Showing papers by "University of Freiburg published in 2015"

U-Net: Convolutional Networks for Biomedical Image Segmentation

Abstract: There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net .

U-Net: Convolutional Networks for Biomedical Image Segmentation

Abstract: There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at this http URL .

FlowNet: Learning Optical Flow with Convolutional Networks

Abstract: Convolutional neural networks (CNNs) have recently been very successful in a variety of computer vision tasks, especially on those linked to recognition. Optical flow estimation has not been among the tasks CNNs succeeded at. In this paper we construct CNNs which are capable of solving the optical flow estimation problem as a supervised learning task. We propose and compare two architectures: a generic architecture and another one including a layer that correlates feature vectors at different image locations. Since existing ground truth data sets are not sufficiently large to train a CNN, we generate a large synthetic Flying Chairs dataset. We show that networks trained on this unrealistic data still generalize very well to existing datasets such as Sintel and KITTI, achieving competitive accuracy at frame rates of 5 to 10 fps.

Biological properties of extracellular vesicles and their physiological functions

Abstract: In the past decade, extracellular vesicles (EVs) have been recognized as potent vehicles of intercellular communication, both in prokaryotes and eukaryotes. This is due to their capacity to transfer proteins, lipids and nucleic acids, thereby influencing various physiological and pathological functions of both recipient and parent cells. While intensive investigation has targeted the role of EVs in different pathological processes, for example, in cancer and autoimmune diseases, the EV-mediated maintenance of homeostasis and the regulation of physiological functions have remained less explored. Here, we provide a comprehensive overview of the current understanding of the physiological roles of EVs, which has been written by crowd-sourcing, drawing on the unique EV expertise of academia-based scientists, clinicians and industry based in 27 European countries, the United States and Australia. This review is intended to be of relevance to both researchers already working on EV biology and to newcomers who will encounter this universal cell biological system. Therefore, here we address the molecular contents and functions of EVs in various tissues and body fluids from cell systems to organs. We also review the physiological mechanisms of EVs in bacteria, lower eukaryotes and plants to highlight the functional uniformity of this emerging communication system.

Striving for Simplicity: The All Convolutional Net

Abstract: Most modern convolutional neural networks (CNNs) used for object recognition are built using the same principles: Alternating convolution and max-pooling layers followed by a small number of fully connected layers. We re-evaluate the state of the art for object recognition from small images with convolutional networks, questioning the necessity of different components in the pipeline. We find that max-pooling can simply be replaced by a convolutional layer with increased stride without loss in accuracy on several image recognition benchmarks. Following this finding -- and building on other recent work for finding simple network structures -- we propose a new architecture that consists solely of convolutional layers and yields competitive or state of the art performance on several object recognition datasets (CIFAR-10, CIFAR-100, ImageNet). To analyze the network we introduce a new variant of the "deconvolution approach" for visualizing features learned by CNNs, which can be applied to a broader range of network structures than existing approaches.

Host microbiota constantly control maturation and function of microglia in the CNS

Abstract: As the tissue macrophages of the CNS, microglia are critically involved in diseases of the CNS. However, it remains unknown what controls their maturation and activation under homeostatic conditions. We observed substantial contributions of the host microbiota to microglia homeostasis, as germ-free (GF) mice displayed global defects in microglia with altered cell proportions and an immature phenotype, leading to impaired innate immune responses. Temporal eradication of host microbiota severely changed microglia properties. Limited microbiota complexity also resulted in defective microglia. In contrast, recolonization with a complex microbiota partially restored microglia features. We determined that short-chain fatty acids (SCFA), microbiota-derived bacterial fermentation products, regulated microglia homeostasis. Accordingly, mice deficient for the SCFA receptor FFAR2 mirrored microglia defects found under GF conditions. These findings suggest that host bacteria vitally regulate microglia maturation and function, whereas microglia impairment can be rectified to some extent by complex microbiota.

An updated evolutionary classification of CRISPR–Cas systems

Abstract: The evolution of CRISPR-cas loci, which encode adaptive immune systems in archaea and bacteria, involves rapid changes, in particular numerous rearrangements of the locus architecture and horizontal transfer of complete loci or individual modules. These dynamics complicate straightforward phylogenetic classification, but here we present an approach combining the analysis of signature protein families and features of the architecture of cas loci that unambiguously partitions most CRISPR-cas loci into distinct classes, types and subtypes. The new classification retains the overall structure of the previous version but is expanded to now encompass two classes, five types and 16 subtypes. The relative stability of the classification suggests that the most prevalent variants of CRISPR-Cas systems are already known. However, the existence of rare, currently unclassifiable variants implies that additional types and subtypes remain to be characterized.

A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation

Abstract: Recent work has shown that optical flow estimation can be formulated as a supervised learning task and can be successfully solved with convolutional networks. Training of the so-called FlowNet was enabled by a large synthetically generated dataset. The present paper extends the concept of optical flow estimation via convolutional networks to disparity and scene flow estimation. To this end, we propose three synthetic stereo video datasets with sufficient realism, variation, and size to successfully train large networks. Our datasets are the first large-scale datasets to enable training and evaluating scene flow methods. Besides the datasets, we present a convolutional network for real-time disparity estimation that provides state-of-the-art results. By combining a flow and disparity estimation network and training it jointly, we demonstrate the first scene flow estimation with a convolutional network.

Combined Measurement of the Higgs Boson Mass in pp Collisions at √s=7 and 8 TeV with the ATLAS and CMS Experiments

Abstract: A measurement of the Higgs boson mass is presented based on the combined data samples of the ATLAS and CMS experiments at the CERN LHC in the H→γγ and H→ZZ→4l decay channels. The results are obtained from a simultaneous fit to the reconstructed invariant mass peaks in the two channels and for the two experiments. The measured masses from the individual channels and the two experiments are found to be consistent among themselves. The combined measured mass of the Higgs boson is mH=125.09±0.21 (stat)±0.11 (syst) GeV.

Efficient and robust automated machine learning

Abstract: The success of machine learning in a broad range of applications has led to an ever-growing demand for machine learning systems that can be used off the shelf by non-experts. To be effective in practice, such systems need to automatically choose a good algorithm and feature preprocessing steps for a new dataset at hand, and also set their respective hyperparameters. Recent work has started to tackle this automated machine learning (AutoML) problem with the help of efficient Bayesian optimization methods. Building on this, we introduce a robust new AutoML system based on scikit-learn (using 15 classifiers, 14 feature preprocessing methods, and 4 data preprocessing methods, giving rise to a structured hypothesis space with 110 hyperparameters). This system, which we dub AUTO-SKLEARN, improves on existing AutoML methods by automatically taking into account past performance on similar datasets, and by constructing ensembles from the models evaluated during the optimization. Our system won the first phase of the ongoing ChaLearn AutoML challenge, and our comprehensive analysis on over 100 diverse datasets shows that it substantially outperforms the previous state of the art in AutoML. We also demonstrate the performance gains due to each of our contributions and derive insights into the effectiveness of the individual components of AUTO-SKLEARN.

Modulation of genetic associations with serum urate levels by body-mass-index in humans

Abstract: We tested for interactions between body mass index (BMI) and common genetic variants affecting serum urate levels, genome-wide, in up to 42569 participants. Both stratified genome-wide association (GWAS) analyses, in lean, overweight and obese individuals, and regression-type analyses in a non BMI-stratified overall sample were performed. The former did not uncover any novel locus with a major main effect, but supported modulation of effects for some known and potentially new urate loci. The latter highlighted a SNP at RBFOX3 reaching genome-wide significant level (effect size 0.014, 95% CI 0.008-0.02, Pinter= 2.6 x 10-8). Two top loci in interaction term analyses, RBFOX3 and ERO1LB-EDARADD, also displayed suggestive differences in main effect size between the lean and obese strata. All top ranking loci for urate effect differences between BMI categories were novel and most had small magnitude but opposite direction effects between strata. They include the locus RBMS1-TANK (men, Pdifflean-overweight= 4.7 x 10-8), a region that has been associated with several obesity related traits, and TSPYL5 (men, Pdifflean-overweight= 9.1 x 10-8), regulating adipocytes-produced estradiol. The top-ranking known urate loci was ABCG2, the strongest known gout risk locus, with an effect halved in obese compared to lean men (Pdifflean-obese= 2 x 10-4). Finally, pathway analysis suggested a role for N-glycan biosynthesis as a prominent urate-associated pathway in the lean stratum. These results illustrate a potentially powerful way to monitor changes occurring in obesogenic environment.

Applying extracellular vesicles based therapeutics in clinical trials - an ISEV position paper

Abstract: Extracellular vesicles (EVs), such as exosomes and microvesicles, are released by different cell types and participate in physiological and pathophysiological processes. EVs mediate intercellular communication as cell-derived extracellular signalling organelles that transmit specific information from their cell of origin to their target cells. As a result of these properties, EVs of defined cell types may serve as novel tools for various therapeutic approaches, including (a) anti-tumour therapy, (b) pathogen vaccination, (c) immune-modulatory and regenerative therapies and (d) drug delivery. The translation of EVs into clinical therapies requires the categorization of EV-based therapeutics in compliance with existing regulatory frameworks. As the classification defines subsequent requirements for manufacturing, quality control and clinical investigation, it is of major importance to define whether EVs are considered the active drug components or primarily serve as drug delivery vehicles. For an effective and particularly safe translation of EV-based therapies into clinical practice, a high level of cooperation between researchers, clinicians and competent authorities is essential. In this position statement, basic and clinical scientists, as members of the International Society for Extracellular Vesicles (ISEV) and of the European Cooperation in Science and Technology (COST) program of the European Union, namely European Network on Microvesicles and Exosomes in Health and Disease (ME-HaD), summarize recent developments and the current knowledge of EV-based therapies. Aspects of safety and regulatory requirements that must be considered for pharmaceutical manufacturing and clinical application are highlighted. Production and quality control processes are discussed. Strategies to promote the therapeutic application of EVs in future clinical studies are addressed.

Ranking treatments in frequentist network meta-analysis works without resampling methods

Abstract: Network meta-analysis is used to compare three or more treatments for the same condition. Within a Bayesian framework, for each treatment the probability of being best, or, more general, the probability that it has a certain rank can be derived from the posterior distributions of all treatments. The treatments can then be ranked by the surface under the cumulative ranking curve (SUCRA). For comparing treatments in a network meta-analysis, we propose a frequentist analogue to SUCRA which we call P-score that works without resampling. P-scores are based solely on the point estimates and standard errors of the frequentist network meta-analysis estimates under normality assumption and can easily be calculated as means of one-sided p-values. They measure the mean extent of certainty that a treatment is better than the competing treatments. Using case studies of network meta-analysis in diabetes and depression, we demonstrate that the numerical values of SUCRA and P-Score are nearly identical. Ranking treatments in frequentist network meta-analysis works without resampling. Like the SUCRA values, P-scores induce a ranking of all treatments that mostly follows that of the point estimates, but takes precision into account. However, neither SUCRA nor P-score offer a major advantage compared to looking at credible or confidence intervals.

Essential versus accessory aspects of cell death: recommendations of the NCCD 2015

Abstract: Cells exposed to extreme physicochemical or mechanical stimuli die in an uncontrollable manner, as a result of their immediate structural breakdown. Such an unavoidable variant of cellular demise is generally referred to as ‘accidental cell death’ (ACD). In most settings, however, cell death is initiated by a genetically encoded apparatus, correlating with the fact that its course can be altered by pharmacologic or genetic interventions. ‘Regulated cell death’ (RCD) can occur as part of physiologic programs or can be activated once adaptive responses to perturbations of the extracellular or intracellular microenvironment fail. The biochemical phenomena that accompany RCD may be harnessed to classify it into a few subtypes, which often (but not always) exhibit stereotyped morphologic features. Nonetheless, efficiently inhibiting the processes that are commonly thought to cause RCD, such as the activation of executioner caspases in the course of apoptosis, does not exert true cytoprotective effects in the mammalian system, but simply alters the kinetics of cellular demise as it shifts its morphologic and biochemical correlates. Conversely, bona fide cytoprotection can be achieved by inhibiting the transduction of lethal signals in the early phases of the process, when adaptive responses are still operational. Thus, the mechanisms that truly execute RCD may be less understood, less inhibitable and perhaps more homogeneous than previously thought. Here, the Nomenclature Committee on Cell Death formulates a set of recommendations to help scientists and researchers to discriminate between essential and accessory aspects of cell death.

Learning to generate chairs with convolutional neural networks

Abstract: We train a generative convolutional neural network which is able to generate images of objects given object type, viewpoint, and color. We train the network in a supervised manner on a dataset of rendered 3D chair models. Our experiments show that the network does not merely learn all images by heart, but rather finds a meaningful representation of a 3D chair model allowing it to assess the similarity of different chairs, interpolate between given viewpoints to generate the missing ones, or invent new chair styles by interpolating between chairs from the training set. We show that the network can be used to find correspondences between different chairs from the dataset, outperforming existing approaches on this task.

Delivery of crop pollination services is an insufficient argument for wild pollinator conservation

Abstract: There is compelling evidence that more diverse ecosystems deliver greater benefits to people, and these ecosystem services have become a key argument for biodiversity conservation. However, it is unclear how much biodiversity is needed to deliver ecosystem services in a cost-effective way. Here we show that, while the contribution of wild bees to crop production is significant, service delivery is restricted to a limited subset of all known bee species. Across crops, years and biogeographical regions, crop-visiting wild bee communities are dominated by a small number of common species, and threatened species are rarely observed on crops. Dominant crop pollinators persist under agricultural expansion and many are easily enhanced by simple conservation measures, suggesting that cost-effective management strategies to promote crop pollination should target a different set of species than management strategies to promote threatened bees. Conserving the biological diversity of bees therefore requires more than just ecosystem-service-based arguments.

Minimum Information about a Biosynthetic Gene cluster.

Abstract: A wide variety of enzymatic pathways that produce specialized metabolites in bacteria, fungi and plants are known to be encoded in biosynthetic gene clusters. Information about these clusters, pathways and metabolites is currently dispersed throughout the literature, making it difficult to exploit. To facilitate consistent and systematic deposition and retrieval of data on biosynthetic gene clusters, we propose the Minimum Information about a Biosynthetic Gene cluster (MIBiG) data standard.

Multimodal deep learning for robust RGB-D object recognition

Abstract: Robust object recognition is a crucial ingredient of many, if not all, real-world robotics applications. This paper leverages recent progress on Convolutional Neural Networks (CNNs) and proposes a novel RGB-D architecture for object recognition. Our architecture is composed of two separate CNN processing streams - one for each modality - which are consecutively combined with a late fusion network. We focus on learning with imperfect sensor data, a typical problem in real-world robotics tasks. For accurate learning, we introduce a multi-stage training methodology and two crucial ingredients for handling depth data with CNNs. The first, an effective encoding of depth information for CNNs that enables learning without the need for large depth datasets. The second, a data augmentation scheme for robust learning with depth images by corrupting them with realistic noise patterns. We present state-of-the-art results on the RGB-D object dataset [15] and show recognition in challenging RGB-D real-world noisy settings.

Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images

Abstract: We introduce Embed to Control (E2C), a method for model learning and control of non-linear dynamical systems from raw pixel images. E2C consists of a deep generative model, belonging to the family of variational autoencoders, that learns to generate image trajectories from a latent space in which the dynamics is constrained to be locally linear. Our model is derived directly from an optimal control formulation in latent space, supports long-term prediction of image sequences and exhibits strong performance on a variety of complex control problems.

Oxygen Defects in Phosphorene

Abstract: Surface reactions with oxygen are a fundamental cause of the degradation of phosphorene. Using first-principles calculations, we show that for each oxygen atom adsorbed onto phosphorene there is an energy release of about 2 eV. Although the most stable oxygen adsorbed forms are electrically inactive and lead only to minor distortions of the lattice, there are low energy metastable forms which introduce deep donor and/or acceptor levels in the gap. We also propose a mechanism for phosphorene oxidation involving reactive dangling oxygen atoms and we suggest that dangling oxygen atoms increase the hydrophilicity of phosphorene.

Oxaliplatin added to fluorouracil-based preoperative chemoradiotherapy and postoperative chemotherapy of locally advanced rectal cancer (the German CAO/ARO/AIO-04 study): final results of the multicentre, open-label, randomised, phase 3 trial

Abstract: Summary Background Preoperative chemoradiotherapy with infusional fluorouracil, total mesorectal excision surgery, and postoperative chemotherapy with fluorouracil was established by the German CAO/ARO/AIO-94 trial as a standard combined modality treatment for locally advanced rectal cancer. Here we compare the previously established regimen with an investigational regimen in which oxaliplatin was added to both preoperative chemoradiotherapy and postoperative chemotherapy. Methods In this multicentre, open-label, randomised, phase 3 study we randomly assigned patients with rectal adenocarcinoma, clinically staged as cT3–4 or any node-positive disease, to two groups: a control group receiving standard fluorouracil-based combined modality treatment, consisting of preoperative radiotherapy of 50·4 Gy in 28 fractions plus infusional fluorouracil (1000 mg/m 2 on days 1–5 and 29–33), followed by surgery and four cycles of bolus fluorouracil (500 mg/m 2 on days 1–5 and 29); or to an investigational group receiving preoperative radiotherapy of 50·4 Gy in 28 fractions plus infusional fluorouracil (250 mg/m 2 on days 1–14 and 22–35) and oxaliplatin (50 mg/m 2 on days 1, 8, 22, and 29), followed by surgery and eight cycles of oxaliplatin (100 mg/m 2 on days 1 and 15), leucovorin (400 mg/m 2 on days 1 and 15), and infusional fluorouracil (2400 mg/m 2 on days 1–2 and 15–16). Randomisation was done with computer-generated block-randomisation codes stratified by centre, clinical T category (cT1–3 vs cT4), and clinical N category (cN0 vs cN1–2) without masking. The primary endpoint was disease-free survival, defined as the time between randomisation and non-radical surgery of the primary tumour (R2 resection), locoregional recurrence after R0/1 resection, metastatic disease or progression, or death from any cause, whichever occurred first. Survival and cumulative incidence of recurrence analyses followed the intention-to-treat principle; toxicity analyses included all patients treated. Enrolment of patients in this trial is completed and follow-up is ongoing. This study is registered with ClinicalTrials.gov, number NCT00349076. Findings Of the 1265 patients initially enrolled, 1236 were assessable (613 in the investigational group and 623 in the control group). With a median follow-up of 50 months (IQR 38–61), disease-free survival at 3 years was 75·9% (95% CI 72·4–79·5) in the investigational group and 71·2% (95% CI 67·6–74·9) in the control group (hazard ratio [HR] 0·79, 95% CI 0·64–0·98; p=0·03). Preoperative grade 3–4 toxic effects occurred in 144 (24%) of 607 patients who actually received fluorouracil and oxaliplatin during chemoradiotherapy and in 128 (20%) of 625 patients who actually received fluorouracil chemoradiotherapy. Of 445 patients who actually received adjuvant fluorouracil and leucovorin and oxaliplatin, 158 (36%) had grade 3–4 toxic effects, as did 170 (36%) of 470 patients who actually received adjuvant fluorouracil. Late grade 3–4 adverse events in patients who received protocol-specified preoperative and postoperative treatment occurred in 112 (25%) of 445 patients in the investigational group, and in 100 (21%) of 470 patients in the control group. Interpretation Adding oxaliplatin to fluorouracil-based neoadjuvant chemoradiotherapy and adjuvant chemotherapy (at the doses and intensities used in this trial) significantly improved disease-free survival of patients with clinically staged cT3–4 or cN1–2 rectal cancer compared with our former fluorouracil-based combined modality regimen (based on CAO/ARO/AIO-94). The regimen established by CAO/ARO/AIO-04 can be deemed a new treatment option for patients with locally advanced rectal cancer. Funding German Cancer Aid (Deutsche Krebshilfe).

Speeding up automatic hyperparameter optimization of deep neural networks by extrapolation of learning curves

Abstract: Deep neural networks (DNNs) show very strong performance on many machine learning problems, but they are very sensitive to the setting of their hyperparameters. Automated hyperparameter optimization methods have recently been shown to yield settings competitive with those found by human experts, but their widespread adoption is hampered by the fact that they require more computational resources than human experts. Humans have one advantage: when they evaluate a poor hyperparameter setting they can quickly detect (after a few steps of stochastic gradient descent) that the resulting network performs poorly and terminate the corresponding evaluation to save time. In this paper, we mimic the early termination of bad runs using a probabilistic model that extrapolates the performance from the first part of a learning curve. Experiments with a broad range of neural network architectures on various prominent object recognition benchmarks show that our resulting approach speeds up state-of-the-art hyperparameter optimization methods for DNNs roughly twofold, enabling them to find DNN settings that yield better performance than those chosen by human experts.

Land use intensification alters ecosystem multifunctionality via loss of biodiversity and changes to functional composition

Abstract: Global change, especially land-use intensification, affects human well-being by impacting the delivery of multiple ecosystem services (multifunctionality). However, whether biodiversity loss is a major component of global change effects on multifunctionality in real-world ecosystems, as in experimental ones, remains unclear. Therefore, we assessed biodiversity, functional composition and 14 ecosystem services on 150 agricultural grasslands differing in land-use intensity. We also introduce five multifunctionality measures in which ecosystem services were weighted according to realistic land-use objectives. We found that indirect land-use effects, i.e. those mediated by biodiversity loss and by changes to functional composition, were as strong as direct effects on average. Their strength varied with land-use objectives and regional context. Biodiversity loss explained indirect effects in a region of intermediate productivity and was most damaging when land-use objectives favoured supporting and cultural services. In contrast, functional composition shifts, towards fast-growing plant species, strongly increased provisioning services in more inherently unproductive grasslands.

Versatile sample environments and automation for biological solution X‐ray scattering experiments at the P12 beamline (PETRA III, DESY)

Abstract: A high-brilliance synchrotron P12 beamline of the EMBL located at the PETRA III storage ring (DESY, Hamburg) is dedicated to biological small-angle X-ray scattering (SAXS) and has been designed and optimized for scattering experiments on macromolecular solutions. Scatterless slits reduce the parasitic scattering, a custom-designed miniature active beamstop ensures accurate data normalization and the photon-counting PILATUS 2M detector enables the background-free detection of weak scattering signals. The high flux and small beam size allow for rapid experiments with exposure time down to 30–50 ms covering the resolution range from about 300 to 0.5 nm. P12 possesses a versatile and flexible sample environment system that caters for the diverse experimental needs required to study macromolecular solutions. These include an in-vacuum capillary mode for standard batch sample analyses with robotic sample delivery and for continuous-flow in-line sample purification and characterization, as well as an in-air capillary time-resolved stopped-flow setup. A novel microfluidic centrifugal mixing device (SAXS disc) is developed for a high-throughput screening mode using sub-microlitre sample volumes. Automation is a key feature of P12; it is controlled by a beamline meta server, which coordinates and schedules experiments from either standard or nonstandard operational setups. The integrated SASFLOW pipeline automatically checks for consistency, and processes and analyses the data, providing near real-time assessments of overall parameters and the generation of low-resolution models within minutes of data collection. These advances, combined with a remote access option, allow for rapid high-throughput analysis, as well as time-resolved and screening experiments for novice and expert biological SAXS users.

Embed to control: a locally Linear Latent dynamics model for control from raw images

Abstract: We introduce Embed to Control (E2C), a method for model learning and control of non-linear dynamical systems from raw pixel images. E2C consists of a deep generative model, belonging to the family of variational autoencoders, that learns to generate image trajectories from a latent space in which the dynamics is constrained to be locally linear. Our model is derived directly from an optimal control formulation in latent space, supports long-term prediction of image sequences and exhibits strong performance on a variety of complex control problems.

Natural history of chronic hepatitis B virus infection

Abstract: Hepatitis B virus infection represents a major global health problem. Currently, there are more than 240 million chronically infected people worldwide. The development of chronic hepatitis B virus-mediated liver disease may lead to liver fibrosis, cirrhosis and eventually hepatocellular carcinoma. Recently, the discovery of the viral entry receptor sodium taurocholate cotransporting polypeptide has facilitated new approaches for a better understanding of viral physiopathology. Hopefully, these novel insights may give rise to the development of more effective antiviral therapy concepts during the next years. In this review, we will discuss the natural history of hepatitis B virus infection including the viral biology, the clinical course of infection and the role of the immune response.