Nils Strodthoff

Markus Wenzel, Erik Grüner, Nils Strodthoff

Motivation: We explored how explainable artificial intelligence (XAI) can help to shed light into the inner workings of neural networks for protein function prediction, by extending the widely used XAI method of integrated gradients such that latent representations inside of transformer models, which were finetuned to Gene Ontology term and Enzyme Commission number prediction, can be inspected too. Results: The approach enabled us to identify amino acids in the sequences that the transformers pay particular attention to, and to show that these relevant sequence parts reflect expectations from biology and chemistry, both in the embedding layer and inside of the model, where we identified transformer heads with a statistically significant correspondence of attribution maps with ground truth sequence annotations (e.g. transmembrane regions, active sites) across many proteins. Availability and Implementation: Source code can be accessed at https://github.com/markuswenzel/xai-proteins .

Juan Miguel Lopez Alcaraz, Nils Strodthoff

The imputation of missing values represents a significant obstacle for many real-world data analysis pipelines. Here, we focus on time series data and put forward SSSD, an imputation model that relies on two emerging technologies, (conditional) diffusion models as state-of-the-art generative models and structured state space models as internal model architecture, which are particularly suited to capture long-term dependencies in time series data. We demonstrate that SSSD matches or even exceeds state-of-the-art probabilistic imputation and forecasting performance on a broad range of data sets and different missingness scenarios, including the challenging blackout-missing scenarios, where prior approaches failed to provide meaningful results.

Temesgen Mehari, Ashish Sundar, Alen Bosnjakovic et al.

Feature importance methods promise to provide a ranking of features according to importance for a given classification task. A wide range of methods exist but their rankings often disagree and they are inherently difficult to evaluate due to a lack of ground truth beyond synthetic datasets. In this work, we put feature importance methods to the test on real-world data in the domain of cardiology, where we try to distinguish three specific pathologies from healthy subjects based on ECG features comparing to features used in cardiologists' decision rules as ground truth. Some methods generally performed well and others performed poorly, while some methods did well on some but not all of the problems considered.

Maximilian Springenberg, Annika Frommholz, Markus Wenzel et al.

While machine learning is currently transforming the field of histopathology, the domain lacks a comprehensive evaluation of state-of-the-art models based on essential but complementary quality requirements beyond a mere classification accuracy. In order to fill this gap, we developed a new methodology to extensively evaluate a wide range of classification models, including recent vision transformers, and convolutional neural networks such as: ConvNeXt, ResNet (BiT), Inception, ViT and Swin transformer, with and without supervised or self-supervised pretraining. We thoroughly tested the models on five widely used histopathology datasets containing whole slide images of breast, gastric, and colorectal cancer and developed a novel approach using an image-to-image translation model to assess the robustness of a cancer classification model against stain variations. Further, we extended existing interpretability methods to previously unstudied models and systematically reveal insights of the models' classifications strategies that can be transferred to future model architectures.

Juan Miguel Lopez Alcaraz, Nils Strodthoff

Synthetic data generation is a promising solution to address privacy issues with the distribution of sensitive health data. Recently, diffusion models have set new standards for generative models for different data modalities. Also very recently, structured state space models emerged as a powerful modeling paradigm to capture long-term dependencies in time series. We put forward SSSD-ECG, as the combination of these two technologies, for the generation of synthetic 12-lead electrocardiograms conditioned on more than 70 ECG statements. Due to a lack of reliable baselines, we also propose conditional variants of two state-of-the-art unconditional generative models. We thoroughly evaluate the quality of the generated samples, by evaluating pretrained classifiers on the generated data and by evaluating the performance of a classifier trained only on synthetic data, where SSSD-ECG clearly outperforms its GAN-based competitors. We demonstrate the soundness of our approach through further experiments, including conditional class interpolation and a clinical Turing test demonstrating the high quality of the SSSD-ECG samples across a wide range of conditions.

Temesgen Mehari, Nils Strodthoff

Deep learning has emerged as the preferred modeling approach for automatic ECG analysis. In this study, we investigate three elements aimed at improving the quantitative accuracy of such systems. These components consistently enhance performance beyond the existing state-of-the-art, which is predominantly based on convolutional models. Firstly, we explore more expressive architectures by exploiting structured state space models (SSMs). These models have shown promise in capturing long-term dependencies in time series data. By incorporating SSMs into our approach, we not only achieve better performance, but also gain insights into long-standing questions in the field. Specifically, for standard diagnostic tasks, we find no advantage in using higher sampling rates such as 500Hz compared to 100Hz. Similarly, extending the input size of the model beyond 3 seconds does not lead to significant improvements. Secondly, we demonstrate that self-supervised learning using contrastive predictive coding can further improve the performance of SSMs. By leveraging self-supervision, we enable the model to learn more robust and representative features, leading to improved analysis accuracy. Lastly, we depart from synthetic benchmarking scenarios and incorporate basic demographic metadata alongside the ECG signal as input. This inclusion of patient metadata departs from the conventional practice of relying solely on the signal itself. Remarkably, this addition consistently yields positive effects on predictive performance. We firmly believe that all three components should be considered when developing next-generation ECG analysis algorithms.

Johanna Vielhaben, Stefan Blücher, Nils Strodthoff

The completeness axiom renders the explanation of a post-hoc XAI method only locally faithful to the model, i.e. for a single decision. For the trustworthy application of XAI, in particular for high-stake decisions, a more global model understanding is required. Recently, concept-based methods have been proposed, which are however not guaranteed to be bound to the actual model reasoning. To circumvent this problem, we propose Multi-dimensional Concept Discovery (MCD) as an extension of previous approaches that fulfills a completeness relation on the level of concepts. Our method starts from general linear subspaces as concepts and does neither require reinforcing concept interpretability nor re-training of model parts. We propose sparse subspace clustering to discover improved concepts and fully leverage the potential of multi-dimensional subspaces. MCD offers two complementary analysis tools for concepts in input space: (1) concept activation maps, that show where a concept is expressed within a sample, allowing for concept characterization through prototypical samples, and (2) concept relevance heatmaps, that decompose the model decision into concept contributions. Both tools together enable a detailed understanding of the model reasoning, which is guaranteed to relate to the model via a completeness relation. This paves the way towards more trustworthy concept-based XAI. We empirically demonstrate the superiority of MCD against more constrained concept definitions.

Urs Hackstein, Jordi Alastruey, Philip Aston et al.

This report is part of the Qumphy project (22HLT01 Qumphy) that is funded by the European Union and is dedicated to the development of measures to quantify the uncertainties associated with Machine Learning algorithms applied to medical problems, in particular the analysis and processing of Photoplethysmography (PPG) signals. In this report, a list of six medical problems that are related to PPG signals and serve as Benchmark Problems is given. Suitable Benchmark datasets and their usage are described also.

Temesgen Mehari, Nils Strodthoff

The field of deep-learning-based ECG analysis has been largely dominated by convolutional architectures. This work explores the prospects of applying the recently introduced structured state space models (SSMs) as a particularly promising approach due to its ability to capture long-term dependencies in time series. We demonstrate that this approach leads to significant improvements over the current state-of-the-art for ECG classification, which we trace back to individual pathologies. Furthermore, the model's ability to capture long-term dependencies allows to shed light on long-standing questions in the literature such as the optimal sampling rate or window size to train classification models. Interestingly, we find no evidence for using data sampled at 500Hz as opposed to 100Hz and no advantages from extending the model's input size beyond 3s. Based on this very promising first assessment, SSMs could develop into a new modeling paradigm for ECG analysis.

Juan Miguel Lopez Alcaraz, Hjalmar Bouma, Nils Strodthoff

Background: AI-driven prediction algorithms have the potential to enhance emergency medicine by enabling rapid and accurate decision-making regarding patient status and potential deterioration. However, the integration of multimodal data, including raw waveform signals, remains underexplored in clinical decision support. Methods: We present a dataset and benchmarking protocol designed to advance multimodal decision support in emergency care. Our models utilize demographics, biometrics, vital signs, laboratory values, and electrocardiogram (ECG) waveforms as inputs to predict both discharge diagnoses and patient deterioration. Results: The diagnostic model achieves area under the receiver operating curve (AUROC) scores above 0.8 for 609 out of 1,428 conditions, covering both cardiac (e.g., myocardial infarction) and non-cardiac (e.g., renal disease, diabetes) diagnoses. The deterioration model attains AUROC scores above 0.8 for 14 out of 15 targets, accurately predicting critical events such as cardiac arrest, mechanical ventilation, ICU admission, and mortality. Conclusions: Our study highlights the positive impact of incorporating raw waveform data into decision support models, improving predictive performance. By introducing a unique, publicly available dataset and baseline models, we provide a foundation for measurable progress in AI-driven decision support for emergency care.

Gabriel Ott, Yannik Schaubelt, Juan Miguel Lopez Alcaraz et al.

Cardiovascular diseases remain the leading global cause of mortality. Age is an important covariate whose effect is most easily investigated in a healthy cohort to properly distinguish the former from disease-related changes. Traditionally, most of such insights have been drawn from the analysis of electrocardiogram (ECG) feature changes in individuals as they age. However, these features, while informative, may potentially obscure underlying data relationships. In this paper we present the following contributions: (1) We employ a deep-learning model and a tree-based model to analyze ECG data from a robust dataset of healthy individuals across varying ages in both raw signals and ECG feature format. (2) We use explainable AI methods to identify the most discriminative ECG features across age groups.(3) Our analysis with tree-based classifiers reveals age-related declines in inferred breathing rates and identifies notably high SDANN values as indicative of elderly individuals, distinguishing them from younger adults. (4) Furthermore, the deep-learning model underscores the pivotal role of the P-wave in age predictions across all age groups, suggesting potential changes in the distribution of different P-wave types with age. These findings shed new light on age-related ECG changes, offering insights that transcend traditional feature-based approaches.

Johanna Vielhaben, Stefan Blücher, Nils Strodthoff

Concepts are key building blocks of higher level human understanding. Explainable AI (XAI) methods have shown tremendous progress in recent years, however, local attribution methods do not allow to identify coherent model behavior across samples and therefore miss this essential component. In this work, we study concept-based explanations and put forward a new definition of concepts as low-dimensional subspaces of hidden feature layers. We novelly apply sparse subspace clustering to discover these concept subspaces. Moving forward, we derive insights from concept subspaces in terms of localized input (concept) maps, show how to quantify concept relevances and lastly, evaluate similarities and transferability between concepts. We empirically demonstrate the soundness of the proposed Sparse Subspace Clustering for Concept Discovery (SSCCD) method for a variety of different image classification tasks. This approach allows for deeper insights into the actual model behavior that would remain hidden from conventional input-level heatmaps.

Juan Miguel Lopez Alcaraz, Nils Strodthoff

Laboratory value represents a cornerstone of medical diagnostics, but suffers from slow turnaround times, and high costs and only provides information about a single point in time. The continuous estimation of laboratory values from non-invasive data such as electrocardiogram (ECG) would therefore mark a significant frontier in healthcare monitoring. Despite its potential, this domain remains relatively underexplored. In this preliminary study, we used a publicly available dataset (MIMIC-IV-ECG) to investigate the feasibility of inferring laboratory values from ECG features and patient demographics using tree-based models (XGBoost). We define the prediction task as a binary problem of whether the lab value falls into low or high abnormalities. We assessed model performance with AUROC. Our findings demonstrate promising results in the estimation of laboratory values related to different organ systems. While further research and validation are warranted to fully assess the clinical utility and generalizability of the approach, our findings lay the groundwork for future investigations for laboratory value estimation using ECG data. Such advancements hold promise for revolutionizing predictive healthcare applications, offering faster, non-invasive, and more affordable means of patient monitoring.

Juan Miguel Lopez Alcaraz, Nils Strodthoff

Ensuring timely and accurate diagnosis of medical conditions is paramount for effective patient care. Electrocardiogram (ECG) signals are fundamental for evaluating a patient's cardiac health and are readily available. Despite this, little attention has been given to the remarkable potential of ECG data in detecting non-cardiac conditions. In our study, we used publicly available datasets (MIMIC-IV-ECG-ICD and ECG-VIEW II) to investigate the feasibility of inferring general diagnostic conditions from ECG features. To this end, we trained a tree-based model (XGBoost) based on ECG features and basic demographic features to estimate a wide range of diagnoses, encompassing both cardiac and non-cardiac conditions. Our results demonstrate the reliability of estimating 23 cardiac as well as 21 non-cardiac conditions above 0.7 AUROC in a statistically significant manner across a wide range of physiological categories. Our findings underscore the predictive potential of ECG data in identifying well-known cardiac conditions. However, even more striking, this research represents a pioneering effort in systematically expanding the scope of ECG-based diagnosis to conditions not traditionally associated with the cardiac system.

Mohammad Moulaeifard, Ciaran Bench, Philip J. Aston et al.

Uncertainty quantification (UQ) is critical for safety-critical domains like healthcare, yet it is rarely evaluated under realistic out-of-distribution (OOD) conditions. Here, we assessed predictive performance and uncertainty reliability for deep learning-based blood pressure (BP) estimation from photoplethysmography (PPG) signals under both in-distribution (ID) and OOD settings. Using an XResNet1D-50 trained on PulseDB and tested on four external datasets, we compared deep ensembles (DE) and Monte Carlo dropout (MCD) with Gaussian negative log-likelihood (GNLL) and mean squared error (MSE) losses, optionally followed by post-hoc recalibration via conformal prediction (CP), temperature scaling (TS), and isotonic regression (IR). The key findings of our study are as follows: (1) DE provides stronger predictive robustness under domain shift than MCD, an advantage that becomes clear primarily under external shift. (2) Recalibrated GNLL-based methods yield the best uncertainty calibration (e.g., GNLL+DE+CP for systolic blood pressure (SBP), GNLL+DE+TS for diastolic blood pressure (DBP)), while MSE-based uncertainty requires recalibration to become practically useful. (3) Across settings, CP and TS offer the most consistent gains, with IR remaining competitive in several cases. Overall, our results identify DE-based methods as most robust for predictive performance under domain shift, GNLL as strongest for native UQ, and recalibration as essential for making MSE-based uncertainty practical. These findings highlight the need to jointly assess predictive accuracy and calibration on external data for trustworthy cuffless BP estimation

Tiezhi Wang, Nils Strodthoff

Scoring sleep stages in polysomnography recordings is a time-consuming task plagued by significant inter-rater variability. Therefore, it stands to benefit from the application of machine learning algorithms. While many algorithms have been proposed for this purpose, certain critical architectural decisions have not received systematic exploration. In this study, we meticulously investigate these design choices within the broad category of encoder-predictor architectures. We identify robust architectures applicable to both time series and spectrogram input representations. These architectures incorporate structured state space models as integral components and achieve statistically significant performance improvements compared to state-of-the-art approaches on the extensive Sleep Heart Health Study dataset. We anticipate that the architectural insights gained from this study along with the refined methodology for architecture search demonstrated herein will not only prove valuable for future research in sleep staging but also hold relevance for other time series annotation tasks.

M A Al-Masud, Nils Strodthoff

Specialized foundation models are beginning to emerge in various medical subdomains, but pretraining methodologies and parametric scaling with the size of the pretraining dataset are rarely assessed systematically and in a like-for-like manner. This work focuses on foundation models for electrocardiography (ECG) data, one of the most widely captured physiological time series world-wide. We present a comprehensive assessment of pretraining methodologies, covering five different contrastive and non-contrastive self-supervised learning objectives for ECG foundation models, and investigate their scaling behavior with pretraining dataset sizes up to 11M input samples, exclusively from publicly available sources. Pretraining strategy has a meaningful and consistent impact on downstream performance, with contrastive predictive coding (slightly ahead of JEPA) yielding the most transferable representations across diverse clinical tasks. Scaling pretraining data continues to yield meaningful improvements up to 11M samples for most objectives. We also compare model architectures across all pretraining methodologies and find evidence for a clear superiority of structured state space models compared to transformers and CNN models. We hypothesize that the strong inductive biases of structured state space models, rather than pretraining scale alone, are the primary driver of effective ECG representation learning, with important implications for future foundation model development in this and potentially other physiological signal domains.

Elias B. Krey, Nils Neukirch, Nils Strodthoff

Intermediate feature representations represent the backbone for the expressivity and adaptability of deep neural networks. However, their geometric structure remains poorly understood. In this submission, we provide indirect insights into this matter by applying a broad selection of manipulations in input space, ranging from geometric and photometric transformations to local masking and semantic manipulations using generative image editing models, and assess the feasibility of learning a mapping in the feature space, mapping from the original to the manipulated feature map. To this end, we devise different types of mappings, from linear to non-linear and local to global mappings and assess both the reconstruction quality of the mapping as well as the semantic content of the mapped representations. We demonstrate the feasibility of learning such mappings for all considered transformations. While global (transformer) models that operate on the full feature map often achieve best results, we show that the same can be achieved with a shared linear model operating on a single feature vector typically with very little degradation in reconstruction quality, even for highly non-trivial semantic manipulations. We analyze the corresponding mappings across different feature layers and characterize them according to dominance of weight vs. bias and the effective rank of the linear transformations. These results provide hints for the hypothesis that the feature space is to a first degree of approximation organized in linear structures. From a broader perspective, the study demonstrates that generative image editing models might open the door to a deeper understanding of the feature space through input manipulation.

Ciaran Bench, Oskar Pfeffer, Vivek Desai et al.

In principle, deep learning models trained on medical time-series, including wearable photoplethysmography (PPG) sensor data, can provide a means to continuously monitor physiological parameters outside of clinical settings. However, there is considerable risk of poor performance when deployed in practical measurement scenarios leading to negative patient outcomes. Reliable uncertainties accompanying predictions can provide guidance to clinicians in their interpretation of the trustworthiness of model outputs. It is therefore of interest to compare the effectiveness of different approaches. Here we implement an unprecedented set of eight uncertainty quantification (UQ) techniques to models trained on two clinically relevant prediction tasks: Atrial Fibrillation (AF) detection (classification), and two variants of blood pressure regression. We formulate a comprehensive evaluation procedure to enable a rigorous comparison of these approaches. We observe a complex picture of uncertainty reliability across the different techniques, where the most optimal for a given task depends on the chosen expression of uncertainty, evaluation metric, and scale of reliability assessed. We find that assessing local calibration and adaptivity provides practically relevant insights about model behaviour that otherwise cannot be acquired using more commonly implemented global reliability metrics. We emphasise that criteria for evaluating UQ techniques should cater to the model's practical use case, where the use of a small number of measurements per patient places a premium on achieving small-scale reliability for the chosen expression of uncertainty, while preserving as much predictive performance as possible.

Mohammad Moulaeifard, Philip J. Aston, Peter H. Charlton et al.

Photoplethysmography (PPG) is one of the most widely captured biosignals for clinical prediction tasks, yet PPG-based algorithms are typically trained on small-scale datasets of uncertain quality, which hinders meaningful algorithm comparisons. We present a comprehensive benchmark for PPG-based clinical prediction using the \dbname~dataset, establishing baselines across the full spectrum of clinically relevant applications: multi-class heart rhythm classification, and regression of physiological parameters including respiratory rate (RR), heart rate (HR), and blood pressure (BP). Most notably, we provide the first comprehensive assessment of PPG for general arrhythmia detection beyond atrial fibrillation (AF) and atrial flutter (AFLT), with performance stratified by BP, HR, and demographic subgroups. Using established deep learning architectures, we achieved strong performance for AF detection (AUROC = 0.96) and accurate physiological parameter estimation (RR MAE: 2.97 bpm; HR MAE: 1.13 bpm; SBP/DBP MAE: 16.13/8.70 mmHg). Cross-dataset validation demonstrates excellent generalizability for AF detection (AUROC = 0.97), while clinical subgroup analysis reveals marked performance differences across subgroups by BP, HR, and demographic strata. These variations appear to reflect population-specific waveform differences rather than systematic bias in model behavior. This framework establishes the first integrated benchmark for multi-task PPG-based clinical prediction, demonstrating that PPG signals can effectively support multiple simultaneous monitoring tasks and providing essential baselines for future algorithm development.

Steffen Meinert, Philipp Schlinge, Nils Strodthoff et al.

XAI gained considerable importance in recent years. Methods based on prototypical case-based reasoning have shown a promising improvement in explainability. However, these methods typically rely on additional post-hoc saliency techniques to explain the semantics of learned prototypes. Multiple critiques have been raised about the reliability and quality of such techniques. For this reason, we study the use of prominent image segmentation foundation models to improve the truthfulness of the mapping between embedding and input space. We aim to restrict the computation area of the saliency map to a predefined semantic image patch to reduce the uncertainty of such visualizations. To perceive the information of an entire image, we use the bounding box from each generated segmentation mask to crop the image. Each mask results in an individual input in our novel model architecture named ProtoMask. We conduct experiments on three popular fine-grained classification datasets with a wide set of metrics, providing a detailed overview on explainability characteristics. The comparison with other popular models demonstrates competitive performance and unique explainability features of our model. https://github.com/uos-sis/quanproto

Stefan Blücher, Johanna Vielhaben, Nils Strodthoff

Feature removal is a central building block for eXplainable AI (XAI), both for occlusion-based explanations (Shapley values) as well as their evaluation (pixel flipping, PF). However, occlusion strategies can vary significantly from simple mean replacement up to inpainting with state-of-the-art diffusion models. This ambiguity limits the usefulness of occlusion-based approaches. For example, PF benchmarks lead to contradicting rankings. This is amplified by competing PF measures: Features are either removed starting with most influential first (MIF) or least influential first (LIF). This study proposes two complementary perspectives to resolve this disagreement problem. Firstly, we address the common criticism of occlusion-based XAI, that artificial samples lead to unreliable model evaluations. We propose to measure the reliability by the R(eference)-Out-of-Model-Scope (OMS) score. The R-OMS score enables a systematic comparison of occlusion strategies and resolves the disagreement problem by grouping consistent PF rankings. Secondly, we show that the insightfulness of MIF and LIF is conversely dependent on the R-OMS score. To leverage this, we combine the MIF and LIF measures into the symmetric relevance gain (SRG) measure. This breaks the inherent connection to the underlying occlusion strategy and leads to consistent rankings. This resolves the disagreement problem, which we verify for a set of 40 different occlusion strategies.

Johanna Vielhaben, Dilyara Bareeva, Jim Berend et al.

Vision transformers (ViTs) can be trained using various learning paradigms, from fully supervised to self-supervised. Diverse training protocols often result in significantly different feature spaces, which are usually compared through alignment analysis. However, current alignment measures quantify this relationship in terms of a single scalar value, obscuring the distinctions between common and unique features in pairs of representations that share the same scalar alignment. We address this limitation by combining alignment analysis with concept discovery, which enables a breakdown of alignment into single concepts encoded in feature space. This fine-grained comparison reveals both universal and unique concepts across different representations, as well as the internal structure of concepts within each of them. Our methodological contributions address two key prerequisites for concept-based alignment: 1) For a description of the representation in terms of concepts that faithfully capture the geometry of the feature space, we define concepts as the most general structure they can possibly form - arbitrary manifolds, allowing hidden features to be described by their proximity to these manifolds. 2) To measure distances between concept proximity scores of two representations, we use a generalized Rand index and partition it for alignment between pairs of concepts. We confirm the superiority of our novel concept definition for alignment analysis over existing linear baselines in a sanity check. The concept-based alignment analysis of representations from four different ViTs reveals that increased supervision correlates with a reduction in the semantic structure of learned representations.

Juan Miguel Lopez Alcaraz, Nils Strodthoff

This study investigates the feasibility of using electrocardiogram (ECG) data combined with basic patient metadata to estimate and monitor prompt laboratory abnormalities. We use the MIMIC-IV dataset to train multimodal deep learning models on ECG waveforms, demographics, biometrics, and vital signs. Our model is a structured state space classifier with late fusion for metadata. We frame the task as individual binary classifications per abnormality and evaluate performance using AUROC. The models achieve strong performance, with AUROCs above 0.70 for 24 lab values in abnormality prediction and up to 24 in abnormality forecasting, across cardiac, renal, hematological, metabolic, immunological, and coagulation categories. NTproBNP (>353 pg/mL) is best predicted (AUROC > 0.90). Other values with AUROC > 0.85 include Hemoglobin (>17.5 g/dL), Albumin (>5.2 g/dL), and Hematocrit (>51%). Our findings show ECG combined with clinical data enables prompt abnormality prediction and forecasting of lab abnormalities, offering a non-invasive, cost-effective alternative to traditional testing. This can support early intervention and enhanced patient monitoring. ECG and clinical data can help estimate and monitor abnormal lab values, potentially improving care while reducing reliance on invasive and costly procedures.

Mohammad Moulaeifard, Peter H. Charlton, Nils Strodthoff

Photoplethysmography (PPG)-based blood pressure (BP) estimation represents a promising alternative to cuff-based BP measurements. Recently, an increasing number of deep learning models have been proposed to infer BP from the raw PPG waveform. However, these models have been predominantly evaluated on in-distribution test sets, which immediately raises the question of the generalizability of these models to external datasets. To investigate this question, we trained five deep learning models on the recently released PulseDB dataset, provided in-distribution benchmarking results on this dataset, and then assessed out-of-distribution performance on several external datasets. The best model (XResNet1d101) achieved in-distribution MAEs of 9.4 and 6.0 mmHg for systolic and diastolic BP respectively on PulseDB (with subject-specific calibration), and 14.0 and 8.5 mmHg respectively without calibration. Equivalent MAEs on external test datasets without calibration ranged from 15.0 to 25.1 mmHg (SBP) and 7.0 to 10.4 mmHg (DBP). Our results indicate that the performance is strongly influenced by the differences in BP distributions between datasets. We investigated a simple way of improving performance through sample-based domain adaptation and put forward recommendations for training models with good generalization properties. With this work, we hope to educate more researchers for the importance and challenges of out-of-distribution generalization.

Juan Miguel Lopez Alcaraz, Wilhelm Haverkamp, Nils Strodthoff

Background: Neoplasms are a major cause of mortality globally, where early diagnosis is essential for improving outcomes. Current diagnostic methods are often invasive, expensive, and inaccessible in resource-limited settings. This study explores the potential of electrocardiogram (ECG) data, a widely available and non-invasive tool for diagnosing neoplasms through cardiovascular changes linked to neoplastic presence. Methods: A diagnostic pipeline combining tree-based machine learning models with Shapley value analysis for explainability was developed. The model was trained and internally validated on a large dataset and externally validated on an independent cohort to ensure robustness and generalizability. Key ECG features contributing to predictions were identified and analyzed. Results: The model achieved high diagnostic accuracy in both internal testing and external validation cohorts. Shapley value analysis highlighted significant ECG features, including novel predictors. The approach is cost-effective, scalable, and suitable for resource-limited settings, offering insights into cardiovascular changes associated with neoplasms and their therapies. Conclusions: This study demonstrates the feasibility of using ECG signals and machine learning for non-invasive neoplasm diagnosis. By providing interpretable insights into cardio-neoplasm interactions, this method addresses gaps in diagnostics and supports integration into broader diagnostic and therapeutic frameworks.

Juan Miguel Lopez Alcaraz, Wilhelm Haverkamp, Nils Strodthoff

Background: Liver diseases present a significant global health challenge and often require costly, invasive diagnostics. Electrocardiography (ECG), a widely available and non-invasive tool, can enable the detection of liver disease by capturing cardiovascular-hepatic interactions. Methods: We trained tree-based machine learning models on ECG features to detect liver diseases using two large datasets: MIMIC-IV-ECG (467,729 patients, 2008-2019) and ECG-View II (775,535 patients, 1994-2013). The task was framed as binary classification, with performance evaluated via the area under the receiver operating characteristic curve (AUROC). To improve interpretability, we applied explainability methods to identify key predictive features. Findings: The models showed strong predictive performance with good generalizability. For example, AUROCs for alcoholic liver disease (K70) were 0.8025 (95% confidence interval (CI), 0.8020-0.8035) internally and 0.7644 (95% CI, 0.7641-0.7649) externally; for hepatic failure (K72), scores were 0.7404 (95% CI, 0.7389-0.7415) and 0.7498 (95% CI, 0.7494-0.7509), respectively. The explainability analysis consistently identified age and prolonged QTc intervals (corrected QT, reflecting ventricular repolarization) as key predictors. Features linked to autonomic regulation and electrical conduction abnormalities were also prominent, supporting known cardiovascular-liver connections and suggesting QTc as a potential biomarker. Interpretation: ECG-based machine learning offers a promising, interpretable approach for liver disease detection, particularly in resource-limited settings. By revealing clinically relevant biomarkers, this method supports non-invasive diagnostics, early detection, and risk stratification prior to targeted clinical assessments.

Juan Miguel Lopez Alcaraz, Nils Strodthoff

Despite the excelling performance of machine learning models, understanding their decisions remains a long-standing goal. Although commonly used attribution methods from explainable AI attempt to address this issue, they typically rely on associational rather than causal relationships. In this study, within the context of time series classification, we introduce a novel model-agnostic attribution method to assess the causal effect of concepts i.e., predefined segments within a time series, on classification outcomes. Our approach compares these causal attributions with closely related associational attributions, both theoretically and empirically. To estimate counterfactual outcomes, we use state-of-the-art diffusion models backed by state space models. We demonstrate the insights gained by our approach for a diverse set of qualitatively different time series classification tasks. Although causal and associational attributions might often share some similarities, in all cases they differ in important details, underscoring the risks associated with drawing causal conclusions from associational data alone. We believe that the proposed approach is also widely applicable in other domains to shed some light on the limits of associational attributions.

Tiezhi Wang, Nils Strodthoff

This study aims to elucidate the significance of long-range correlations for deep-learning-based sleep staging. It is centered around S4Sleep(TS), a recently proposed model for automated sleep staging. This model utilizes electroencephalography (EEG) as raw time series input and relies on structured state space sequence (S4) models as essential model component. Although the model already surpasses state-of-the-art methods for a moderate number of 15 input epochs, recent literature results suggest potential benefits from incorporating very long correlations spanning hundreds of input epochs. In this submission, we explore the possibility of achieving further enhancements by systematically scaling up the model's input size, anticipating potential improvements in prediction accuracy. In contrast to findings in literature, our results demonstrate that augmenting the input size does not yield a significant enhancement in the performance of S4Sleep(TS). These findings, coupled with the distinctive ability of S4 models to capture long-range dependencies in time series data, cast doubt on the diagnostic relevance of very long-range interactions for sleep staging.

Mohammad Moulaeifard, Loic Coquelin, Mantas Rinkevičius et al.

Photoplethysmography (PPG) is a widely used non-invasive physiological sensing technique, suitable for various clinical applications. Such clinical applications are increasingly supported by machine learning methods, raising the question of the most appropriate input representation and model choice. Comprehensive comparisons, in particular across different input representations, are scarce. We address this gap in the research landscape by a comprehensive benchmarking study covering three kinds of input representations, interpretable features, image representations and raw waveforms, across prototypical regression and classification use cases: blood pressure and atrial fibrillation prediction. In both cases, the best results are achieved by deep neural networks operating on raw time series as input representations. Within this model class, best results are achieved by modern convolutional neural networks (CNNs). but depending on the task setup, shallow CNNs are often also very competitive. We envision that these results will be insightful for researchers to guide their choice on machine learning tasks for PPG data, even beyond the use cases presented in this work.

M A Al-Masud, Juan Miguel Lopez Alcaraz, Nils Strodthoff

The 12-lead electrocardiogram (ECG) is a long-standing diagnostic tool. Yet machine learning for ECG interpretation remains fragmented, often limited to narrow tasks or datasets. Foundation models promise broader adaptability, but their generalization across diverse ECG tasks is not well understood. We benchmarked eight ECG foundation models on 26 clinically relevant tasks using 12 public datasets comprising 1,650 regression and classification targets. Models were evaluated under fine-tuning and frozen settings, with scaling analyses across dataset sizes. Results show heterogeneous performance across domains: in the most widely studied domain, adult ECG interpretation, three foundation models consistently outperformed strong supervised baselines. In contrast, ECG-CPC, a compact structured state-space model pretrained on HEEDB, dominated other categories where most foundation models failed to surpass supervised learning. Foundation models also displayed distinct scaling behaviors with dataset size, which are critical for small-scale clinical applications. Overall, while foundation models show promise for adult ECG analysis, substantial gaps remain in cardiac structure, outcome prediction, and patient characterization. Notably, ECG-CPC's strong performance despite being orders of magnitude smaller and consuming minimal computational resources highlights untapped opportunities for advancing ECG foundation models.

Nils Neukirch, Johanna Vielhaben, Nils Strodthoff

Internal representations are crucial for understanding deep neural networks, such as their properties and reasoning patterns, but remain difficult to interpret. While mapping from feature space to input space aids in interpreting the former, existing approaches often rely on crude approximations. We propose using a conditional diffusion model - a pretrained high-fidelity diffusion model conditioned on spatially resolved feature maps - to learn such a mapping in a probabilistic manner. We demonstrate the feasibility of this approach across various pretrained image classifiers from CNNs to ViTs, showing excellent reconstruction capabilities. Through qualitative comparisons and robustness analysis, we validate our method and showcase possible applications, such as the visualization of concept steering in input space or investigations of the composite nature of the feature space. This approach has broad potential for improving feature space understanding in computer vision models.

Zahra Mansour, Verena Uslar, Dirk Weyhe et al.

The development of electronic stethoscopes and wearable recording sensors opened the door to the automated analysis of bowel sound (BS) signals. This enables a data-driven analysis of bowel sound patterns, their interrelations, and their correlation to different pathologies. This work leverages a BS dataset collected from 16 healthy subjects that was annotated according to four established BS patterns. This dataset is used to evaluate the performance of machine learning models to detect and/or classify BS patterns. The selection of considered models covers models using tabular features, convolutional neural networks based on spectrograms and models pre-trained on large audio datasets. The results highlight the clear superiority of pre-trained models, particularly in detecting classes with few samples, achieving an AUC of 0.89 in distinguishing BS from non-BS using a HuBERT model and an AUC of 0.89 in differentiating bowel sound patterns using a Wav2Vec 2.0 model. These results pave the way for an improved understanding of bowel sounds in general and future machine-learning-driven diagnostic applications for gastrointestinal examinations

Nils Strodthoff, Juan Miguel Lopez Alcaraz, Wilhelm Haverkamp

Current deep learning algorithms designed for automatic ECG analysis have exhibited notable accuracy. However, akin to traditional electrocardiography, they tend to be narrowly focused and typically address a singular diagnostic condition. In this exploratory study, we specifically investigate the capability of a single model to predict a diverse range of both cardiac and non-cardiac discharge diagnoses based on a sole ECG collected in the emergency department. We find that 253, 81 cardiac, and 172 non-cardiac, ICD codes can be reliably predicted in the sense of exceeding an AUROC score of 0.8 in a statistically significant manner. This underscores the model's proficiency in handling a wide array of cardiac and non-cardiac diagnostic scenarios which demonstrates potential as a screening tool for diverse medical encounters.

Zahra Mansour, Verena Uslar, Dirk Weyhe et al.

Bowel sounds (BS) are typically momentary and have low amplitude, making them difficult to detect accurately through manual auscultation. This leads to significant variability in clinical assessment. Digital acoustic sensors allow the acquisition of high-quality BS and enable automated signal analysis, offering the potential to provide clinicians with both objective and quantitative feedback on bowel activity. This study presents an automated pipeline for bowel sound segmentation and classification using a wearable acoustic SonicGuard sensor. BS signals from 83 subjects were recorded using a SonicGuard sensor. Data from 40 subjects were manually annotated by clinical experts and used to train an automatic annotation algorithm, while the remaining subjects were used for further model evaluation. An energy-based event detection algorithm was developed to detect BS events. Detected sound segments were then classified into BS patterns using a pretrained Audio Spectrogram Transformer (AST) model. Model performance was evaluated separately for healthy individuals and patients. The best configuration used two specialized models, one trained on healthy subjects and one on patients, achieving (accuracy: 0.97, AUROC: 0.98) for healthy group and (accuracy: 0.96, AUROC: 0.98) for patient group. The auto-annotation method reduced manual labeling time by approximately 70%, and expert review showed that less than 12% of automatically detected segments required correction. The proposed automated segmentation and classification system enables quantitative assessment of bowel activity, providing clinicians with an objective diagnostic tool that may improve the diagnostic of gastrointestinal function and support the annotation of large-scale datasets.

Richard Koebe, Noah Saibel, Juan Miguel Lopez Alcaraz et al.

Hypotension in critically ill ICU patients is common and life-threatening. Escalation to catecholamine therapy marks a key management step, with both undertreatment and overtreatment posing risks. Most machine learning (ML) models predict hypotension using fixed MAP thresholds or MAP forecasting, overlooking the clinical decision behind treatment escalation. Predicting catecholamine initiation, the start of vasoactive or inotropic agent administration offers a more clinically actionable target reflecting real decision-making. Using the MIMIC-III database, we modeled catecholamine initiation as a binary event within a 15-minute prediction window. Input features included statistical descriptors from a two-hour sliding MAP context window, along with demographics, biometrics, comorbidities, and ongoing treatments. An Extreme Gradient Boosting (XGBoost) model was trained and interpreted via SHapley Additive exPlanations (SHAP). The model achieved an AUROC of 0.822 (0.813-0.830), outperforming the hypotension baseline (MAP < 65, AUROC 0.686 [0.675-0.699]). SHAP analysis highlighted recent MAP values, MAP trends, and ongoing treatments (e.g., sedatives, electrolytes) as dominant predictors. Subgroup analysis showed higher performance in males, younger patients (<53 years), those with higher BMI (>32), and patients without comorbidities or concurrent medications. Predicting catecholamine initiation based on MAP dynamics, treatment context, and patient characteristics supports the critical decision of when to escalate therapy, shifting focus from threshold-based alarms to actionable decision support. This approach is feasible across a broad ICU cohort under natural event imbalance. Future work should enrich temporal and physiological context, extend label definitions to include therapy escalation, and benchmark against existing hypotension prediction systems.

Lara Ahrens, Wilhelm Haverkamp, Nils Strodthoff

Domain-adapted open-weight large language models (LLMs) offer promising healthcare applications, from queryable knowledge bases to multimodal assistants, with the crucial advantage of local deployment for privacy preservation. However, optimal adaptation strategies, evaluation methodologies, and performance relative to general-purpose LLMs remain poorly characterized. We investigated these questions in electrocardiography, an important area of cardiovascular medicine, by finetuning open-weight models on domain-specific literature and implementing a multi-layered evaluation framework comparing finetuned models, retrieval-augmented generation (RAG), and Claude Sonnet 3.7 as a representative general-purpose model. Finetuned Llama 3.1 70B achieved superior performance on multiple-choice evaluations and automatic text metrics, ranking second to Claude 3.7 in LLM-as-a-judge assessments. Human expert evaluation favored Claude 3.7 and RAG approaches for complex queries. Finetuned models significantly outperformed their base counterparts across nearly all evaluation modes. Our findings reveal substantial performance heterogeneity across evaluation methodologies, underscoring assessment complexity. Nevertheless, domain-specific adaptation through finetuning and RAG achieves competitive performance with proprietary models, supporting the viability of privacy-preserving, locally deployable clinical solutions.

Tiezhi Wang, Wilhelm Haverkamp, Nils Strodthoff

The electrocardiogram (ECG) exemplifies biosignal-based time series with continuous, temporally ordered structure reflecting cardiac physiological and pathophysiological dynamics. Detailed analysis of these dynamics has proven challenging, as conventional methods capture either global trends or local waveform features but rarely their simultaneous interplay at high temporal resolution. To bridge global and local signal analysis, we introduce S4ECG, a novel deep learning architecture leveraging structured state space models for multi-epoch arrhythmia classification. Our joint multi-epoch predictions significantly outperform single-epoch approaches by 1.0-11.6% in macro-AUROC, with atrial fibrillation specificity improving from 0.718-0.979 to 0.967-0.998, demonstrating superior performance in-distribution and enhanced out-of-distribution robustness. Systematic investigation reveals optimal temporal dependency windows spanning 10-20 minutes for peak performance. This work contributes to a paradigm shift toward temporally-aware arrhythmia detection algorithms, opening new possibilities for ECG interpretation, in particular for complex arrhythmias like atrial fibrillation and atrial flutter.

Julia Matejas, Olaf Żurawski, Nils Strodthoff et al.

Purpose: Chest X-rays are essential for diagnosing pulmonary conditions, but limited access in resource-constrained settings can delay timely diagnosis. Electrocardiograms (ECGs), in contrast, are widely available, non-invasive, and often acquired earlier in clinical workflows. This study aims to assess whether ECG features and patient demographics can predict chest radiograph findings using an interpretable machine learning approach. Methods: Using the MIMIC-IV database, Extreme Gradient Boosting (XGBoost) classifiers were trained to predict diverse chest radiograph findings from ECG-derived features and demographic variables. Recursive feature elimination was performed independently for each target to identify the most predictive features. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC) with bootstrapped 95% confidence intervals. Shapley Additive Explanations (SHAP) were applied to interpret feature contributions. Results: Models successfully predicted multiple chest radiograph findings with varying accuracy. Feature selection tailored predictors to each target, and including demographic variables consistently improved performance. SHAP analysis revealed clinically meaningful contributions from ECG features to radiographic predictions. Conclusion: ECG-derived features combined with patient demographics can serve as a proxy for certain chest radiograph findings, enabling early triage or pre-screening in settings where radiographic imaging is limited. Interpretable machine learning demonstrates potential to support radiology workflows and improve patient care.

Juan Miguel Lopez Alcaraz, Ebenezer Oloyede, David Taylor et al.

Background: Electrocardiogram (ECG) analysis has emerged as a promising tool for detecting physiological changes linked to non-cardiac disorders. Given the close connection between cardiovascular and neurocognitive health, ECG abnormalities may be present in individuals with co-occurring neurocognitive conditions. This highlights the potential of ECG as a biomarker to improve detection, therapy monitoring, and risk stratification in patients with neurocognitive disorders, an area that remains underexplored. Methods: We aim to demonstrate the feasibility to predict neurocognitive disorders from ECG features across diverse patient populations. We utilized ECG features and demographic data to predict neurocognitive disorders defined by ICD-10 codes, focusing on dementia, delirium, and Parkinson's disease. Internal and external validations were performed using the MIMIC-IV and ECG-View datasets. Predictive performance was assessed using AUROC scores, and Shapley values were used to interpret feature contributions. Results: Significant predictive performance was observed for disorders within the neurcognitive disorders. Significantly, the disorders with the highest predictive performance is F03: Dementia, with an internal AUROC of 0.848 (95% CI: 0.848-0.848) and an external AUROC of 0.865 (0.864-0.965), followed by G30: Alzheimer's, with an internal AUROC of 0.809 (95% CI: 0.808-0.810) and an external AUROC of 0.863 (95% CI: 0.863-0.864). Feature importance analysis revealed both known and novel ECG correlates. ECGs hold promise as non-invasive, explainable biomarkers for selected neurocognitive disorders. This study demonstrates robust performance across cohorts and lays the groundwork for future clinical applications, including early detection and personalized monitoring.

Patrick Wagner, Temesgen Mehari, Wilhelm Haverkamp et al.

Deep neural networks have become increasingly popular for analyzing ECG data because of their ability to accurately identify cardiac conditions and hidden clinical factors. However, the lack of transparency due to the black box nature of these models is a common concern. To address this issue, explainable AI (XAI) methods can be employed. In this study, we present a comprehensive analysis of post-hoc XAI methods, investigating the local (attributions per sample) and global (based on domain expert concepts) perspectives. We have established a set of sanity checks to identify sensible attribution methods, and we provide quantitative evidence in accordance with expert rules. This dataset-wide analysis goes beyond anecdotal evidence by aggregating data across patient subgroups. Furthermore, we demonstrate how these XAI techniques can be utilized for knowledge discovery, such as identifying subtypes of myocardial infarction. We believe that these proposed methods can serve as building blocks for a complementary assessment of the internal validity during a certification process, as well as for knowledge discovery in the field of ECG analysis.

Vignesh Srinivasan, Nils Strodthoff, Jackie Ma et al.

There is an increasing number of medical use-cases where classification algorithms based on deep neural networks reach performance levels that are competitive with human medical experts. To alleviate the challenges of small dataset sizes, these systems often rely on pretraining. In this work, we aim to assess the broader implications of these approaches. For diabetic retinopathy grading as exemplary use case, we compare the impact of different training procedures including recently established self-supervised pretraining methods based on contrastive learning. To this end, we investigate different aspects such as quantitative performance, statistics of the learned feature representations, interpretability and robustness to image distortions. Our results indicate that models initialized from ImageNet pretraining report a significant increase in performance, generalization and robustness to image distortions. In particular, self-supervised models show further benefits to supervised models. Self-supervised models with initialization from ImageNet pretraining not only report higher performance, they also reduce overfitting to large lesions along with improvements in taking into account minute lesions indicative of the progression of the disease. Understanding the effects of pretraining in a broader sense that goes beyond simple performance comparisons is of crucial importance for the broader medical imaging community beyond the use-case considered in this work.

Johanna Vielhaben, Markus Wenzel, Eva Weicken et al.

Predicting the binding of viral peptides to the major histocompatibility complex with machine learning can potentially extend the computational immunology toolkit for vaccine development, and serve as a key component in the fight against a pandemic. In this work, we adapt and extend USMPep, a recently proposed, conceptually simple prediction algorithm based on recurrent neural networks. Most notably, we combine regressors (binding affinity data) and classifiers (mass spectrometry data) from qualitatively different data sources to obtain a more comprehensive prediction tool. We evaluate the performance on a recently released SARS-CoV-2 dataset with binding stability measurements. USMPep not only sets new benchmarks on selected single alleles, but consistently turns out to be among the best-performing methods or, for some metrics, to be even the overall best-performing method for this task.

Temesgen Mehari, Nils Strodthoff

Clinical 12-lead electrocardiography (ECG) is one of the most widely encountered kinds of biosignals. Despite the increased availability of public ECG datasets, label scarcity remains a central challenge in the field. Self-supervised learning represents a promising way to alleviate this issue. In this work, we put forward the first comprehensive assessment of self-supervised representation learning from clinical 12-lead ECG data. To this end, we adapt state-of-the-art self-supervised methods based on instance discrimination and latent forecasting to the ECG domain. In a first step, we learn contrastive representations and evaluate their quality based on linear evaluation performance on a recently established, comprehensive, clinical ECG classification task. In a second step, we analyze the impact of self-supervised pretraining on finetuned ECG classifiers as compared to purely supervised performance. For the best-performing method, an adaptation of contrastive predictive coding, we find a linear evaluation performance only 0.5% below supervised performance. For the finetuned models, we find improvements in downstream performance of roughly 1% compared to supervised performance, label efficiency, as well as robustness against physiological noise. This work clearly establishes the feasibility of extracting discriminative representations from ECG data via self-supervised learning and the numerous advantages when finetuning such representations on downstream tasks as compared to purely supervised training. As first comprehensive assessment of its kind in the ECG domain carried out exclusively on publicly available datasets, we hope to establish a first step towards reproducible progress in the rapidly evolving field of representation learning for biosignals.

Stefan Blücher, Johanna Vielhaben, Nils Strodthoff

PredDiff is a model-agnostic, local attribution method that is firmly rooted in probability theory. Its simple intuition is to measure prediction changes while marginalizing features. In this work, we clarify properties of PredDiff and its close connection to Shapley values. We stress important differences between classification and regression, which require a specific treatment within both formalisms. We extend PredDiff by introducing a new, well-founded measure for interaction effects between arbitrary feature subsets. The study of interaction effects represents an inevitable step towards a comprehensive understanding of black-box models and is particularly important for science applications. Equipped with our novel interaction measure, PredDiff is a promising model-agnostic approach for obtaining reliable, numerically inexpensive and theoretically sound attributions.

Johanna Vielhaben, Nils Strodthoff

Generative neural samplers offer a complementary approach to Monte Carlo methods for problems in statistical physics and quantum field theory. This work tests the ability of generative neural samplers to estimate observables for real-world low-dimensional spin systems. It maps out how autoregressive models can sample configurations of a quantum Heisenberg chain via a classical approximation based on the Suzuki-Trotter transformation. We present results for energy, specific heat and susceptibility for the isotropic XXX and the anisotropic XY chain that are in good agreement with Monte Carlo results within the same approximation scheme.

Nils Strodthoff, Claas Strodthoff, Tobias Becher et al.

Electrical impedance tomography (EIT) is a noninvasive imaging modality that allows a continuous assessment of changes in regional bioimpedance of different organs. One of its most common biomedical applications is monitoring regional ventilation distribution in critically ill patients treated in intensive care units. In this work, we put forward a proof-of-principle study that demonstrates how one can reconstruct synchronously measured respiratory or circulatory parameters from the EIT image sequence using a deep learning model trained in an end-to-end fashion. We demonstrate that one can accurately infer absolute volume, absolute flow, normalized airway pressure and within certain limitations even the normalized arterial blood pressure from the EIT signal alone, in a way that generalizes to unseen patients without prior calibration. As an outlook with direct clinical relevance, we furthermore demonstrate the feasibility of reconstructing the absolute transpulmonary pressure from a combination of EIT and absolute airway pressure, as a way to potentially replace the invasive measurement of esophageal pressure. With these results, we hope to stimulate further studies building on the framework put forward in this work.

Nils Strodthoff, Patrick Wagner, Tobias Schaeffter et al.

Electrocardiography is a very common, non-invasive diagnostic procedure and its interpretation is increasingly supported by automatic interpretation algorithms. The progress in the field of automatic ECG interpretation has up to now been hampered by a lack of appropriate datasets for training as well as a lack of well-defined evaluation procedures to ensure comparability of different algorithms. To alleviate these issues, we put forward first benchmarking results for the recently published, freely accessible PTB-XL dataset, covering a variety of tasks from different ECG statement prediction tasks over age and gender prediction to signal quality assessment. We find that convolutional neural networks, in particular resnet- and inception-based architectures, show the strongest performance across all tasks outperforming feature-based algorithms by a large margin. These results are complemented by deeper insights into the classification algorithm in terms of hidden stratification, model uncertainty and an exploratory interpretability analysis. We also put forward benchmarking results for the ICBEB2018 challenge ECG dataset and discuss prospects of transfer learning using classifiers pretrained on PTB-XL. With this resource, we aim to establish the PTB-XL dataset as a resource for structured benchmarking of ECG analysis algorithms and encourage other researchers in the field to join these efforts.

Stefan Bluecher, Lukas Kades, Jan M. Pawlowski et al.

Machine learning has the potential to aid our understanding of phase structures in lattice quantum field theories through the statistical analysis of Monte Carlo samples. Available algorithms, in particular those based on deep learning, often demonstrate remarkable performance in the search for previously unidentified features, but tend to lack transparency if applied naively. To address these shortcomings, we propose representation learning in combination with interpretability methods as a framework for the identification of observables. More specifically, we investigate action parameter regression as a pretext task while using layer-wise relevance propagation (LRP) to identify the most important observables depending on the location in the phase diagram. The approach is put to work in the context of a scalar Yukawa model in (2+1)d. First, we investigate a multilayer perceptron to determine an importance hierarchy of several predefined, standard observables. The method is then applied directly to the raw field configurations using a convolutional network, demonstrating the ability to reconstruct all order parameters from the learned filter weights. Based on our results, we argue that due to its broad applicability, attribution methods such as LRP could prove a useful and versatile tool in our search for new physical insights. In the case of the Yukawa model, it facilitates the construction of an observable that characterises the symmetric phase.

Kim A. Nicoli, Shinichi Nakajima, Nils Strodthoff et al.

We propose a general framework for the estimation of observables with generative neural samplers focusing on modern deep generative neural networks that provide an exact sampling probability. In this framework, we present asymptotically unbiased estimators for generic observables, including those that explicitly depend on the partition function such as free energy or entropy, and derive corresponding variance estimators. We demonstrate their practical applicability by numerical experiments for the 2d Ising model which highlight the superiority over existing methods. Our approach greatly enhances the applicability of generative neural samplers to real-world physical systems.