Patrick Schwab

Mathieu Chevalley, Yusuf Roohani, Arash Mehrjou et al.

Causal inference is a vital aspect of multiple scientific disciplines and is routinely applied to high-impact applications such as medicine. However, evaluating the performance of causal inference methods in real-world environments is challenging due to the need for observations under both interventional and control conditions. Traditional evaluations conducted on synthetic datasets do not reflect the performance in real-world systems. To address this, we introduce CausalBench, a benchmark suite for evaluating network inference methods on real-world interventional data from large-scale single-cell perturbation experiments. CausalBench incorporates biologically-motivated performance metrics, including new distribution-based interventional metrics. A systematic evaluation of state-of-the-art causal inference methods using our CausalBench suite highlights how poor scalability of current methods limits performance. Moreover, methods that use interventional information do not outperform those that only use observational data, contrary to what is observed on synthetic benchmarks. Thus, CausalBench opens new avenues in causal network inference research and provides a principled and reliable way to track progress in leveraging real-world interventional data.

Mathieu Chevalley, Jacob Sackett-Sanders, Yusuf Roohani et al.

In drug discovery, mapping interactions between genes within cellular systems is a crucial early step. Such maps are not only foundational for understanding the molecular mechanisms underlying disease biology but also pivotal for formulating hypotheses about potential targets for new medicines. Recognizing the need to elevate the construction of these gene-gene interaction networks, especially from large-scale, real-world datasets of perturbed single cells, the CausalBench Challenge was initiated. This challenge aimed to inspire the machine learning community to enhance state-of-the-art methods, emphasizing better utilization of expansive genetic perturbation data. Using the framework provided by the CausalBench benchmark, participants were tasked with refining the current methodologies or proposing new ones. This report provides an analysis and summary of the methods submitted during the challenge to give a partial image of the state of the art at the time of the challenge. Notably, the winning solutions significantly improved performance compared to previous baselines, establishing a new state of the art for this critical task in biology and medicine.

Amin Abyaneh, Nino Scherrer, Patrick Schwab et al.

Causal discovery serves a pivotal role in mitigating model uncertainty through recovering the underlying causal mechanisms among variables. In many practical domains, such as healthcare, access to the data gathered by individual entities is limited, primarily for privacy and regulatory constraints. However, the majority of existing causal discovery methods require the data to be available in a centralized location. In response, researchers have introduced federated causal discovery. While previous federated methods consider distributed observational data, the integration of interventional data remains largely unexplored. We propose FedCDI, a federated framework for inferring causal structures from distributed data containing interventional samples. In line with the federated learning framework, FedCDI improves privacy by exchanging belief updates rather than raw samples. Additionally, it introduces a novel intervention-aware method for aggregating individual updates. We analyze scenarios with shared or disjoint intervened covariates, and mitigate the adverse effects of interventional data heterogeneity. The performance and scalability of FedCDI is rigorously tested across a variety of synthetic and real-world graphs.

Rahil Mehrizi, Arash Mehrjou, Maryana Alegro et al.

High-content cellular imaging, transcriptomics, and proteomics data provide rich and complementary views on the molecular layers of biology that influence cellular states and function. However, the biological determinants through which changes in multi-omics measurements influence cellular morphology have not yet been systematically explored, and the degree to which cell imaging could potentially enable the prediction of multi-omics directly from cell imaging data is therefore currently unclear. Here, we address the question of whether it is possible to predict bulk multi-omics measurements directly from cell images using Image2Omics - a deep learning approach that predicts multi-omics in a cell population directly from high-content images of cells stained with multiplexed fluorescent dyes. We perform an experimental evaluation in gene-edited macrophages derived from human induced pluripotent stem cells (hiPSC) under multiple stimulation conditions and demonstrate that Image2Omics achieves significantly better performance in predicting transcriptomics and proteomics measurements directly from cell images than predictions based on the mean observed training set abundance. We observed significant predictability of abundances for 4927 (18.72%; 95% CI: 6.52%, 35.52%) and 3521 (13.38%; 95% CI: 4.10%, 32.21%) transcripts out of 26137 in M1 and M2-stimulated macrophages respectively and for 422 (8.46%; 95% CI: 0.58%, 25.83%) and 697 (13.98%; 95% CI: 2.41%, 32.83%) proteins out of 4986 in M1 and M2-stimulated macrophages respectively. Our results show that some transcript and protein abundances are predictable from cell imaging and that cell imaging may potentially, in some settings and depending on the mechanisms of interest and desired performance threshold, even be a scalable and resource-efficient substitute for multi-omics measurements.

Yichen Zhou, Jonathan Golob, Amir Karimi et al.

Protein language models (pLMs) have shown strong potential for zero-shot prediction of missense variant effects, yet systematic benchmarking on viral proteins remains limited, a critical gap given the need for proactive tools that can anticipate emerging mutations ahead of experimental validation. Here we introduce ViroGym, a comprehensive benchmark evaluating pLMs across three tasks: 79 deep mutational scanning (DMS) assays covering eukaryotic viruses with 552,065 mutated sequences across 7 phenotypic readouts, 21 influenza neutralisation tasks, and a real-world pandemic prediction task for SARS-CoV-2. We benchmark well-established pLMs on fitness landscapes, antigenic diversity, and pandemic forecasting, and find that the ProGen2 family consistently achieves the strongest performance across all three tasks. Crucially, DMS and neutralisation performance reliably identifies models that generalise to real-world emergence, even though the mutation sets they surface barely overlap, revealing that complementary in vitro benchmarks capture the evolutionary constraints needed for real-world mutation forecasting.

Arash Mehrjou, Ashkan Soleymani, Annika Buchholz et al.

Federated learning (FL) has been proposed as a method to train a model on different units without exchanging data. This offers great opportunities in the healthcare sector, where large datasets are available but cannot be shared to ensure patient privacy. We systematically investigate the effectiveness of FL on the publicly available eICU dataset for predicting the survival of each ICU stay. We employ Federated Averaging as the main practical algorithm for FL and show how its performance changes by altering three key hyper-parameters, taking into account that clients can significantly vary in size. We find that in many settings, a large number of local training epochs improves the performance while at the same time reducing communication costs. Furthermore, we outline in which settings it is possible to have only a low number of hospitals participating in each federated update round. When many hospitals with low patient counts are involved, the effect of overfitting can be avoided by decreasing the batchsize. This study thus contributes toward identifying suitable settings for running distributed algorithms such as FL on clinical datasets.

Mathieu Chevalley, Arash Mehrjou, Patrick Schwab

We investigate theoretical guarantees for the false-negative rate (FNR) -- the fraction of true causal edges whose orientation is not recovered, under single-variable random interventions and an $ε$-interventional faithfulness assumption that accommodates latent confounding. For sparse Erdős--Rényi directed acyclic graphs, where the edge probability scales as $p_e = Θ(1/d)$, we show that the FNR concentrates around its mean at rate $O(\frac{\log d}{\sqrt d})$, implying that large deviations above the expected error become exponentially unlikely as dimensionality increases. This concentration ensures that derived upper bounds hold with high probability in large-scale settings. Extending the analysis to generalized Barabási--Albert graphs reveals an even stronger phenomenon: when the degree exponent satisfies $γ> 3$, the deviation width scales as $O(d^{β- \frac{1}{2}})$ with $β= 1/(γ- 1) < \frac{1}{2}$, and hence vanishes in the limit. This demonstrates that realistic scale-free topologies intrinsically regularize causal discovery, reducing variability in orientation error. These finite-dimension results provide the first dimension-adaptive, faithfulness-robust guarantees for causal structure recovery, and challenge the intuition that high dimensionality and network heterogeneity necessarily hinder accurate discovery. Our simulation results corroborate these theoretical predictions, showing that the FNR indeed concentrates and often vanishes in practice as dimensionality grows.

Arash Mehrjou, Ashkan Soleymani, Andrew Jesson et al.

In vitro cellular experimentation with genetic interventions, using for example CRISPR technologies, is an essential step in early-stage drug discovery and target validation that serves to assess initial hypotheses about causal associations between biological mechanisms and disease pathologies. With billions of potential hypotheses to test, the experimental design space for in vitro genetic experiments is extremely vast, and the available experimental capacity - even at the largest research institutions in the world - pales in relation to the size of this biological hypothesis space. Machine learning methods, such as active and reinforcement learning, could aid in optimally exploring the vast biological space by integrating prior knowledge from various information sources as well as extrapolating to yet unexplored areas of the experimental design space based on available data. However, there exist no standardised benchmarks and data sets for this challenging task and little research has been conducted in this area to date. Here, we introduce GeneDisco, a benchmark suite for evaluating active learning algorithms for experimental design in drug discovery. GeneDisco contains a curated set of multiple publicly available experimental data sets as well as open-source implementations of state-of-the-art active learning policies for experimental design and exploration.

August DuMont Schütte, Jürgen Hetzel, Sergios Gatidis et al.

Privacy concerns around sharing personally identifiable information are a major practical barrier to data sharing in medical research. However, in many cases, researchers have no interest in a particular individual's information but rather aim to derive insights at the level of cohorts. Here, we utilize Generative Adversarial Networks (GANs) to create derived medical imaging datasets consisting entirely of synthetic patient data. The synthetic images ideally have, in aggregate, similar statistical properties to those of a source dataset but do not contain sensitive personal information. We assess the quality of synthetic data generated by two GAN models for chest radiographs with 14 different radiology findings and brain computed tomography (CT) scans with six types of intracranial hemorrhages. We measure the synthetic image quality by the performance difference of predictive models trained on either the synthetic or the real dataset. We find that synthetic data performance disproportionately benefits from a reduced number of unique label combinations. Our open-source benchmark also indicates that at low number of samples per class, label overfitting effects start to dominate GAN training. We additionally conducted a reader study in which trained radiologists do not perform better than random on discriminating between synthetic and real medical images for intermediate levels of resolutions. In accordance with our benchmark results, the classification accuracy of radiologists increases at higher spatial resolution levels. Our study offers valuable guidelines and outlines practical conditions under which insights derived from synthetic medical images are similar to those that would have been derived from real imaging data. Our results indicate that synthetic data sharing may be an attractive and privacy-preserving alternative to sharing real patient-level data in the right settings.

Clare Lyle, Arash Mehrjou, Pascal Notin et al. · deepmind

The discovery of therapeutics to treat genetically-driven pathologies relies on identifying genes involved in the underlying disease mechanisms. Existing approaches search over the billions of potential interventions to maximize the expected influence on the target phenotype. However, to reduce the risk of failure in future stages of trials, practical experiment design aims to find a set of interventions that maximally change a target phenotype via diverse mechanisms. We propose DiscoBAX, a sample-efficient method for maximizing the rate of significant discoveries per experiment while simultaneously probing for a wide range of diverse mechanisms during a genomic experiment campaign. We provide theoretical guarantees of approximate optimality under standard assumptions, and conduct a comprehensive experimental evaluation covering both synthetic as well as real-world experimental design tasks. DiscoBAX outperforms existing state-of-the-art methods for experimental design, selecting effective and diverse perturbations in biological systems.

Djordje Miladinovic, Tobias Höppe, Mathieu Chevalley et al.

Data generated in perturbation experiments link perturbations to the changes they elicit and therefore contain information relevant to numerous biological discovery tasks -- from understanding the relationships between biological entities to developing therapeutics. However, these data encompass diverse perturbations and readouts, and the complex dependence of experimental outcomes on their biological context makes it challenging to integrate insights across experiments. Here, we present the Large Perturbation Model (LPM), a deep-learning model that integrates multiple, heterogeneous perturbation experiments by representing perturbation, readout, and context as disentangled dimensions. LPM outperforms existing methods across multiple biological discovery tasks, including in predicting post-perturbation transcriptomes of unseen experiments, identifying shared molecular mechanisms of action between chemical and genetic perturbations, and facilitating the inference of gene-gene interaction networks.

Frederik Träuble, Lachlan Stuart, Andreas Georgiou et al.

Understanding how molecular changes caused by genetic variation drive disease risk is crucial for deciphering disease mechanisms. However, interpreting genome sequences is challenging because of the vast size of the human genome, and because its consequences manifest across a wide range of cells, tissues and scales -- spanning from molecular to whole organism level. Here, we present Phenformer, a multi-scale genetic language model that learns to generate mechanistic hypotheses as to how differences in genome sequence lead to disease-relevant changes in expression across cell types and tissues directly from DNA sequences of up to 88 million base pairs. Using whole genome sequencing data from more than 150 000 individuals, we show that Phenformer generates mechanistic hypotheses about disease-relevant cell and tissue types that match literature better than existing state-of-the-art methods, while using only sequence data. Furthermore, disease risk predictors enriched by Phenformer show improved prediction performance and generalisation to diverse populations. Accurate multi-megabase scale interpretation of whole genomes without additional experimental data enables both a deeper understanding of molecular mechanisms involved in disease and improved disease risk prediction at the level of individuals.

Vinod Kumar Chauhan, Lei Clifton, Achille Salaün et al. · oxford

While machine learning algorithms hold promise for personalised medicine, their clinical adoption remains limited, partly due to biases that can compromise the reliability of predictions. In this paper, we focus on sample selection bias (SSB), a specific type of bias where the study population is less representative of the target population, leading to biased and potentially harmful decisions. Despite being well-known in the literature, SSB remains scarcely studied in machine learning for healthcare. Moreover, the existing machine learning techniques try to correct the bias mostly by balancing distributions between the study and the target populations, which may result in a loss of predictive performance. To address these problems, our study illustrates the potential risks associated with SSB by examining SSB's impact on the performance of machine learning algorithms. Most importantly, we propose a new research direction for addressing SSB, based on the target population identification rather than the bias correction. Specifically, we propose two independent networks(T-Net) and a multitasking network (MT-Net) for addressing SSB, where one network/task identifies the target subpopulation which is representative of the study population and the second makes predictions for the identified subpopulation. Our empirical results with synthetic and semi-synthetic datasets highlight that SSB can lead to a large drop in the performance of an algorithm for the target population as compared with the study population, as well as a substantial difference in the performance for the target subpopulations that are representative of the selected and the non-selected patients from the study population. Furthermore, our proposed techniques demonstrate robustness across various settings, including different dataset sizes, event rates, and selection rates, outperforming the existing bias correction techniques.

Mathieu Chevalley, Arash Mehrjou, Patrick Schwab

In the algorithm Intersort, Chevalley et al. (2024) proposed a score-based method to discover the causal order of variables in a Directed Acyclic Graph (DAG) model, leveraging interventional data to outperform existing methods. However, as a score-based method over the permutahedron, Intersort is computationally expensive and non-differentiable, limiting its ability to be utilised in problems involving large-scale datasets, such as those in genomics and climate models, or to be integrated into end-to-end gradient-based learning frameworks. We address this limitation by reformulating Intersort using differentiable sorting and ranking techniques. Our approach enables scalable and differentiable optimization of causal orderings, allowing the continuous score function to be incorporated as a regularizer in downstream tasks. Empirical results demonstrate that causal discovery algorithms benefit significantly from regularizing on the causal order, underscoring the effectiveness of our method. Our work opens the door to efficiently incorporating regularization for causal order into the training of differentiable models and thereby addresses a long-standing limitation of purely associational supervised learning.

Xiao Gu, Wei Tang, Jinpei Han et al. · oxford

Cardiac biosignals, such as electrocardiograms (ECG) and photoplethysmograms (PPG), are of paramount importance for the diagnosis, prevention, and management of cardiovascular diseases, and have been extensively used in a variety of clinical tasks. Conventional deep learning approaches for analyzing these signals typically rely on homogeneous datasets and static bespoke models, limiting their robustness and generalizability across diverse clinical settings and acquisition protocols. In this study, we present a cardiac sensing foundation model (CSFM) that leverages advanced transformer architectures and a generative, masked pretraining strategy to learn unified representations from vast, heterogeneous health records. Our model is pretrained on an innovative multi-modal integration of data from multiple large-scale datasets (including MIMIC-III-WDB, MIMIC-IV-ECG, and CODE), comprising cardiac signals and the corresponding clinical or machine-generated text reports from approximately 1.7 million individuals. We demonstrate that the embeddings derived from our CSFM not only serve as effective feature extractors across diverse cardiac sensing scenarios, but also enable seamless transfer learning across varying input configurations and sensor modalities. Extensive evaluations across diagnostic tasks, demographic information recognition, vital sign measurement, clinical outcome prediction, and ECG question answering reveal that CSFM consistently outperforms traditional one-modal-one-task approaches. Notably, CSFM exhibits robust performance across multiple ECG lead configurations from standard 12-lead systems to single-lead setups, and in scenarios where only ECG, only PPG, or a combination thereof is available. These findings highlight the potential of CSFM as a versatile and scalable solution, for comprehensive cardiac monitoring.

Dimitris Spathis, Aaqib Saeed, Ali Etemad et al.

This non-archival index is not complete, as some accepted papers chose to opt-out of inclusion. The list of all accepted papers is available on the workshop website.

Simon Bing, Andrea Dittadi, Stefan Bauer et al.

The widespread adoption of electronic health records (EHRs) and subsequent increased availability of longitudinal healthcare data has led to significant advances in our understanding of health and disease with direct and immediate impact on the development of new diagnostics and therapeutic treatment options. However, access to EHRs is often restricted due to their perceived sensitive nature and associated legal concerns, and the cohorts therein typically are those seen at a specific hospital or network of hospitals and therefore not representative of the wider population of patients. Here, we present HealthGen, a new approach for the conditional generation of synthetic EHRs that maintains an accurate representation of real patient characteristics, temporal information and missingness patterns. We demonstrate experimentally that HealthGen generates synthetic cohorts that are significantly more faithful to real patient EHRs than the current state-of-the-art, and that augmenting real data sets with conditionally generated cohorts of underrepresented subpopulations of patients can significantly enhance the generalisability of models derived from these data sets to different patient populations. Synthetic conditionally generated EHRs could help increase the accessibility of longitudinal healthcare data sets and improve the generalisability of inferences made from these data sets to underrepresented populations.

Nino Scherrer, Olexa Bilaniuk, Yashas Annadani et al.

Discovering causal structures from data is a challenging inference problem of fundamental importance in all areas of science. The appealing properties of neural networks have recently led to a surge of interest in differentiable neural network-based methods for learning causal structures from data. So far, differentiable causal discovery has focused on static datasets of observational or fixed interventional origin. In this work, we introduce an active intervention targeting (AIT) method which enables a quick identification of the underlying causal structure of the data-generating process. Our method significantly reduces the required number of interactions compared with random intervention targeting and is applicable for both discrete and continuous optimization formulations of learning the underlying directed acyclic graph (DAG) from data. We examine the proposed method across multiple frameworks in a wide range of settings and demonstrate superior performance on multiple benchmarks from simulated to real-world data.

Sonali Parbhoo, Stefan Bauer, Patrick Schwab

Estimating an individual's potential response to interventions from observational data is of high practical relevance for many domains, such as healthcare, public policy or economics. In this setting, it is often the case that combinations of interventions may be applied simultaneously, for example, multiple prescriptions in healthcare or different fiscal and monetary measures in economics. However, existing methods for counterfactual inference are limited to settings in which actions are not used simultaneously. Here, we present Neural Counterfactual Relation Estimation (NCoRE), a new method for learning counterfactual representations in the combination treatment setting that explicitly models cross-treatment interactions. NCoRE is based on a novel branched conditional neural representation that includes learnt treatment interaction modulators to infer the potential causal generative process underlying the combination of multiple treatments. Our experiments show that NCoRE significantly outperforms existing state-of-the-art methods for counterfactual treatment effect estimation that do not account for the effects of combining multiple treatments across several synthetic, semi-synthetic and real-world benchmarks.

Patrick Schwab, Arash Mehrjou, Sonali Parbhoo et al.

Coronavirus Disease 2019 (COVID-19) is an emerging respiratory disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) with rapid human-to-human transmission and a high case fatality rate particularly in older patients. Due to the exponential growth of infections, many healthcare systems across the world are under pressure to care for increasing amounts of at-risk patients. Given the high number of infected patients, identifying patients with the highest mortality risk early is critical to enable effective intervention and optimal prioritisation of care. Here, we present the COVID-19 Early Warning System (CovEWS), a clinical risk scoring system for assessing COVID-19 related mortality risk. CovEWS provides continuous real-time risk scores for individual patients with clinically meaningful predictive performance up to 192 hours (8 days) in advance, and is automatically derived from patients' electronic health records (EHRs) using machine learning. We trained and evaluated CovEWS using de-identified data from a cohort of 66430 COVID-19 positive patients seen at over 69 healthcare institutions in the United States (US), Australia, Malaysia and India amounting to an aggregated total of over 2863 years of patient observation time. On an external test cohort of 5005 patients, CovEWS predicts COVID-19 related mortality from $78.8\%$ ($95\%$ confidence interval [CI]: $76.0$, $84.7\%$) to $69.4\%$ ($95\%$ CI: $57.6, 75.2\%$) specificity at a sensitivity greater than $95\%$ between respectively 1 and 192 hours prior to observed mortality events - significantly outperforming existing generic and COVID-19 specific clinical risk scores. CovEWS could enable clinicians to intervene at an earlier stage, and may therefore help in preventing or mitigating COVID-19 related mortality.

Patrick Schwab, August DuMont Schütte, Benedikt Dietz et al.

Coronavirus Disease 2019 (COVID-19) is a rapidly emerging respiratory disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Due to the rapid human-to-human transmission of SARS-CoV-2, many healthcare systems are at risk of exceeding their healthcare capacities, in particular in terms of SARS-CoV-2 tests, hospital and intensive care unit (ICU) beds and mechanical ventilators. Predictive algorithms could potentially ease the strain on healthcare systems by identifying those who are most likely to receive a positive SARS-CoV-2 test, be hospitalised or admitted to the ICU. Here, we study clinical predictive models that estimate, using machine learning and based on routinely collected clinical data, which patients are likely to receive a positive SARS-CoV-2 test, require hospitalisation or intensive care. To evaluate the predictive performance of our models, we perform a retrospective evaluation on clinical and blood analysis data from a cohort of 5644 patients. Our experimental results indicate that our predictive models identify (i) patients that test positive for SARS-CoV-2 a priori at a sensitivity of 75% (95% CI: 67%, 81%) and a specificity of 49% (95% CI: 46%, 51%), (ii) SARS-CoV-2 positive patients that require hospitalisation with 0.92 AUC (95% CI: 0.81, 0.98), and (iii) SARS-CoV-2 positive patients that require critical care with 0.98 AUC (95% CI: 0.95, 1.00). In addition, we determine which clinical features are predictive to what degree for each of the aforementioned clinical tasks. Our results indicate that predictive models trained on routinely collected clinical data could be used to predict clinical pathways for COVID-19, and therefore help inform care and prioritise resources.

Patrick Schwab, Walter Karlen

Multiple sclerosis (MS) affects the central nervous system with a wide range of symptoms. MS can, for example, cause pain, changes in mood and fatigue, and may impair a person's movement, speech and visual functions. Diagnosis of MS typically involves a combination of complex clinical assessments and tests to rule out other diseases with similar symptoms. New technologies, such as smartphone monitoring in free-living conditions, could potentially aid in objectively assessing the symptoms of MS by quantifying symptom presence and intensity over long periods of time. Here, we present a deep-learning approach to diagnosing MS from smartphone-derived digital biomarkers that uses a novel combination of a multilayer perceptron with neural soft attention to improve learning of patterns in long-term smartphone monitoring data. Using data from a cohort of 774 participants, we demonstrate that our deep-learning models are able to distinguish between people with and without MS with an area under the receiver operating characteristic curve of 0.88 (95% CI: 0.70, 0.88). Our experimental results indicate that digital biomarkers derived from smartphone data could in the future be used as additional diagnostic criteria for MS.

Patrick Schwab, Walter Karlen

Feature importance estimates that inform users about the degree to which given inputs influence the output of a predictive model are crucial for understanding, validating, and interpreting machine-learning models. However, providing fast and accurate estimates of feature importance for high-dimensional data, and quantifying the uncertainty of such estimates remain open challenges. Here, we frame the task of providing explanations for the decisions of machine-learning models as a causal learning task, and train causal explanation (CXPlain) models that learn to estimate to what degree certain inputs cause outputs in another machine-learning model. CXPlain can, once trained, be used to explain the target model in little time, and enables the quantification of the uncertainty associated with its feature importance estimates via bootstrap ensembling. We present experiments that demonstrate that CXPlain is significantly more accurate and faster than existing model-agnostic methods for estimating feature importance. In addition, we confirm that the uncertainty estimates provided by CXPlain ensembles are strongly correlated with their ability to accurately estimate feature importance on held-out data.

Patrick Schwab, Lorenz Linhardt, Stefan Bauer et al.

Estimating what would be an individual's potential response to varying levels of exposure to a treatment is of high practical relevance for several important fields, such as healthcare, economics and public policy. However, existing methods for learning to estimate counterfactual outcomes from observational data are either focused on estimating average dose-response curves, or limited to settings with only two treatments that do not have an associated dosage parameter. Here, we present a novel machine-learning approach towards learning counterfactual representations for estimating individual dose-response curves for any number of treatments with continuous dosage parameters with neural networks. Building on the established potential outcomes framework, we introduce performance metrics, model selection criteria, model architectures, and open benchmarks for estimating individual dose-response curves. Our experiments show that the methods developed in this work set a new state-of-the-art in estimating individual dose-response.

Patrick Schwab, Lorenz Linhardt, Walter Karlen

Learning representations for counterfactual inference from observational data is of high practical relevance for many domains, such as healthcare, public policy and economics. Counterfactual inference enables one to answer "What if...?" questions, such as "What would be the outcome if we gave this patient treatment $t_1$?". However, current methods for training neural networks for counterfactual inference on observational data are either overly complex, limited to settings with only two available treatments, or both. Here, we present Perfect Match (PM), a method for training neural networks for counterfactual inference that is easy to implement, compatible with any architecture, does not add computational complexity or hyperparameters, and extends to any number of treatments. PM is based on the idea of augmenting samples within a minibatch with their propensity-matched nearest neighbours. Our experiments demonstrate that PM outperforms a number of more complex state-of-the-art methods in inferring counterfactual outcomes across several benchmarks, particularly in settings with many treatments.

Patrick Schwab, Walter Karlen

Parkinson's disease is a neurodegenerative disease that can affect a person's movement, speech, dexterity, and cognition. Clinicians primarily diagnose Parkinson's disease by performing a clinical assessment of symptoms. However, misdiagnoses are common. One factor that contributes to misdiagnoses is that the symptoms of Parkinson's disease may not be prominent at the time the clinical assessment is performed. Here, we present a machine-learning approach towards distinguishing between people with and without Parkinson's disease using long-term data from smartphone-based walking, voice, tapping and memory tests. We demonstrate that our attentive deep-learning models achieve significant improvements in predictive performance over strong baselines (area under the receiver operating characteristic curve = 0.85) in data from a cohort of 1853 participants. We also show that our models identify meaningful features in the input data. Our results confirm that smartphone data collected over extended periods of time could in the future potentially be used as a digital biomarker for the diagnosis of Parkinson's disease.

Patrick Schwab, Emanuela Keller, Carl Muroi et al.

Patients in the intensive care unit (ICU) require constant and close supervision. To assist clinical staff in this task, hospitals use monitoring systems that trigger audiovisual alarms if their algorithms indicate that a patient's condition may be worsening. However, current monitoring systems are extremely sensitive to movement artefacts and technical errors. As a result, they typically trigger hundreds to thousands of false alarms per patient per day - drowning the important alarms in noise and adding to the exhaustion of clinical staff. In this setting, data is abundantly available, but obtaining trustworthy annotations by experts is laborious and expensive. We frame the problem of false alarm reduction from multivariate time series as a machine-learning task and address it with a novel multitask network architecture that utilises distant supervision through multiple related auxiliary tasks in order to reduce the number of expensive labels required for training. We show that our approach leads to significant improvements over several state-of-the-art baselines on real-world ICU data and provide new insights on the importance of task selection and architectural choices in distantly supervised multitask learning.

Patrick Schwab, Djordje Miladinovic, Walter Karlen

Knowledge of the importance of input features towards decisions made by machine-learning models is essential to increase our understanding of both the models and the underlying data. Here, we present a new approach to estimating feature importance with neural networks based on the idea of distributing the features of interest among experts in an attentive mixture of experts (AME). AMEs use attentive gating networks trained with a Granger-causal objective to learn to jointly produce accurate predictions as well as estimates of feature importance in a single model. Our experiments show (i) that the feature importance estimates provided by AMEs compare favourably to those provided by state-of-the-art methods, (ii) that AMEs are significantly faster at estimating feature importance than existing methods, and (iii) that the associations discovered by AMEs are consistent with those reported by domain experts.

Patrick Schwab, Gaetano Scebba, Jia Zhang et al.

With tens of thousands of electrocardiogram (ECG) records processed by mobile cardiac event recorders every day, heart rhythm classification algorithms are an important tool for the continuous monitoring of patients at risk. We utilise an annotated dataset of 12,186 single-lead ECG recordings to build a diverse ensemble of recurrent neural networks (RNNs) that is able to distinguish between normal sinus rhythms, atrial fibrillation, other types of arrhythmia and signals that are too noisy to interpret. In order to ease learning over the temporal dimension, we introduce a novel task formulation that harnesses the natural segmentation of ECG signals into heartbeats to drastically reduce the number of time steps per sequence. Additionally, we extend our RNNs with an attention mechanism that enables us to reason about which heartbeats our RNNs focus on to make their decisions. Through the use of attention, our model maintains a high degree of interpretability, while also achieving state-of-the-art classification performance with an average F1 score of 0.79 on an unseen test set (n=3,658).