Search

In many medical applications, interpretable models with high prediction performance are sought. Often, those models are required to handle semi-structured data like tabular and image data. We show how to apply deep transformation models (DTMs) for distributional regression which fulfill these requirements. DTMs allow the data analyst to specify (deep) neural networks for different input modalities making them applicable to various research questions. Like statistical models, DTMs can provide interpretable effect estimates while achieving the state-of-the-art prediction performance of deep neural networks. In addition, the construction of ensembles of DTMs that retain model structure and interpretability allows quantifying epistemic and aleatoric uncertainty. In this study, we compare several DTMs, including baseline-adjusted models, trained on a semi-structured data set of 407 stroke patients with the aim to predict ordinal functional outcome three months after stroke. We follow statistical principles of model-building to achieve an adequate trade-off between interpretability and flexibility while assessing the relative importance of the involved data modalities. We evaluate the models for an ordinal and dichotomized version of the outcome as used in clinical practice. We show that both, tabular clinical and brain imaging data, are useful for functional outcome prediction, while models based on tabular data only outperform those based on imaging data only. There is no substantial evidence for improved prediction when combining both data modalities. Overall, we highlight that DTMs provide a powerful, interpretable approach to analyzing semi-structured data and that they have the potential to support clinical decision making., Preprint under review

We present a deep transformation model for probabilistic regression. Deep learning is known for outstandingly accurate predictions on complex data but in regression tasks, it is predominantly used to just predict a single number. This ignores the non-deterministic character of most tasks. Especially if crucial decisions are based on the predictions, like in medical applications, it is essential to quantify the prediction uncertainty. The presented deep learning transformation model estimates the whole conditional probability distribution, which is the most thorough way to capture uncertainty about the outcome. We combine ideas from a statistical transformation model (most likely transformation) with recent transformation models from deep learning (normalizing flows) to predict complex outcome distributions. The core of the method is a parameterized transformation function which can be trained with the usual maximum likelihood framework using gradient descent. The method can be combined with existing deep learning architectures. For small machine learning benchmark datasets, we report state of the art performance for most dataset and partly even outperform it. Our method works for complex input data, which we demonstrate by employing a CNN architecture on image data.

In this paper, we describe how we created a meta-classifier to detect the message-level sentiment of tweets. We participated in SemEval-2014 Task 9B by combining the results of several existing classifiers using a random forest. The results of 5 other teams from the competition as well as from 7 general purpose commercial classifiers were used to train the algorithm. This way, we were able to get a boost of up to 3.24 F1 score points.

Kramers relaxation times $\tau_{K}$ and relaxation times $\tau_{R}$ and $\tau_{G}$ for the end-to-end distances and for center of mass diffusion are calculated for dense systems of athermal lattice chains. $\tau_{K}$ is defined from the response of the radius of gyration to a Kramers potential which approximately describes the effect of a stationary shear flow. It is shown that within an intermediate range of chain lengths N the relaxation times $\tau_{R}$ and $\tau_{K}$ exhibit the same scaling with N, suggesting that N-dependent melt-viscosities for non-entangled chains can be obtained from the Kramers equilibrium concept., Comment: submitted to: Journal of Chemical Physics

Structural disorder is an inherent property of solid materials, which can support a macroscopic ionic current. Many transport phenomena in these solid ionic conductors appear to be related to concepts from percolation theory. We demonstrate this for three classes of materials, namely (i) dispersed ionic conductors, which show conductance properties that can be related to random electrical networks, (ii) ion-doped network glasses, whose concentration-dependent diffusion properties are accessible by critical path analysis, and (iii) polymer ionic conductors. For the latter we discuss Monte Carlo simulations which indicate the applicability of dynamic bond percolation theory.

Fast ion conduction is known to occur in many solid materials, allowing us to perform fundamental studies of diffusion processes in a variety of structures. After briefly discussing classes of solid ionic conductors with increasing complexity in their structural and transport behaviour, we shall focus on polymer electrolytes. A lattice model is described which emphasizes the asymmetry in the cation-polymer and anion-polymer interactions. Monte-Carlo simulations of the equation of state are supplemented by calculations based on the quasi-chemical approximation. Simulated diffusion constants and their dependence on temperature and ion concentration are compared with recent experiments on PEO (polyethylene oxid)-type conductors.