences. However, how vulnerable they are to data and model perturbations remains an open research question. We introduce several theoretical bounds for evaluating the robustness of partial dependence plots and accumulated local effects. Our experimental results with synthetic and real-world datasets quantify the gap between the best and worst-case scenarios of (mis)interpreting machine learning predictions globally.

Paulina Tomaszewska, Mateusz Sperkowski, Przemysław Biecek

The development of Artificial Intelligence for healthcare is of great importance. Models can sometimes achieve even superior performance to human experts, however, they can reason based on spurious features. This is not acceptable to the experts as it is expected that the models catch the valid patterns in the data following domain expertise. In the work, we analyse whether Deep Learning (DL) models for vision follow the histopathologists' practice so that when diagnosing a part of a lesion, they take into account also the surrounding tissues which serve as context. It turns out that the performance of DL models significantly decreases when the amount of contextual information is limited, therefore contextual information is valuable at prediction time. Moreover, we show that the models sometimes behave in an unstable way as for some images, they change the predictions many times depending on the size of the context. It may suggest that partial contextual information can be misleading.

Przemysław Bombiński, Patryk Szatkowski, Bartłomiej Sobieski et al.

Lung mask creation lacks well-defined criteria and standardized guidelines, leading to a high degree of subjectivity between annotators. In this study, we assess the underestimation of lung regions on chest X-ray segmentation masks created according to the current state-of-the-art method, by comparison with total lung volume evaluated on computed tomography (CT). We show, that lung X-ray masks created by following the contours of the heart, mediastinum, and diaphragm significantly underestimate lung regions and exclude substantial portions of the lungs from further assessment, which may result in numerous clinical errors.

Paulina Tomaszewska, Elżbieta Sienkiewicz, Mai P. Hoang et al.

We propose 'Deep spatial context' (DSCon) method, which serves for investigation of the attention-based vision models using the concept of spatial context. It was inspired by histopathologists, however, the method can be applied to various domains. The DSCon allows for a quantitative measure of the spatial context's role using three Spatial Context Measures: $SCM_{features}$, $SCM_{targets}$, $SCM_{residuals}$ to distinguish whether the spatial context is observable within the features of neighboring regions, their target values (attention scores) or residuals, respectively. It is achieved by integrating spatial regression into the pipeline. The DSCon helps to verify research questions. The experiments reveal that spatial relationships are much bigger in the case of the classification of tumor lesions than normal tissues. Moreover, it turns out that the larger the size of the neighborhood taken into account within spatial regression, the less valuable contextual information is. Furthermore, it is observed that the spatial context measure is the largest when considered within the feature space as opposed to the targets and residuals.

Weronika Hryniewska, Piotr Czarnecki, Jakub Wiśniewski et al.

Prevention is better than cure. This old truth applies not only to the prevention of diseases but also to the prevention of issues with AI models used in medicine. The source of malfunctioning of predictive models often lies not in the training process but reaches the data acquisition phase or design of the experiment phase. In this paper, we analyze in detail a single use case - a Kaggle competition related to the detection of abnormalities in X-ray lung images. We demonstrate how a series of simple tests for data imbalance exposes faults in the data acquisition and annotation process. Complex models are able to learn such artifacts and it is difficult to remove this bias during or after the training. Errors made at the data collection stage make it difficult to validate the model correctly. Based on this use case, we show how to monitor data and model balance (fairness) throughout the life cycle of a predictive model, from data acquisition to parity analysis of model scores.