No abstract

The nal publication is available at Springer via http://dx.doi.org/10.1007/978-3-319-46604-011Additional information: Use policyThe full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that:• a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders.Please consult the full DRO policy for further details. Abstract. Real-time road-scene understanding is a challenging computer vision task with recent advances in convolutional neural networks (CNN) achieving results that notably surpass prior traditional feature driven approaches. Here, we take an existing CNN architecture, pre-trained for urban road-scene understanding, and retrain it towards the task of classifying off-road scenes, assessing the network performance during the training cycle. Within the paradigm of transfer learning we analyse the effects on CNN classification, by training and assessing varying levels of prior training on varying sub-sets of our off-road training data. For each of these configurations, we evaluate the network at multiple points during its training cycle, allowing us to analyse in depth exactly how the training process is affected by these variations. Finally, we compare this CNN to a more traditional approach using a feature-driven Support Vector Machine (SVM) classifier and demonstrate state-of-the-art results in this particularly challenging problem of off-road scene understanding.

The number, size and severity of aquatic low-oxygen dead zones are increasing worldwide. Microbial processes in low-oxygen environments have important ecosystem-level consequences, such as denitrification, greenhouse gas production and acidification. To identify key microbial processes occurring in lowoxygen bottom waters of the Chesapeake Bay, we sequenced both 16S rRNA genes and shotgun metagenomic libraries to determine the identity, functional potential and spatiotemporal distribution of microbial populations in the water column. Unsupervised clustering algorithms grouped samples into three clusters using water chemistry or microbial communities, with extensive overlap of cluster composition between methods. Clusters were strongly differentiated by temperature, salinity and oxygen. Sulfuroxidizing microorganisms were found to be enriched in the low-oxygen bottom water and predictive of hypoxic conditions. Metagenome-assembled genomes demonstrate that some of these sulfuroxidizing populations are capable of partial denitrification and transcriptionally active in a prior study. These results suggest that microorganisms capable of oxidizing reduced sulfur compounds are a previously unidentified microbial indicator of low oxygen in the Chesapeake Bay and reveal ties between the sulfur, nitrogen and oxygen cycles that could be important to capture when predicting the ecosystem response to remediation efforts or climate change.