Filters








626 Hits in 1.7 sec

Learning by Association - A versatile semi-supervised training method for neural networks [article]

Philip Häusser and Alexander Mordvintsev and Daniel Cremers
2017 arXiv   pre-print
In many real-world scenarios, labeled data for a specific machine learning task is costly to obtain. Semi-supervised training methods make use of abundantly available unlabeled data and a smaller number of labeled examples. We propose a new framework for semi-supervised training of deep neural networks inspired by learning in humans. "Associations" are made from embeddings of labeled samples to those of unlabeled ones and back. The optimization schedule encourages correct association cycles
more » ... end up at the same class from which the association was started and penalizes wrong associations ending at a different class. The implementation is easy to use and can be added to any existing end-to-end training setup. We demonstrate the capabilities of learning by association on several data sets and show that it can improve performance on classification tasks tremendously by making use of additionally available unlabeled data. In particular, for cases with few labeled data, our training scheme outperforms the current state of the art on SVHN.
arXiv:1706.00909v1 fatcat:ydjimu3firhlbk5j65u5tfkjdu

FlowNet: Learning Optical Flow with Convolutional Networks [article]

Philipp Fischer, Alexey Dosovitskiy, Eddy Ilg, Philip Häusser, Caner Hazırbaş, Vladimir Golkov, Patrick van der Smagt, Daniel Cremers, Thomas Brox
2015 arXiv   pre-print
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 where CNNs were successful. In this paper we construct appropriate 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
more » ... ctors at different image locations. Since existing ground truth data sets are not sufficiently large to train a CNN, we generate a 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.
arXiv:1504.06852v2 fatcat:po46vuso2renfdy7hnivq5g7zy

Evidence of an oncogenic gammaherpesvirus in domestic dogs

Shih-Hung Huang, Philip J. Kozak, Jessica Kim, Georges Habineza-Ndikuyeze, Charles Meade, Anita Gaurnier-Hausser, Reema Patel, Erle Robertson, Nicola J. Mason
2012 Virology  
In addition, the canonical NF-κB pathway is constitutively active in malignant lymphocytes of dogs with DLBCL (Gaurnier-Hausser et al., 2011) , a finding that is shared with EBVassociated lymphoid malignancies  ... 
doi:10.1016/j.virol.2012.02.013 pmid:22405628 pmcid:PMC4592777 fatcat:x5f6hn3qmvcy3dhxgbdk42krme

FlowNet: Learning Optical Flow with Convolutional Networks

Alexey Dosovitskiy, Philipp Fischer, Eddy Ilg, Philip Hausser, Caner Hazirbas, Vladimir Golkov, Patrick van der Smagt, Daniel Cremers, Thomas Brox
2015 2015 IEEE International Conference on Computer Vision (ICCV)  
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 where CNNs were successful. In this paper we construct appropriate 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
more » ... ctors at different image locations. Since existing ground truth data sets are not sufficiently large to train a CNN, we generate a 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.
doi:10.1109/iccv.2015.316 dblp:conf/iccv/DosovitskiyFIHH15 fatcat:mehno2azlbcmxlhvthtnpdqzni

Loss of Bardet-Biedl syndrome proteins causes synaptic aberrations in principal neurons [article]

Naila Haq, Christoph Schmidt-Hieber, Fernando J. Sialana, Lorenza Ciani, Janosch P. Heller, Michelle Stewart, Liz Bentley, Sara Wells, Richard J. Rodenburg, Patrick M Nolan, Elizabeth Forsythe, Michael C. Wu (+4 others)
2019 bioRxiv   pre-print
AbstractBardet-Biedl syndrome (BBS), a ciliopathy, is a rare genetic condition characterised by retina degeneration, obesity, kidney failure and cognitive impairment. In spite of a progress made in general understanding of BBS aetiology, the molecular mechanism of cognitive impairment remains elusive. Here we report that loss of Bardet-Biedl syndrome proteins causes synaptic dysfunction in principal neurons providing possible explanation for cognitive impairment phenotype in BBS patients. Using
more » ... synaptosomal proteomics and immunocytochemistry we demonstrate the presence of Bbs in postsynaptic density of hippocampal neurons. Loss of Bbs results in the significant reduction of dendritic spines in principal neurons of Bbs mice models. Furthermore, we demonstrate that spine deficiency correlates with events that destabilize spine architecture, such as, impaired spine membrane receptors signalling known to be involved in the maintenance of dendritic spines. Finally, we show that voluntary exercise rescues spine deficiency in the neurons. Based on our data, we propose a model in which Bbs proteins, similar to their function in primary cilia, regulate trafficking of signalling receptors into and out of the membrane of dendritic spines, thus providing the basis for synaptic plasticity.
doi:10.1101/580399 fatcat:qvnrfj2mwfhdvhmhymys4me234

The use of extracorporeal shock wave therapy for the treatment of bone marrow oedema — a systematic review and meta-analysis

Jonathan Häußer, Juliane Wieber, Philip Catalá-Lehnen
2021 Journal of Orthopaedic Surgery and Research  
Background Extracorporeal shock wave therapy (ESWT) has been used for various pathologies associated with bone marrow oedema (BME). However, it is still not clear whether ESWT may be favourable in the treatment of BME. Therefore, the aim of this systematic review was to assess the efficacy of ESWT for the treatment of BME. Methods MEDLINE was searched for relevant literature with no time constraints. Both randomized and non-randomized trials were included. Case reports and conference abstracts
more » ... ere excluded. Titles and abstracts were screened and full-text articles of included studies were retrieved. Data on the effect of ESWT on pain, function, and the BME area on magnet resonance imaging were extracted. Results Pain, function, and magnet resonance imaging results all improved across the studies — regardless of whether it was a randomized or non-randomized study. This effect was consistent across multiple pathologies such as osteonecrosis of the femoral head, BME associated with knee osteoarthritis, Kienböck's disease, and osteitis pubis. The meta-analysis showed that pain (after 1 month: weighted mean difference (WMD) = − 2.23, 95% CI − 2.58 to − 1.88, P < 0.0001; after 3–6 month: WMD = − 1.72, 95% CI − 2.52 to − 0.92, P < 0.00001) and function (after 1 month: WMD = − 1.59, 95% CI − 2.04 to − 1.14, P < 0.0001; after 3–6 month: WMD = − 2.06, 95% CI − 3.16 to − 0.96, P = 0.0002; after ≥ 12 month: WMD = − 1.20, 95% CI − 1.83 to − 0.56, P = 0.0002) was reduced in terms of ESWT treatment compared to a control group. Conclusions Based on the available evidence, ESWT may be an adequate option for conservative therapy in pathologies involving BME. Trial registration PROSPERO, CRD42021201719. Registered 23 December 2020
doi:10.1186/s13018-021-02484-5 pmid:34107978 fatcat:5u4cdfnuy5at5ek5ayjldyflzm

Loss of Bardet-Biedl syndrome proteins causes synaptic aberrations in principal neurons

Naila Haq, Christoph Schmidt-Hieber, Fernando J. Sialana, Lorenza Ciani, Janosch P. Heller, Michelle Stewart, Liz Bentley, Sara Wells, Richard J. Rodenburg, Patrick M. Nolan, Elizabeth Forsythe, Michael C. Wu (+6 others)
2019 PLoS Biology  
Bardet-Biedl syndrome (BBS), a ciliopathy, is a rare genetic condition characterised by retinal degeneration, obesity, kidney failure, and cognitive impairment. In spite of progress made in our general understanding of BBS aetiology, the molecular and cellular mechanisms underlying cognitive impairment in BBS remain elusive. Here, we report that the loss of BBS proteins causes synaptic dysfunction in principal neurons, providing a possible explanation for the cognitive impairment phenotype
more » ... ved in BBS patients. Using synaptosomal proteomics and immunocytochemistry, we demonstrate the presence of Bbs proteins in the postsynaptic density (PSD) of hippocampal neurons. Loss of Bbs results in a significant reduction of dendritic spines in principal neurons of Bbs mouse models. Furthermore, we show that spine deficiency correlates with events that destabilise spine architecture, such as impaired spine membrane receptor signalling, known to be involved in the maintenance of dendritic spines. Our findings suggest a role for BBS proteins in dendritic spine homeostasis that may be linked to the cognitive phenotype observed in BBS.
doi:10.1371/journal.pbio.3000414 pmid:31479441 pmcid:PMC6743795 fatcat:25xcyvho5zfjngpju3rqulkxju

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

Nikolaus Mayer, Eddy Ilg, Philip Hausser, Philipp Fischer, Daniel Cremers, Alexey Dosovitskiy, Thomas Brox
2016 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
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
more » ... lly 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.
doi:10.1109/cvpr.2016.438 dblp:conf/cvpr/MayerIHFCDB16 fatcat:okgccilwlbf2pd2phi4ltsbiry

Mutations in CSTA, Encoding Cystatin A, Underlie Exfoliative Ichthyosis and Reveal a Role for This Protease Inhibitor in Cell-Cell Adhesion

Diana C. Blaydon, Daniela Nitoiu, Katja-Martina Eckl, Rita M. Cabral, Philip Bland, Ingrid Hausser, David A. van Heel, Shefali Rajpopat, Judith Fischer, Vinzenz Oji, Alex Zvulunov, Heiko Traupe (+2 others)
2011 American Journal of Human Genetics  
Autosomal-recessive exfoliative ichthyosis presents shortly after birth as dry, scaly skin over most of the body with coarse peeling of nonerythematous skin on the palms and soles, which is exacerbated by excessive moisture and minor trauma. Using whole-genome homozygosity mapping, candidate-gene analysis and deep sequencing, we have identified loss-of-function mutations in the gene for protease inhibitor cystatin A (CSTA) as the underlying genetic cause of exfoliative ichthyosis. We found two
more » ... omozygous mutations, a splice-site and a nonsense mutation, in two consanguineous families of Bedouin and Turkish origin. Electron microscopy of skin biopsies from affected individuals revealed that the level of detachment occurs in the basal and lower suprabasal layers. In addition, in vitro modeling suggests that in the absence of cystatin A protein, there is a cell-cell adhesion defect in human keratinocytes that is particularly prominent when cells are subject to mechanical stress. We show here evidence of a key role for a protease inhibitor in epidermal adhesion within the lower layers of the human epidermis.
doi:10.1016/j.ajhg.2011.09.001 pmid:21944047 pmcid:PMC3188842 fatcat:vqceetot2zbghf54jazkdgk6nq

Aberrant Cortical Activity In Multiple GCaMP6-Expressing Transgenic Mouse Lines [article]

Nicholas A. Steinmetz, Christina Buetfering, Jerome Lecoq, Christian R. Lee, Andrew J. Peters, Elina A. K. Jacobs, Philip Coen, Douglas R. Ollerenshaw, Matthew T. Valley, Saskia E. J. de Vries, Marina Garrett, Jun Zhuang (+22 others)
2017 bioRxiv   pre-print
Treated and untreated animals were imaged as described above (Widefield imaging in the Häusser laboratory).  ...  At the Häusser Laboratory (UCL) all surgical procedures were carried out under a license from the UK Home Office in accordance with the Animal (Scientific Procedures) Act 1986.  ... 
doi:10.1101/138511 fatcat:etzht3dvk5bdzbhdqun232w2aq

Aberrant Cortical Activity in Multiple GCaMP6-Expressing Transgenic Mouse Lines

Nicholas A. Steinmetz, Christina Buetfering, Jerome Lecoq, Christian R. Lee, Andrew J. Peters, Elina A. K. Jacobs, Philip Coen, Douglas R. Ollerenshaw, Matthew T. Valley, Saskia E. J. de Vries, Marina Garrett, Jun Zhuang (+22 others)
2017 eNeuro  
Treated and untreated animals were imaged as described above (Widefield imaging in the Häusser laboratory).  ...  At the Häusser Laboratory (UCL) all surgical procedures were carried out under a license from the UK Home Office in accordance with the Animal (Scientific Procedures) Act 1986.  ... 
doi:10.1523/eneuro.0207-17.2017 pmid:28932809 pmcid:PMC5604087 fatcat:2e3ob5qb4ncsvn5pmneqygvrba

Revised nomenclature and classification of inherited ichthyoses: Results of the First Ichthyosis Consensus Conference in Sorèze 2009

Vinzenz Oji, Gianluca Tadini, Masashi Akiyama, Claudine Blanchet Bardon, Christine Bodemer, Emmanuelle Bourrat, Philippe Coudiere, John J. DiGiovanna, Peter Elias, Judith Fischer, Philip Fleckman, Michal Gina (+25 others)
2010 Journal of American Academy of Dermatology  
doi:10.1016/j.jaad.2009.11.020 pmid:20643494 fatcat:vtcffbjg7fhkhlrmo4re7jqpx4

A One Health overview, facilitating advances in comparative medicine and translational research

Cheryl Stroud, Igor Dmitriev, Elena Kashentseva, Jeffrey N. Bryan, David T. Curiel, Hans Rindt, Carol Reinero, Carolyn J. Henry, Philip J. Bergman, Nicola J. Mason, Josephine S. Gnanandarajah, Julie B. Engiles (+33 others)
2016 Clinical and Translational Medicine  
A One Health overview, facilitating advances in comparative medicine and translational research. Permalink https://escholarship.org/uc/item/1bj665t0 Journal Clinical and translational medicine, 5(Suppl 1)
doi:10.1186/s40169-016-0107-4 pmid:27558513 pmcid:PMC4996801 fatcat:dejan2ufivbnfjzt6lf7vrqulq

Page 461 of The Methodist Review Vol. 50, Issue 3 [page]

1868 The Methodist Review  
MUHL- HAUSSER, Review of the Posthumous Works of Ernst Friedrich Fink.  ...  Kinp, Philip Gallicius. 2. FickER, The Moak Mar- cus, a Reformer of the Fifth Century. 3. Scumrpt, Confession of Johann Fried- rich the Magnanimous, on the Interieur. 4.  ... 

Page 4216 of Psychological Abstracts Vol. 84, Issue 9 [page]

1997 Psychological Abstracts  
Vol 84:31694] Section III: Interaction and its socio-cultural context The relevance of a macrosociological perspective on sexuality for an undestanding of the risks of HIV infection - Fred Deven and Philip  ...  Van Cam- penhoudt, Luc; Cohen, Mitchell; Guizzardi, Gustavo & Hausser, Dominique (Eds.). Taylor & Francis: Washington, DC, 1997. xii, 266 pp.  ... 
« Previous Showing results 1 — 15 out of 626 results