As Large Language Models (LLMs) are increasingly integrated in diverse applications, obtaining reliable measures of their predictive uncertainty has become critically important. A precise distinction between aleatoric uncertainty, arising from inherent ambiguities within input data, and epistemic uncertainty, originating exclusively from model limitations, is essential to effectively address each uncertainty source. In this paper, we introduce Spectral Uncertainty, a novel approach to quantifying and decomposing uncertainties in LLMs. Leveraging the Von Neumann entropy from quantum information theory, Spectral Uncertainty provides a rigorous theoretical foundation for separating total uncertainty into distinct aleatoric and epistemic components. Unlike existing baseline methods, our approach incorporates a fine-grained representation of semantic similarity, enabling nuanced differentiation among various semantic interpretations in model responses. Empirical evaluations demonstrate that Spectral Uncertainty outperforms state-of-the-art methods in estimating both aleatoric and total uncertainty across diverse models and benchmark datasets.

We study the problem of monitoring machine learning models under gradual distribution shifts, where circumstances change slowly over time, often leading to unnoticed yet significant declines in accuracy. To address this, we propose Incremental Uncertainty-aware Performance Monitoring (IUPM), a novel label-free method that estimates performance changes by modeling gradual shifts using optimal transport. In addition, IUPM quantifies the uncertainty in the performance prediction and introduces an active labeling procedure to restore a reliable estimate under a limited labeling budget. Our experiments show that IUPM outperforms existing performance estimation baselines in various gradual shift scenarios and that its uncertainty awareness guides label acquisition more effectively compared to other strategies.

Shortcut learning, i.e., a model's reliance on undesired features not directly relevant to the task, is a major challenge that severely limits the applications of machine learning algorithms, particularly when deploying them to assist in making sensitive decisions, such as in medical diagnostics. In this work, we leverage recent advancements in machine learning to create an unsupervised framework that is capable of both detecting and mitigating shortcut learning in transformers. We validate our method on multiple datasets. Results demonstrate that our framework significantly improves both worst-group accuracy (samples misclassified due to shortcuts) and average accuracy, while minimizing human annotation effort. Moreover, we demonstrate that the detected shortcuts are meaningful and informative to human experts, and that our framework is computationally efficient, allowing it to be run on consumer hardware.

Given the ability to model more realistic and dynamic problems, Federated Continual Learning (FCL) has been increasingly investigated recently. A well-known problem encountered in this setting is the so-called catastrophic forgetting, for which the learning model is inclined to focus on more recent tasks while forgetting the previously learned knowledge. The majority of the current approaches in FCL propose generative-based solutions to solve said problem. However, this setting requires multiple training epochs over the data, implying an offline setting where datasets are stored locally and remain unchanged over time. Furthermore, the proposed solutions are tailored for vision tasks solely. To overcome these limitations, we propose a new approach to deal with different modalities in the online scenario where new data arrive in streams of mini-batches that can only be processed once. To solve catastrophic forgetting, we propose an uncertainty-aware memory-based approach. Specifically, we suggest using an estimator based on the Bregman Information (BI) to compute the model's variance at the sample level. Through measures of predictive uncertainty, we retrieve samples with specific characteristics, and – by retraining the model on such samples – we demonstrate the potential of this approach to reduce the forgetting effect in realistic settings while maintaining data confidentiality and competitive communication efficiency compared to state-of-the-art approaches.

Perturbation-based explanations are widely utilized to enhance the transparency of machine-learning models in practice. However, their reliability is often compromised by the unknown model behavior under the specific perturbations used. This paper investigates the relationship between uncertainty calibration - the alignment of model confidence with actual accuracy - and perturbation-based explanations. We show that models systematically produce unreliable probability estimates when subjected to explainability-specific perturbations and theoretically prove that this directly undermines global and local explanation quality. To address this, we introduce ReCalX, a novel approach to recalibrate models for improved explanations while preserving their original predictions. Empirical evaluations across diverse models and datasets demonstrate that ReCalX consistently reduces perturbation-specific miscalibration most effectively while enhancing explanation robustness and the identification of globally important input features.

Single-cell transcriptomics offers critical insights into the molecular mechanisms of heart failure (HF) with reduced or preserved ejection fraction. However, understanding these mechanisms is hindered by the growing complexity of single-cell data and the difficulty in unmasking meaningful differential gene signatures among HF types. Machine learning, particularly deep neural networks (NNs), address these challenges by learning transcriptional patterns, reconstructing expression profiles and effectively classifying cells but often lacks interpretability. Recent advances in explainable AI (XAI) offer tools to clarify model decisions. Yet pinpointing differentially regulated genes with these tools remains challenging. We introduce a novel method to identify differentially explained genes (DXGs) based on importance scores derived from custom-built NNs. We highlight the superiority of DXGs in identifying HF subtypes-specific pathways that provide new insights into different types of HF. Offering a robust foundation for future research and therapeutic exploration in expanding transcriptome atlases.

Current deep learning-based solutions for image analysis tasks are commonly incapable of handling problems to which multiple different plausible solutions exist. In response, posterior-based methods such as conditional Diffusion Models and Invertible Neural Networks have emerged; however, their translation is hampered by a lack of research on adequate validation. In other words, the way progress is measured often does not reflect the needs of the driving practical application. Closing this gap in the literature, we present the first systematic framework for the application-driven validation of posterior-based methods in inverse problems. As a methodological novelty, it adopts key principles from the field of object detection validation, which has a long history of addressing the question of how to locate and match multiple object instances in an image. Treating modes as instances enables us to perform mode-centric validation, using well-interpretable metrics from the application perspective. We demonstrate the value of our framework through instantiations for a synthetic toy example and two medical vision use cases: pose estimation in surgery and imaging-based quantification of functional tissue parameters for diagnostics. Our framework offers key advantages over common approaches to posterior validation in all three examples and could thus revolutionize performance assessment in inverse problems.

Proper scoring rules evaluate the quality of probabilistic predictions, playing an essential role in the pursuit of accurate and well-calibrated models. Every proper score decomposes into two fundamental components – proper calibration error and refinement – utilizing a Bregman divergence. While uncertainty calibration has gained significant attention, current literature lacks a general estimator for these quantities with known statistical properties. To address this gap, we propose a method that allows consistent, and asymptotically unbiased estimation of all proper calibration errors and refinement terms. In particular, we introduce Kullback-Leibler calibration error, induced by the commonly used cross-entropy loss. As part of our results, we prove the relation between refinement and f-divergences, which implies information monotonicity in neural networks, regardless of which proper scoring rule is optimized. Our experiments validate empirically the claimed properties of the proposed estimator and suggest that the selection of a post-hoc calibration method should be determined by the particular calibration error of interest.

To ensure the trustworthiness and interpretability of AI systems, it is essential to align machine learning models with human domain knowledge. This can be a challenging and time-consuming endeavor that requires close communication between data scientists and domain experts. Recent leaps in the capabilities of Large Language Models (LLMs) can help alleviate this burden. In this paper, we propose a Mixture of Rule Experts guided by a Large Language Model (MoRE-LLM) which combines a data-driven black-box model with knowledge extracted from an LLM to enable domain knowledge-aligned and transparent predictions. While the introduced Mixture of Rule Experts (MoRE) steers the discovery of local rule-based surrogates during training and their utilization for the classification task, the LLM is responsible for enhancing the domain knowledge alignment of the rules by correcting and contextualizing them. Importantly, our method does not rely on access to the LLM during test time and ensures interpretability while not being prone to LLM-based confabulations. We evaluate our method on several tabular data sets and compare its performance with interpretable and non-interpretable baselines. Besides performance, we evaluate our grey-box method with respect to the utilization of interpretable rules. In addition to our quantitative evaluation, we shed light on how the LLM can provide additional context to strengthen the comprehensibility and trustworthiness of the model's reasoning process.

Using feature attributions for post-hoc explanations is a common practice to understand and verify the predictions of opaque machine learning models. Despite the numerous techniques available, individual methods often produce inconsistent and unstable results, putting their overall reliability into question. In this work, we aim to systematically improve the quality of feature attributions by combining multiple explanations across distinct methods or their variations. For this purpose, we propose a novel approach to derive optimal convex combinations of feature attributions that yield provable improvements of desired quality criteria such as robustness or faithfulness to the model behavior. Through extensive experiments involving various model architectures and popular feature attribution techniques, we demonstrate that our combination strategy consistently outperforms individual methods and existing baselines.

Generative models, like large language models, are becoming increasingly relevant in our daily lives, yet a theoretical framework to assess their generalization behavior and uncertainty does not exist. Particularly, the problem of uncertainty estimation is commonly solved in an ad-hoc and task-dependent manner. For example, natural language approaches cannot be transferred to image generation. In this paper, we introduce the first bias-variance-covariance decomposition for kernel scores. This decomposition represents a theoretical framework from which we derive a kernel-based variance and entropy for uncertainty estimation. We propose unbiased and consistent estimators for each quantity which only require generated samples but not the underlying model itself. Based on the wide applicability of kernels, we demonstrate our framework via generalization and uncertainty experiments for image, audio, and language generation. Specifically, kernel entropy for uncertainty estimation is more predictive of performance on CoQA and TriviaQA question answering datasets than existing baselines and can also be applied to closed-source models.

Monitoring and maintaining machine learning models are among the most critical challenges in translating recent advances in the field into real-world applications. However, current monitoring methods lack the capability of provide actionable insights answering the question of why the performance of a particular model really degraded. In this work, we propose a novel approach to explain the behavior of a black-box model under feature shifts by attributing an estimated performance change to interpretable input characteristics. We refer to our method that combines concepts from Optimal Transport and Shapley Values as Explanatory Performance Estimation (XPE). We analyze the underlying assumptions and demonstrate the superiority of our approach over several baselines on different data sets across various data modalities such as images, audio, and tabular data. We also indicate how the generated results can lead to valuable insights, enabling explanatory model monitoring by revealing potential root causes for model deterioration and guiding toward actionable countermeasures.

Communicating the predictive uncertainty of deep neural networks transparently and reliably is important in many safety-critical applications such as medicine. However, modern neural networks tend to be poorly calibrated, resulting in wrong predictions made with a high confidence. While existing post-hoc calibration methods like temperature scaling or isotonic regression yield strongly calibrated predictions in artificial experimental settings, their efficiency can significantly reduce in real-world applications, where scarcity of labeled data or domain drifts are commonly present. In this paper, we first investigate the impact of these characteristics on post-hoc calibration and introduce an easy-to-implement extension of common post-hoc calibration methods based on test time augmentation. In extensive experiments, we demonstrate that our approach results in substantially better calibration on various architectures. We demonstrate the robustness of our proposed approach on a real-world application for skin cancer classification and show that it facilitates safe decision-making under real-world uncertainties.

Reliably estimating the uncertainty of a prediction throughout the model lifecycle is crucial in many safety-critical applications. The most common way to measure this uncertainty is via the predicted confidence. While this tends to work well for in-domain samples, these estimates are unreliable under domain drift and restricted to classification. Alternatively, proper scores can be used for most predictive tasks but a bias-variance decomposition for model uncertainty does not exist in the current literature. In this work we introduce a general bias-variance decomposition for proper scores, giving rise to the Bregman Information as the variance term. We discover how exponential families and the classification log-likelihood are special cases and provide novel formulations. Surprisingly, we can express the classification case purely in the logit space. We showcase the practical relevance of this decomposition on several downstream tasks, including model ensembles and confidence regions. Further, we demonstrate how different approximations of the instance-level Bregman Information allow reliable out-of-distribution detection for all degrees of domain drift.

Many real-world systems are described not only by data from a single source but via multiple data views. In genomic medicine, for instance, patients can be characterized by data from different molecular layers. Latent variable models with structured sparsity are a commonly used tool for disentangling variation within and across data views. However, their interpretability is cumbersome since it requires a direct inspection and interpretation of each factor from domain experts. Here, we propose MuVI, a novel multi-view latent variable model based on a modified horseshoe prior for modeling structured sparsity. This facilitates the incorporation of limited and noisy domain knowledge, thereby allowing for an analysis of multi-view data in an inherently explainable manner. We demonstrate that our model (i) outperforms state-of-the-art approaches for modeling structured sparsity in terms of the reconstruction error and the precision/recall, (ii) robustly integrates noisy domain expertise in the form of feature sets, (iii) promotes the identifiability of factors and (iv) infers interpretable and biologically meaningful axes of variation in a real-world multi-view dataset of cancer patients.

Built on the success of the past five years, KDD DSHealth 2023 will further catalyze the development of links between academic and industrial data science groups. The workshop aims to stimulate discussion on strategic areas for development and to facilitate future cross-disciplinary collaborations. In accordance with the multi-year goal to continue fostering this community via timely topics, this year the workshop will focus on the applications and new development of generative models in healthcare, including the new development and application of LLMs. The workshop invites full papers, as well as work-in-progress on the application of data science in healthcare. The workshop will feature two invited talks from eminent speakers, spanning academia, industry, clinical researchers, and governmental regulatory bodies. In addition, we will invite community members to submit their research works and bring them for discussion. The summary gives a brief description of the half-day workshop to be held on August 7th, 2023.

KDD DSHealth 2022, aims to build on the success of the past four years to further catalyze the development of links between academic and commercial data science groups and the rapidly developing translational medicine informatics community. The workshop will stimulate discussion as to strategic areas for development and will lead to future cross-disciplinary collaborations. In accordance with the multi-year goal to continue fostering this community as a series of KDD workshops via timely topics, this year the workshop will focus on the transparency and human-centered AI in healthcare. The workshop invites full papers, as well as work-in-progress on the application of data science in healthcare. The workshop will feature four invited talks from eminent speakers, spanning academia, industry, clinical researchers, and governmental regulatory bodies. In addition, selected papers will be invited to publish in a special issue of Journal of Healthcare Informatics Research. The summary gives a brief description of the full-day workshop to be held on August 14th, 2022.

With model trustworthiness being crucial for sensitive real-world applications, practitioners are putting more and more focus on improving the uncertainty calibration of deep neural networks.Calibration errors are designed to quantify the reliability of probabilistic predictions but their estimators are usually biased and inconsistent.In this work, we introduce the framework of \textit{proper calibration errors}, which relates every calibration error to a proper score and provides a respective upper bound with optimal estimation properties.This relationship can be used to reliably quantify the model calibration improvement.We theoretically and empirically demonstrate the shortcomings of commonly used estimators compared to our approach.Due to the wide applicability of proper scores, this gives a natural extension of recalibration beyond classification.

To facilitate a wide-spread acceptance of AI systems guiding decision making in real-world applications, trustworthiness of deployed models is key. That is, it is crucial for predictive models to be uncertainty-aware and yield well-calibrated (and thus trustworthy) predictions for both in-domain samples as well as under domain shift. Recent efforts to account for predictive uncertainty include post-processing steps for trained neural networks, Bayesian neural networks as well as alternative non-Bayesian approaches such as ensemble approaches and evidential deep learning. Here, we propose an efficient yet general modelling approach for obtaining well-calibrated, trustworthy probabilities for samples obtained after a domain shift. We introduce a new training strategy combining an entropy-encouraging loss term with an adversarial calibration loss term and demonstrate that this results in well-calibrated and technically trustworthy predictions for a wide range of domain drifts. We comprehensively evaluate previously proposed approaches on different data modalities, a large range of data sets including sequence data, network architectures and perturbation strategies. We observe that our modelling approach substantially outperforms existing state-of-the-art approaches, yielding well-calibrated predictions under domain drift.

We address the problem of uncertainty calibration. While standard deep neural networks typically yield uncalibrated predictions, calibrated confidence scores that are representative of the true likelihood of a prediction can be achieved using post-hoc calibration methods. However, to date, the focus of these approaches has been on in-domain calibration. Our contribution is two-fold. First, we show that existing post-hoc calibration methods yield highly over-confident predictions under domain shift. Second, we introduce a simple strategy where perturbations are applied to samples in the validation set before performing the post-hoc calibration step. In extensive experiments, we demonstrate that this perturbation step results in substantially better calibration under domain shift on a wide range of architectures and modelling tasks.

KDD Health Day/DSHealth 2021, aims to build on the success of the past 3 years to further catalyze the development of links between academic and commercial data science groups and the rapidly developing translational medicine informatics community. The workshop will stimulate discussion as to strategic areas for development and will lead to future cross-disciplinary collaborations. In accordance with the multi-year goal to continue fostering this community as a series of KDD workshops via timely topics, this year the workshop will focus on the state of explainability and trustworthiness in healthcare. The workshop invites full papers, as well as work-in-progress on the application of data science in healthcare. The workshop will feature 8 invited talks from eminent speakers across academia, industry, clinical researchers, and governmental regulatory bodies. In addition, selected papers will be invited to publish in a special issue of Artificial Intelligence in Medicine journal. The summary gives a brief description of the full-day workshop to be held on August, 2021 virtually.

Recommender systems are often designed based on a collaborative filtering approach, where user preferences are predicted by modelling interactions between users and items. Many common approaches to solve the collaborative filtering task are based on learning representations of users and items, including simple matrix factorization, Gaussian process latent variable models, and neural-network based embeddings. While matrix factorization approaches fail to model nonlinear relations, neural networks can potentially capture such complex relations with unprecedented predictive power and are highly scalable. However, neither of them is able to model predictive uncertainties. In contrast, Gaussian Process based models can generate a predictive distribution, but cannot scale to large amounts of data. In this manuscript, we propose a novel approach combining the representation learning paradigm of collaborative filtering with multi-output Gaussian processes in a joint framework to generate uncertainty-aware recommendations. We introduce an efficient strategy for model training and inference, resulting in a model that scales to very large and sparse datasets and achieves competitive performance in terms of classical metrics quantifying the reconstruction error. In addition to accurately predicting user preferences, our model also provides meaningful uncertainty estimates about that prediction.

Diagnosing diseases such as leukemia or anemia requires reliable counts of blood cells. Hematologists usually label and count microscopy images of blood cells manually. In many cases, however, cells in different maturity states are difficult to distinguish, and in combination with image noise and subjectivity, humans are prone to make labeling mistakes. This results in labels that are often not reproducible, which can directly affect the diagnoses. We introduce TIMELY, a probabilistic model that combines pseudotime inference methods with inhomogeneous hidden Markov trees, which addresses this challenge of label inconsistency. We show first on simulation data that TIMELY is able to identify and correct wrong labels with higher precision and recall than baseline methods for labeling correction. We then apply our method to two real-world datasets of blood cell data and show that TIMELY successfully finds inconsistent labels, thereby improving the quality of human-generated labels.

We address two challenges in topic models: (1) Context information around words helps in determining their actual meaning, e.g., “networks” used in the contexts artificial neural networks vs. biological neuron networks. Generative topic models infer topic-word distributions, taking no or only little context into account. Here, we extend a neural autoregressive topic model to exploit the full context information around words in a document in a language modeling fashion. The proposed model is named as iDocNADE. (2) Due to the small number of word occurrences (i.e., lack of context) in short text and data sparsity in a corpus of few documents, the application of topic models is challenging on such texts. Therefore, we propose a simple and efficient way of incorporating external knowledge into neural autoregressive topic models: we use embeddings as a distributional prior. The proposed variants are named as DocNADEe and iDocNADEe. We present novel neural autoregressive topic model variants that consistently outperform state-of-the-art generative topic models in terms of generalization, interpretability (topic coherence) and applicability (retrieval and classification) over 7 long-text and 8 short-text datasets from diverse domains.

A reduction in salivary flow and xerostomia are common side-effects after radiotherapy of head and neck tumours. Xerostomia can be modeled based on the dose to the parotid glands. To date, all spatial information has been discarded and dose-response models are usually reduced to the mean dose. We present novel morphological dose-response models and use multivariate Bayesian logistic regression to model xerostomia. We use 3D invariant statistical moments as morphometric descriptors to quantify the shape of the 3D dose distribution. As this results in a very high number of potential predictors, we apply a Bayesian variable-selection algorithm to find the best model based on any subset of all potential predictors. To do this, we determine the posterior probabilities of being the best model for all potential models and calculate the marginal probabilities that a variable should be included in a model. This was done using a Reversible Jump Markov Chain Monte Carlo algorithm. The performance of the best model was quantified using the deviance information criterion and a leave-one-out cross-validation (LOOCV). This methodology was applied to 64 head and neck cancer patients treated with either intensity-modulated radiotherapy (IMRT) or conventional radiotherapy. Results show a substantial increase in both model-fit and area under the curve (AUC) when including morphological information compared to conventional mean-dose models. The best mean-dose model for IMRT patients only resulted in an AUC of 0.63 after LOOCV while the best morphological model had an AUC of 0.90. For conventional patients the mean-dose model and the morphological model had AUC of 0.55 and 0.86 respectively. For a joint model with all patients pooled together, the mean dose model had an AUC of 0.75 and the morphological model an AUC of 0.88. We have shown that invariant statistical moments are a good morphometric descriptor and by using Bayesian variable selection we were able to identify models with a substantially higher predictive power than conventional mean-dose models.

Radiotherapy treatments of cancer patients are planned using dose-volume constraints. These constraints limit the volume of organs receiving a given threshold dose. We propose a new framework to predict radiation-induced toxicities and evaluate dosimetric constraints using Bayesian logistic regression with high-order interactions. The predictive power of 2 sets of rectal dose-volume constraints proposed in the recent literature was evaluated using follow-up data from the RT01 prostate radiotherapy trial. Toxicities considered were rectal bleeding and loose stools. Furthermore we derived a new type of geometrical dosimetric constraint and assessed the predictive power. % using the Bayesian logistic regression model. Bayesian logistic regression with high-order interactions using dosimetric constraints successfully predicted radiation-induced rectal bleeding and loose stools. Literature-based dose-volume constraints had less predictive power than our new type of geometrical constraint. Imposing the latter type of constraints when generating a treatment plan would be beneficial for outcome.