- AbstractURLBibTeXEndNoteBibSonomy1.Luo, C., Zhao, P., Chen, C., Qiao, B., Du, C., Zhang, H., Wu, W., Cai, S., He, B., Rajmohan, S., and Lin, Q. (2020) PULNS: Positive-Unlabeled Learning with Effective Negative Sample Selector. In , pp. 8784–8792.Positive-unlabeled learning (PU learning) is an important case of binary classification where the training data only contains positive and unlabeled samples. The current state-of-the-art approach for PU learning is the cost-sensitive approach, which casts PU learning as a cost-sensitive classification problem and relies on unbiased risk estimator for correcting the bias introduced by the unlabeled samples. However, this approach requires the knowledge of class prior and is subject to the potential label noise. In this paper, we propose a novel PU learning approach dubbed PULNS, equipped with an effective negative sample selector, which is optimized by reinforcement learning. Our PULNS approach employs an effective negative sample selector as the agent responsible for selecting negative samples from the unlabeled data. While the selected, likely negative samples can be used to improve the classifier, the performance of classifier is also used as the reward to improve the selector through the REINFORCE algorithm. By alternating the updates of the selector and the classifier, the performance of both is improved. Extensive experimental studies on 7 real-world application benchmarks demonstrate that PULNS consistently outperforms the current state-of-the-art methods in PU learning, and our experimental results also confirm the effectiveness of the negative sample selector underlying PULNS.
@inproceedings{luo20212020pulns,
abstract = {Positive-unlabeled learning (PU learning) is an important case of binary classification where the training data only contains positive and unlabeled samples. The current state-of-the-art approach for PU learning is the cost-sensitive approach, which casts PU learning as a cost-sensitive classification problem and relies on unbiased risk estimator for correcting the bias introduced by the unlabeled samples. However, this approach requires the knowledge of class prior and is subject to the potential label noise. In this paper, we propose a novel PU learning approach dubbed PULNS, equipped with an effective negative sample selector, which is optimized by reinforcement learning. Our PULNS approach employs an effective negative sample selector as the agent responsible for selecting negative samples from the unlabeled data. While the selected, likely negative samples can be used to improve the classifier, the performance of classifier is also used as the reward to improve the selector through the REINFORCE algorithm. By alternating the updates of the selector and the classifier, the performance of both is improved. Extensive experimental studies on 7 real-world application benchmarks demonstrate that PULNS consistently outperforms the current state-of-the-art methods in PU learning, and our experimental results also confirm the effectiveness of the negative sample selector underlying PULNS.},
author = {Luo, Chuan and Zhao, Pu and Chen, Chen and Qiao, Bo and Du, Chao and Zhang, Hongyu and Wu, Wei and Cai, Shaowei and He, Bing and Rajmohan, Saravanakumar and Lin, Qingwei},
keywords = {l3s learning leibnizailab},
number = 10,
pages = {8784-8792},
title = {PULNS: Positive-Unlabeled Learning with Effective Negative Sample Selector},
type = {Publication},
volume = 35,
year = {2021, 2020}
}%0 Conference Paper
%1 luo20212020pulns
%A Luo, Chuan
%A Zhao, Pu
%A Chen, Chen
%A Qiao, Bo
%A Du, Chao
%A Zhang, Hongyu
%A Wu, Wei
%A Cai, Shaowei
%A He, Bing
%A Rajmohan, Saravanakumar
%A Lin, Qingwei
%D 2021, 2020
%N 10
%P 8784-8792
%T PULNS: Positive-Unlabeled Learning with Effective Negative Sample Selector
%U https://ojs.aaai.org/index.php/AAAI/article/view/17064
%V 35
%X Positive-unlabeled learning (PU learning) is an important case of binary classification where the training data only contains positive and unlabeled samples. The current state-of-the-art approach for PU learning is the cost-sensitive approach, which casts PU learning as a cost-sensitive classification problem and relies on unbiased risk estimator for correcting the bias introduced by the unlabeled samples. However, this approach requires the knowledge of class prior and is subject to the potential label noise. In this paper, we propose a novel PU learning approach dubbed PULNS, equipped with an effective negative sample selector, which is optimized by reinforcement learning. Our PULNS approach employs an effective negative sample selector as the agent responsible for selecting negative samples from the unlabeled data. While the selected, likely negative samples can be used to improve the classifier, the performance of classifier is also used as the reward to improve the selector through the REINFORCE algorithm. By alternating the updates of the selector and the classifier, the performance of both is improved. Extensive experimental studies on 7 real-world application benchmarks demonstrate that PULNS consistently outperforms the current state-of-the-art methods in PU learning, and our experimental results also confirm the effectiveness of the negative sample selector underlying PULNS. - BibTeXEndNoteBibSonomy1.Olatunji, I. E., Nejdl, W., and Khosla, M. (2021) Membership inference attack on graph neural networks, ICLR workshop on Distributed and Private Machine Learning (DPML) 2021.
@article{olatunji2021membership,
author = {Olatunji, Iyiola E and Nejdl, Wolfgang and Khosla, Megha},
journal = {ICLR workshop on Distributed and Private Machine Learning (DPML) 2021},
keywords = {l3s leibnizailab myown},
title = {Membership inference attack on graph neural networks},
year = 2021
}%0 Journal Article
%1 olatunji2021membership
%A Olatunji, Iyiola E
%A Nejdl, Wolfgang
%A Khosla, Megha
%D 2021
%J ICLR workshop on Distributed and Private Machine Learning (DPML) 2021
%T Membership inference attack on graph neural networks - BibTeXEndNoteBibSonomy1.Decker, M., Lammens, T., Ferster, A., Erlacher, M., Yoshimi, A., Niemeyer, C. M., Ernst, M. P. T., Raaijmakers, M. H. G. P., Duployez, N., Flaum, A., Steinemann, D., Schlegelberger, B., Illig, T., and Ripperger, T. (2021) Functional classification of RUNX1 variants in familial platelet disorder with associated myeloid malignancies, Leukemia.
@article{display:block).nova-c-button--color-blue.nova-c-button--theme-solid:focus-visible-webkit-box-shadow:0,
author = {Decker, Melanie and Lammens, Tim and Ferster, Alina and Erlacher, Miriam and Yoshimi, Ayami and Niemeyer, Charlotte M. and Ernst, Martijn P. T. and Raaijmakers, Marc H. G. P. and Duployez, Nicolas and Flaum, Andreas and Steinemann, Doris and Schlegelberger, Brigitte and Illig, Thomas and Ripperger, Tim},
journal = {Leukemia},
keywords = {l3s leibnizailab leukemia},
month = {jul},
title = {Functional classification of RUNX1 variants in familial platelet disorder with associated myeloid malignancies},
year = 2021
}%0 Journal Article
%1 display:block).nova-c-button--color-blue.nova-c-button--theme-solid:focus-visible-webkit-box-shadow:0
%A Decker, Melanie
%A Lammens, Tim
%A Ferster, Alina
%A Erlacher, Miriam
%A Yoshimi, Ayami
%A Niemeyer, Charlotte M.
%A Ernst, Martijn P. T.
%A Raaijmakers, Marc H. G. P.
%A Duployez, Nicolas
%A Flaum, Andreas
%A Steinemann, Doris
%A Schlegelberger, Brigitte
%A Illig, Thomas
%A Ripperger, Tim
%D 2021
%J Leukemia
%T Functional classification of RUNX1 variants in familial platelet disorder with associated myeloid malignancies - AbstractURLBibTeXEndNoteBibSonomy1.Zhao, B., van der Aa, H., Nguyen, T. T., Nguyen, Q. V. H., and Weidlich, M. (2021) EIRES: Efficient Integration of Remote Data in Event Stream Processing. In Proceedings of the 2021 International Conference on Management of Data, ACM.To support reactive and predictive applications, complex event processing (CEP) systems detect patterns in event streams based on predefined queries. To determine the events that constitute a query match, their payload data may need to be assessed together with data from remote sources. Such dependencies are problematic, since waiting for remote data to be fetched interrupts the processing of the stream. Yet, without event selection based on remote data, the query state to maintain may grow exponentially. In either case, the performance of the CEP system degrades drastically. To tackle these issues, we present EIRES, a framework for efficient integration of static data from remote sources in CEP. It employs a cost-model to determine when to fetch certain remote data elements and how long to keep them in a cache for future use. EIRES combines strategies for (i) prefetching that queries remote data based on anticipated use and (ii) lazy evaluation that postpones the event selection based on remote data without interrupting the stream processing. Our experiments indicate that the combination of these strategies improves the latency of query evaluation by up to 3,725x for synthetic data and 47x for real-world data.
@inproceedings{Zhao_2021,
abstract = {To support reactive and predictive applications, complex event processing (CEP) systems detect patterns in event streams based on predefined queries. To determine the events that constitute a query match, their payload data may need to be assessed together with data from remote sources. Such dependencies are problematic, since waiting for remote data to be fetched interrupts the processing of the stream. Yet, without event selection based on remote data, the query state to maintain may grow exponentially. In either case, the performance of the CEP system degrades drastically. To tackle these issues, we present EIRES, a framework for efficient integration of static data from remote sources in CEP. It employs a cost-model to determine when to fetch certain remote data elements and how long to keep them in a cache for future use. EIRES combines strategies for (i) prefetching that queries remote data based on anticipated use and (ii) lazy evaluation that postpones the event selection based on remote data without interrupting the stream processing. Our experiments indicate that the combination of these strategies improves the latency of query evaluation by up to 3,725x for synthetic data and 47x for real-world data.},
author = {Zhao, Bo and van der Aa, Han and Nguyen, Thanh Tam and Nguyen, Quoc Viet Hung and Weidlich, Matthias},
booktitle = {Proceedings of the 2021 International Conference on Management of Data},
keywords = {l3s leibnizailab},
month = {june},
publisher = {ACM},
title = {EIRES: Efficient Integration of Remote Data in Event Stream Processing},
year = 2021
}%0 Conference Paper
%1 Zhao_2021
%A Zhao, Bo
%A van der Aa, Han
%A Nguyen, Thanh Tam
%A Nguyen, Quoc Viet Hung
%A Weidlich, Matthias
%B Proceedings of the 2021 International Conference on Management of Data
%D 2021
%I ACM
%R 10.1145/3448016.3457304
%T EIRES: Efficient Integration of Remote Data in Event Stream Processing
%U https://doi.org/10.1145%2F3448016.3457304
%X To support reactive and predictive applications, complex event processing (CEP) systems detect patterns in event streams based on predefined queries. To determine the events that constitute a query match, their payload data may need to be assessed together with data from remote sources. Such dependencies are problematic, since waiting for remote data to be fetched interrupts the processing of the stream. Yet, without event selection based on remote data, the query state to maintain may grow exponentially. In either case, the performance of the CEP system degrades drastically. To tackle these issues, we present EIRES, a framework for efficient integration of static data from remote sources in CEP. It employs a cost-model to determine when to fetch certain remote data elements and how long to keep them in a cache for future use. EIRES combines strategies for (i) prefetching that queries remote data based on anticipated use and (ii) lazy evaluation that postpones the event selection based on remote data without interrupting the stream processing. Our experiments indicate that the combination of these strategies improves the latency of query evaluation by up to 3,725x for synthetic data and 47x for real-world data. - URLBibTeXEndNoteBibSonomy1.Luo, C., Zhao, P., Qiao, B., Wu, Y., Zhang, H., Wu, W., Lu, W., Dang, Y., Rajmohan, S., Lin, Q., and Zhang, D. (2021) NTAM: Neighborhood-Temporal Attention Model for Disk Failure Prediction in Cloud Platforms. In Proceedings of the Web Conference 2021, ACM.
@inproceedings{Luo_2021,
author = {Luo, Chuan and Zhao, Pu and Qiao, Bo and Wu, Youjiang and Zhang, Hongyu and Wu, Wei and Lu, Weihai and Dang, Yingnong and Rajmohan, Saravanakumar and Lin, Qingwei and Zhang, Dongmei},
booktitle = {Proceedings of the Web Conference 2021},
keywords = {cloudplatforms l3s leibnizailab},
month = {apr},
publisher = {ACM},
title = {NTAM: Neighborhood-Temporal Attention Model for Disk Failure Prediction in Cloud Platforms},
year = 2021
}%0 Conference Paper
%1 Luo_2021
%A Luo, Chuan
%A Zhao, Pu
%A Qiao, Bo
%A Wu, Youjiang
%A Zhang, Hongyu
%A Wu, Wei
%A Lu, Weihai
%A Dang, Yingnong
%A Rajmohan, Saravanakumar
%A Lin, Qingwei
%A Zhang, Dongmei
%B Proceedings of the Web Conference 2021
%D 2021
%I ACM
%R 10.1145/3442381.3449867
%T NTAM: Neighborhood-Temporal Attention Model for Disk Failure Prediction in Cloud Platforms
%U https://doi.org/10.1145%2F3442381.3449867 - AbstractURLBibTeXEndNoteBibSonomy1.Luo, C., Lin, J., Cai, S., Chen, X., He, B., Qiao, B., Zhao, P., Lin, Q., Zhang, H., Wu, W., Rajmohan, S., and Zhang, D. (2021) AutoCCAG: An Automated Approach to Constrained Covering Array Generation. In 2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE), pp. 201–212.Combinatorial interaction testing (CIT) is an important technique for testing highly configurable software systems with demonstrated effectiveness in practice. The goal of CIT is to generate test cases covering the interactions of configuration options, under certain hard constraints. In this context, constrained covering arrays (CCAs) are frequently used as test cases in CIT. Constrained Covering Array Generation (CCAG) is an NP-hard combinatorial optimization problem, solving which requires an effective method for generating small CCAs. In particular, effectively solving t-way CCAG with t>=4 is even more challenging. Inspired by the success of automated algorithm configuration and automated algorithm selection in solving combinatorial optimization problems, in this paper, we investigate the efficacy of automated algorithm configuration and automated algorithm selection for the CCAG problem, and propose a novel, automated CCAG approach called AutoCCAG. Extensive experiments on public benchmarks show that AutoCCAG can find much smaller-sized CCAs than current state-of-the-art approaches, indicating the effectiveness of AutoCCAG. More encouragingly, to our best knowledge, our paper reports the first results for CCAG with a high coverage strength (i.e., 5-way CCAG) on public benchmarks. Our results demonstrate that AutoCCAG can bring considerable benefits in testing highly configurable software systems.
@inproceedings{9402109,
abstract = {Combinatorial interaction testing (CIT) is an important technique for testing highly configurable software systems with demonstrated effectiveness in practice. The goal of CIT is to generate test cases covering the interactions of configuration options, under certain hard constraints. In this context, constrained covering arrays (CCAs) are frequently used as test cases in CIT. Constrained Covering Array Generation (CCAG) is an NP-hard combinatorial optimization problem, solving which requires an effective method for generating small CCAs. In particular, effectively solving t-way CCAG with t>=4 is even more challenging. Inspired by the success of automated algorithm configuration and automated algorithm selection in solving combinatorial optimization problems, in this paper, we investigate the efficacy of automated algorithm configuration and automated algorithm selection for the CCAG problem, and propose a novel, automated CCAG approach called AutoCCAG. Extensive experiments on public benchmarks show that AutoCCAG can find much smaller-sized CCAs than current state-of-the-art approaches, indicating the effectiveness of AutoCCAG. More encouragingly, to our best knowledge, our paper reports the first results for CCAG with a high coverage strength (i.e., 5-way CCAG) on public benchmarks. Our results demonstrate that AutoCCAG can bring considerable benefits in testing highly configurable software systems.},
author = {Luo, Chuan and Lin, Jinkun and Cai, Shaowei and Chen, Xin and He, Bing and Qiao, Bo and Zhao, Pu and Lin, Qingwei and Zhang, Hongyu and Wu, Wei and Rajmohan, Saravanakumar and Zhang, Dongmei},
booktitle = {2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE)},
keywords = {l3s leibnizailab},
month = {may},
pages = {201-212},
title = {AutoCCAG: An Automated Approach to Constrained Covering Array Generation},
year = 2021
}%0 Conference Paper
%1 9402109
%A Luo, Chuan
%A Lin, Jinkun
%A Cai, Shaowei
%A Chen, Xin
%A He, Bing
%A Qiao, Bo
%A Zhao, Pu
%A Lin, Qingwei
%A Zhang, Hongyu
%A Wu, Wei
%A Rajmohan, Saravanakumar
%A Zhang, Dongmei
%B 2021 IEEE/ACM 43rd International Conference on Software Engineering (ICSE)
%D 2021
%P 201-212
%R 10.1109/ICSE43902.2021.00030
%T AutoCCAG: An Automated Approach to Constrained Covering Array Generation
%U https://ieeexplore.ieee.org/document/9402109
%X Combinatorial interaction testing (CIT) is an important technique for testing highly configurable software systems with demonstrated effectiveness in practice. The goal of CIT is to generate test cases covering the interactions of configuration options, under certain hard constraints. In this context, constrained covering arrays (CCAs) are frequently used as test cases in CIT. Constrained Covering Array Generation (CCAG) is an NP-hard combinatorial optimization problem, solving which requires an effective method for generating small CCAs. In particular, effectively solving t-way CCAG with t>=4 is even more challenging. Inspired by the success of automated algorithm configuration and automated algorithm selection in solving combinatorial optimization problems, in this paper, we investigate the efficacy of automated algorithm configuration and automated algorithm selection for the CCAG problem, and propose a novel, automated CCAG approach called AutoCCAG. Extensive experiments on public benchmarks show that AutoCCAG can find much smaller-sized CCAs than current state-of-the-art approaches, indicating the effectiveness of AutoCCAG. More encouragingly, to our best knowledge, our paper reports the first results for CCAG with a high coverage strength (i.e., 5-way CCAG) on public benchmarks. Our results demonstrate that AutoCCAG can bring considerable benefits in testing highly configurable software systems. - BibTeXEndNoteBibSonomy1.Wandt, B., Rudolph, M., Zell, P., Rhodin, H., and Rosenhahn, B. (2021) CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild. In Computer Vision and Pattern Recognition (CVPR).
@inproceedings{WanRud2021a,
author = {Wandt, Bastian and Rudolph, Marco and Zell, Petrissa and Rhodin, Helge and Rosenhahn, Bodo},
booktitle = {Computer Vision and Pattern Recognition (CVPR)},
keywords = {l3s leibnizailab},
month = {june},
title = {CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild},
year = 2021
}%0 Conference Paper
%1 WanRud2021a
%A Wandt, Bastian
%A Rudolph, Marco
%A Zell, Petrissa
%A Rhodin, Helge
%A Rosenhahn, Bodo
%B Computer Vision and Pattern Recognition (CVPR)
%D 2021
%T CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild - AbstractURLBibTeXEndNoteBibSonomy1.Hachmann, H., Krüger, B., Rosenhahn, B., and Nogueira, W. (2021) Localization Of Cochlear Implant Electrodes From Cone Beam Computed Tomography Using Particle Belief Propagation. In 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), pp. 593–597.Cochlear implants (CIs) are implantable medical devices that can restore the hearing sense of people suffering from profound hearing loss. The CI uses a set of electrode contacts placed inside the cochlea to stimulate the auditory nerve with current pulses. The exact location of these electrodes may be an important parameter to improve and predict the performance with these devices. Currently the methods used in clinics to characterize the geometry of the cochlea as well as to estimate the electrode positions are manual, error-prone and time consuming.We propose a Markov random field (MRF) model for CI electrode localization for cone beam computed tomography (CBCT) data-sets. Intensity and shape of electrodes are included as prior knowledge as well as distance and angles between contacts. MRF inference is based on slice sampling particle belief propagation and guided by several heuristics. A stochastic search finds the best maximum a posteriori estimation among sampled MRF realizations.We evaluate our algorithm on synthetic and real CBCT data-sets and compare its performance with two state of the art algorithms. An increase of localization precision up to 31.5% (mean), or 48.6% (median) respectively, on real CBCT data-sets is shown.
@inproceedings{9433845,
abstract = {Cochlear implants (CIs) are implantable medical devices that can restore the hearing sense of people suffering from profound hearing loss. The CI uses a set of electrode contacts placed inside the cochlea to stimulate the auditory nerve with current pulses. The exact location of these electrodes may be an important parameter to improve and predict the performance with these devices. Currently the methods used in clinics to characterize the geometry of the cochlea as well as to estimate the electrode positions are manual, error-prone and time consuming.We propose a Markov random field (MRF) model for CI electrode localization for cone beam computed tomography (CBCT) data-sets. Intensity and shape of electrodes are included as prior knowledge as well as distance and angles between contacts. MRF inference is based on slice sampling particle belief propagation and guided by several heuristics. A stochastic search finds the best maximum a posteriori estimation among sampled MRF realizations.We evaluate our algorithm on synthetic and real CBCT data-sets and compare its performance with two state of the art algorithms. An increase of localization precision up to 31.5% (mean), or 48.6% (median) respectively, on real CBCT data-sets is shown.},
author = {Hachmann, Hendrik and Krüger, Benjamin and Rosenhahn, Bodo and Nogueira, Waldo},
booktitle = {2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)},
keywords = {cochlear l3s leibnizailab},
month = {April},
pages = {593-597},
title = {Localization Of Cochlear Implant Electrodes From Cone Beam Computed Tomography Using Particle Belief Propagation},
year = 2021
}%0 Conference Paper
%1 9433845
%A Hachmann, Hendrik
%A Krüger, Benjamin
%A Rosenhahn, Bodo
%A Nogueira, Waldo
%B 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI)
%D 2021
%P 593-597
%R 10.1109/ISBI48211.2021.9433845
%T Localization Of Cochlear Implant Electrodes From Cone Beam Computed Tomography Using Particle Belief Propagation
%U https://ieeexplore.ieee.org/abstract/document/9433845
%X Cochlear implants (CIs) are implantable medical devices that can restore the hearing sense of people suffering from profound hearing loss. The CI uses a set of electrode contacts placed inside the cochlea to stimulate the auditory nerve with current pulses. The exact location of these electrodes may be an important parameter to improve and predict the performance with these devices. Currently the methods used in clinics to characterize the geometry of the cochlea as well as to estimate the electrode positions are manual, error-prone and time consuming.We propose a Markov random field (MRF) model for CI electrode localization for cone beam computed tomography (CBCT) data-sets. Intensity and shape of electrodes are included as prior knowledge as well as distance and angles between contacts. MRF inference is based on slice sampling particle belief propagation and guided by several heuristics. A stochastic search finds the best maximum a posteriori estimation among sampled MRF realizations.We evaluate our algorithm on synthetic and real CBCT data-sets and compare its performance with two state of the art algorithms. An increase of localization precision up to 31.5% (mean), or 48.6% (median) respectively, on real CBCT data-sets is shown. - AbstractBibTeXEndNoteBibSonomy1.Becker, M., Strengert, M., Junker, D., Kaiser, P. D., Kerrinnes, T., Traenkle, B., Dinter, H., Häring, J., Ghozzi, S., Zeck, A., Weise, F., Peter, A., Hörber, S., Fink, S., Ruoff, F., Dulovic, A., Bakchoul, T., Baillot, A., Lohse, S., Cornberg, M., Illig, T., Gottlieb, J., Smola, S., Karch, A., Berger, K., Rammensee, H.-G., Schenke-Layland, K., Nelde, A., Märklin, M., Heitmann, J. S., Walz, J. S., Templin, M., Joos, T. O., Rothbauer, U., Krause, G., and Schneiderhan-Marra, N. (2021) Exploring beyond clinical routine SARS-CoV-2 serology using MultiCoV-Ab to evaluate endemic coronavirus cross-reactivity, Nat Commun 12.The humoral immune response to SARS-CoV-2 is a benchmark for immunity and detailed analysis is required to understand the manifestation and progression of COVID-19, monitor seroconversion within the general population, and support vaccine development. The majority of currently available commercial serological assays only quantify the SARS-CoV-2 antibody response against individual antigens, limiting our understanding of the immune response. To overcome this, we have developed a multiplex immunoassay (MultiCoV-Ab) including spike and nucleocapsid proteins of SARS-CoV-2 and the endemic human coronaviruses. Compared to three broadly used commercial in vitro diagnostic tests, our MultiCoV-Ab achieves a higher sensitivity and specificity when analyzing a well-characterized sample set of SARS-CoV-2 infected and uninfected individuals. We find a high response against endemic coronaviruses in our sample set, but no consistent cross-reactive IgG response patterns against SARS-CoV-2. Here we show a robust, high-content-enabled, antigen-saving multiplex assay suited to both monitoring vaccination studies and facilitating epidemiologic screenings for humoral immunity towards pandemic and endemic coronaviruses.
@article{noauthororeditor,
abstract = {The humoral immune response to SARS-CoV-2 is a benchmark for immunity and detailed analysis is required to understand the manifestation and progression of COVID-19, monitor seroconversion within the general population, and support vaccine development. The majority of currently available commercial serological assays only quantify the SARS-CoV-2 antibody response against individual antigens, limiting our understanding of the immune response. To overcome this, we have developed a multiplex immunoassay (MultiCoV-Ab) including spike and nucleocapsid proteins of SARS-CoV-2 and the endemic human coronaviruses. Compared to three broadly used commercial in vitro diagnostic tests, our MultiCoV-Ab achieves a higher sensitivity and specificity when analyzing a well-characterized sample set of SARS-CoV-2 infected and uninfected individuals. We find a high response against endemic coronaviruses in our sample set, but no consistent cross-reactive IgG response patterns against SARS-CoV-2. Here we show a robust, high-content-enabled, antigen-saving multiplex assay suited to both monitoring vaccination studies and facilitating epidemiologic screenings for humoral immunity towards pandemic and endemic coronaviruses.},
author = {Becker, Matthias and Strengert, Monika and Junker, Daniel and Kaiser, Philipp D. and Kerrinnes, Tobias and Traenkle, Bjoern and Dinter, Heiko and Häring, Julia and Ghozzi, Stéphane and Zeck, Anne and Weise, Frank and Peter, Andreas and Hörber, Sebastian and Fink, Simon and Ruoff, Felix and Dulovic, Alex and Bakchoul, Tamam and Baillot, Armin and Lohse, Stefan and Cornberg, Markus and Illig, Thomas and Gottlieb, Jens and Smola, Sigrun and Karch, André and Berger, Klaus and Rammensee, Hans-Georg and Schenke-Layland, Katja and Nelde, Annika and Märklin, Melanie and Heitmann, Jonas S. and Walz, Juliane S. and Templin, Markus and Joos, Thomas O. and Rothbauer, Ulrich and Krause, Gérard and Schneiderhan-Marra, Nicole},
journal = {Nat Commun},
keywords = {covid l3s leibnizailab},
month = {feb},
number = 1152,
title = {Exploring beyond clinical routine SARS-CoV-2 serology using MultiCoV-Ab to evaluate endemic coronavirus cross-reactivity},
volume = 12,
year = 2021
}%0 Journal Article
%1 noauthororeditor
%A Becker, Matthias
%A Strengert, Monika
%A Junker, Daniel
%A Kaiser, Philipp D.
%A Kerrinnes, Tobias
%A Traenkle, Bjoern
%A Dinter, Heiko
%A Häring, Julia
%A Ghozzi, Stéphane
%A Zeck, Anne
%A Weise, Frank
%A Peter, Andreas
%A Hörber, Sebastian
%A Fink, Simon
%A Ruoff, Felix
%A Dulovic, Alex
%A Bakchoul, Tamam
%A Baillot, Armin
%A Lohse, Stefan
%A Cornberg, Markus
%A Illig, Thomas
%A Gottlieb, Jens
%A Smola, Sigrun
%A Karch, André
%A Berger, Klaus
%A Rammensee, Hans-Georg
%A Schenke-Layland, Katja
%A Nelde, Annika
%A Märklin, Melanie
%A Heitmann, Jonas S.
%A Walz, Juliane S.
%A Templin, Markus
%A Joos, Thomas O.
%A Rothbauer, Ulrich
%A Krause, Gérard
%A Schneiderhan-Marra, Nicole
%D 2021
%J Nat Commun
%N 1152
%R https://doi.org/10.1038/s41467-021-20973-3
%T Exploring beyond clinical routine SARS-CoV-2 serology using MultiCoV-Ab to evaluate endemic coronavirus cross-reactivity
%V 12
%X The humoral immune response to SARS-CoV-2 is a benchmark for immunity and detailed analysis is required to understand the manifestation and progression of COVID-19, monitor seroconversion within the general population, and support vaccine development. The majority of currently available commercial serological assays only quantify the SARS-CoV-2 antibody response against individual antigens, limiting our understanding of the immune response. To overcome this, we have developed a multiplex immunoassay (MultiCoV-Ab) including spike and nucleocapsid proteins of SARS-CoV-2 and the endemic human coronaviruses. Compared to three broadly used commercial in vitro diagnostic tests, our MultiCoV-Ab achieves a higher sensitivity and specificity when analyzing a well-characterized sample set of SARS-CoV-2 infected and uninfected individuals. We find a high response against endemic coronaviruses in our sample set, but no consistent cross-reactive IgG response patterns against SARS-CoV-2. Here we show a robust, high-content-enabled, antigen-saving multiplex assay suited to both monitoring vaccination studies and facilitating epidemiologic screenings for humoral immunity towards pandemic and endemic coronaviruses. - BibTeXEndNoteBibSonomy1.Singh, J., Wang, Z., Khosla, M., and Anand, A. (2021) Extracting per Query Valid Explanations for Blackbox Learning-to-Rank Models. In International Conference on the Theory of Information Retrieval.
@inproceedings{noauthororeditor2021extracting,
author = {Singh, Jaspreet and Wang, Zhenye and Khosla, Megha and Anand, Avishek},
booktitle = {International Conference on the Theory of Information Retrieval},
keywords = {leibnizailab myown},
title = {Extracting per Query Valid Explanations for Blackbox Learning-to-Rank Models},
year = 2021
}%0 Conference Paper
%1 noauthororeditor2021extracting
%A Singh, Jaspreet
%A Wang, Zhenye
%A Khosla, Megha
%A Anand, Avishek
%B International Conference on the Theory of Information Retrieval
%D 2021
%T Extracting per Query Valid Explanations for Blackbox Learning-to-Rank Models - BibTeXEndNoteBibSonomy1.Bellinghausen, C., Pletz, M. W., Rupp, J., Witzenrath, M., Welsch, C., Zeuzem, S., Trebicka, J., Rohde, G. G. U., and of the CAPNETZ study group, M. (2021) Chronic liver disease negatively affects outcome in hospitalised patients with community-acquired pneumonia, Gut 70, 221–222.
@article{noauthororeditor,
author = {Bellinghausen, Carla and Pletz, Mathias W. and Rupp, Jan and Witzenrath, Martin and Welsch, Christoph and Zeuzem, Stefan and Trebicka, Jonel and Rohde, Gernot G. U. and of the CAPNETZ study group, Members},
journal = {Gut},
keywords = {l3s leibnizailab},
month = {jan},
number = 1,
pages = {221-222},
title = {Chronic liver disease negatively affects outcome in hospitalised patients with community-acquired pneumonia},
volume = 70,
year = 2021
}%0 Journal Article
%1 noauthororeditor
%A Bellinghausen, Carla
%A Pletz, Mathias W.
%A Rupp, Jan
%A Witzenrath, Martin
%A Welsch, Christoph
%A Zeuzem, Stefan
%A Trebicka, Jonel
%A Rohde, Gernot G. U.
%A of the CAPNETZ study group, Members
%D 2021
%J Gut
%N 1
%P 221-222
%R doi: 10.1136/gutjnl-2020-320876
%T Chronic liver disease negatively affects outcome in hospitalised patients with community-acquired pneumonia
%V 70 - AbstractURLBibTeXEndNoteBibSonomy1.Holzapfel, C., Sag, S., Graf-Schindler, J., Fischer, M., Drabsch, T., Illig, T., Grallert, H., Stecher, L., Strack, C., Caterson, I., Jebb, S., Hauner, H., and Baessler, A. (2021) Association between single nucleotide polymorphisms and weight reduction in behavioural interventions—a pooled analysis, Nutrients, MDPI 13.Knowledge of the association between single nucleotide polymorphisms (SNPs) and weight loss is limited. The aim was to analyse whether selected obesity-associated SNPs within the fat mass and obesity-associated (FTO), transmembrane protein 18 (TMEM18), melanocortin-4 receptor (MC4R), SEC16 homolog B (SEC16B), and brain-derived neurotrophic factor (BDNF) gene are associated with anthropometric changes during behavioural intervention for weight loss. genetic and anthropometric data from 576 individuals with overweight and obesity from four lifestyle interventions were obtained. A genetic predisposition score (GPS) was calculated. Our results show that study participants had a mean age of 48.2 ± 12.6 years and a mean baseline body mass index of 33.9 ± 6.4 kg/m2. Mean weight reduction after 12 months was −7.7 ± 10.9 kg. After 12 months of intervention, the MC4R SNPs rs571312 and rs17782313 were significantly associated with a greater decrease in body weight and BMI (p = 0.012, p = 0.011, respectively). The investigated SNPs within the other four genetic loci showed no statistically significant association with changes in anthropometric parameters. The GPS showed no statistically significant association with weight reduction. In conclusion there was no consistent evidence for statistically significant associations of SNPs with anthropometric changes during a behavioural intervention. It seems that other factors play a more significant in weight management than the investigated SNPs.
@article{holzapfel2021association,
abstract = {Knowledge of the association between single nucleotide polymorphisms (SNPs) and weight loss is limited. The aim was to analyse whether selected obesity-associated SNPs within the fat mass and obesity-associated (FTO), transmembrane protein 18 (TMEM18), melanocortin-4 receptor (MC4R), SEC16 homolog B (SEC16B), and brain-derived neurotrophic factor (BDNF) gene are associated with anthropometric changes during behavioural intervention for weight loss. genetic and anthropometric data from 576 individuals with overweight and obesity from four lifestyle interventions were obtained. A genetic predisposition score (GPS) was calculated. Our results show that study participants had a mean age of 48.2 ± 12.6 years and a mean baseline body mass index of 33.9 ± 6.4 kg/m2. Mean weight reduction after 12 months was −7.7 ± 10.9 kg. After 12 months of intervention, the MC4R SNPs rs571312 and rs17782313 were significantly associated with a greater decrease in body weight and BMI (p = 0.012, p = 0.011, respectively). The investigated SNPs within the other four genetic loci showed no statistically significant association with changes in anthropometric parameters. The GPS showed no statistically significant association with weight reduction. In conclusion there was no consistent evidence for statistically significant associations of SNPs with anthropometric changes during a behavioural intervention. It seems that other factors play a more significant in weight management than the investigated SNPs.},
author = {Holzapfel, C and Sag, S and Graf-Schindler, J and Fischer, M and Drabsch, T and Illig, T and Grallert, H and Stecher, L and Strack, C and Caterson, ID and Jebb, SA and Hauner, H and Baessler, A},
journal = {Nutrients},
keywords = {l3s leibnizailab},
number = 3,
publisher = {MDPI},
title = {Association between single nucleotide polymorphisms and weight reduction in behavioural interventions—a pooled analysis},
type = {Publication},
volume = 13,
year = 2021
}%0 Journal Article
%1 holzapfel2021association
%A Holzapfel, C
%A Sag, S
%A Graf-Schindler, J
%A Fischer, M
%A Drabsch, T
%A Illig, T
%A Grallert, H
%A Stecher, L
%A Strack, C
%A Caterson, ID
%A Jebb, SA
%A Hauner, H
%A Baessler, A
%D 2021
%I MDPI
%J Nutrients
%N 3
%R 10.3390/nu13030819
%T Association between single nucleotide polymorphisms and weight reduction in behavioural interventions—a pooled analysis
%U https://ora.ox.ac.uk/objects/uuid:98987391-ba79-4686-a819-5fbe2a79ff45
%V 13
%X Knowledge of the association between single nucleotide polymorphisms (SNPs) and weight loss is limited. The aim was to analyse whether selected obesity-associated SNPs within the fat mass and obesity-associated (FTO), transmembrane protein 18 (TMEM18), melanocortin-4 receptor (MC4R), SEC16 homolog B (SEC16B), and brain-derived neurotrophic factor (BDNF) gene are associated with anthropometric changes during behavioural intervention for weight loss. genetic and anthropometric data from 576 individuals with overweight and obesity from four lifestyle interventions were obtained. A genetic predisposition score (GPS) was calculated. Our results show that study participants had a mean age of 48.2 ± 12.6 years and a mean baseline body mass index of 33.9 ± 6.4 kg/m2. Mean weight reduction after 12 months was −7.7 ± 10.9 kg. After 12 months of intervention, the MC4R SNPs rs571312 and rs17782313 were significantly associated with a greater decrease in body weight and BMI (p = 0.012, p = 0.011, respectively). The investigated SNPs within the other four genetic loci showed no statistically significant association with changes in anthropometric parameters. The GPS showed no statistically significant association with weight reduction. In conclusion there was no consistent evidence for statistically significant associations of SNPs with anthropometric changes during a behavioural intervention. It seems that other factors play a more significant in weight management than the investigated SNPs. - AbstractBibTeXEndNoteBibSonomy1.Warnstorf, D., Bawadi, R., Schienke, A., Strasser, R., Schmidt, G., Illig, T., Tauscher, M., Thol, F., Heuser, M., Steinemann, D., Davenport, C., Schlegelberger, B., Behrens, Y. L., and Göhring, G. (2021) Unbalanced translocation der(5;17) resulting in a TP53 loss as recurrent aberration in myelodysplastic syndrome and acute myeloid leukemia with complex karyotype, Genes Chromosomes Cancer 60, 452–457.A complex karyotype, detected in myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML), is associated with a reduced median survival. The most frequent chromosomal aberrations in complex karyotypes are deletions of 5q and 17p harboring the tumor suppressor gene TP53. The unbalanced translocation der(5;17) involving chromosome 5q and 17p is a recurrent aberration in MDS/AML, resulting in TP53 loss. We analyzed the karyotypes of 178 patients with an unbalanced translocation der(5;17) using fluorescence R−/G-banding analysis. Whenever possible, fluorescence in situ hybridization (FISH) (n = 138/141), multicolor FISH (n = 8), telomere length measurement (n = 9), targeted DNA sequencing (n = 13), array-CGH (n = 7) and targeted RNA sequencing (n = 2) were conducted. The der(5;17) aberration was accompanied with loss of genetic material in 7q (53%), −7 (27%), gain of 21q (29%), +8 (17%) and − 18 (16%) and all analyzed patients (n = 13) showed a (likely) pathogenic variant inTP53. The der(5;17) cohort showed significantly shortened telomeres in comparison to the healthy age-matched controls (P < .05), but there was no significant telomere shortening in comparison to MDS/AML patients with a complex karyotype without der(5;17). No fusion genes resulted from the unbalanced translocation. This study demonstrates that the unbalanced translocation der(5;17) is associated with a biallelic inactivation of TP53 due to a deletion of TP53 in one allele and a pathogenic variant of the second TP53 allele. Since the breakpoints are located within (near-) heterochromatic regions, alterations of DNA methylation or histone modifications may be involved in the generation of der(5;17).
@article{unbalanced,
abstract = {A complex karyotype, detected in myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML), is associated with a reduced median survival. The most frequent chromosomal aberrations in complex karyotypes are deletions of 5q and 17p harboring the tumor suppressor gene TP53. The unbalanced translocation der(5;17) involving chromosome 5q and 17p is a recurrent aberration in MDS/AML, resulting in TP53 loss. We analyzed the karyotypes of 178 patients with an unbalanced translocation der(5;17) using fluorescence R−/G-banding analysis. Whenever possible, fluorescence in situ hybridization (FISH) (n = 138/141), multicolor FISH (n = 8), telomere length measurement (n = 9), targeted DNA sequencing (n = 13), array-CGH (n = 7) and targeted RNA sequencing (n = 2) were conducted. The der(5;17) aberration was accompanied with loss of genetic material in 7q (53%), −7 (27%), gain of 21q (29%), +8 (17%) and − 18 (16%) and all analyzed patients (n = 13) showed a (likely) pathogenic variant inTP53. The der(5;17) cohort showed significantly shortened telomeres in comparison to the healthy age-matched controls (P < .05), but there was no significant telomere shortening in comparison to MDS/AML patients with a complex karyotype without der(5;17). No fusion genes resulted from the unbalanced translocation. This study demonstrates that the unbalanced translocation der(5;17) is associated with a biallelic inactivation of TP53 due to a deletion of TP53 in one allele and a pathogenic variant of the second TP53 allele. Since the breakpoints are located within (near-) heterochromatic regions, alterations of DNA methylation or histone modifications may be involved in the generation of der(5;17).},
author = {Warnstorf, Daria and Bawadi, Randa and Schienke, Andrea and Strasser, Renate and Schmidt, Gunnar and Illig, Thomas and Tauscher, Marcel and Thol, Felicitas and Heuser, Michael and Steinemann, Doris and Davenport, Claudia and Schlegelberger, Brigitte and Behrens, Yvonne Lisa and Göhring, Gudrun},
journal = {Genes Chromosomes Cancer},
keywords = {l3s leibnizailab},
month = {jan},
number = 6,
pages = {452-457},
title = {Unbalanced translocation der(5;17) resulting in a TP53 loss as recurrent aberration in myelodysplastic syndrome and acute myeloid leukemia with complex karyotype},
volume = 60,
year = 2021
}%0 Journal Article
%1 unbalanced
%A Warnstorf, Daria
%A Bawadi, Randa
%A Schienke, Andrea
%A Strasser, Renate
%A Schmidt, Gunnar
%A Illig, Thomas
%A Tauscher, Marcel
%A Thol, Felicitas
%A Heuser, Michael
%A Steinemann, Doris
%A Davenport, Claudia
%A Schlegelberger, Brigitte
%A Behrens, Yvonne Lisa
%A Göhring, Gudrun
%D 2021
%J Genes Chromosomes Cancer
%N 6
%P 452-457
%R https://doi.org/10.1002/gcc.22938
%T Unbalanced translocation der(5;17) resulting in a TP53 loss as recurrent aberration in myelodysplastic syndrome and acute myeloid leukemia with complex karyotype
%V 60
%X A complex karyotype, detected in myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML), is associated with a reduced median survival. The most frequent chromosomal aberrations in complex karyotypes are deletions of 5q and 17p harboring the tumor suppressor gene TP53. The unbalanced translocation der(5;17) involving chromosome 5q and 17p is a recurrent aberration in MDS/AML, resulting in TP53 loss. We analyzed the karyotypes of 178 patients with an unbalanced translocation der(5;17) using fluorescence R−/G-banding analysis. Whenever possible, fluorescence in situ hybridization (FISH) (n = 138/141), multicolor FISH (n = 8), telomere length measurement (n = 9), targeted DNA sequencing (n = 13), array-CGH (n = 7) and targeted RNA sequencing (n = 2) were conducted. The der(5;17) aberration was accompanied with loss of genetic material in 7q (53%), −7 (27%), gain of 21q (29%), +8 (17%) and − 18 (16%) and all analyzed patients (n = 13) showed a (likely) pathogenic variant inTP53. The der(5;17) cohort showed significantly shortened telomeres in comparison to the healthy age-matched controls (P < .05), but there was no significant telomere shortening in comparison to MDS/AML patients with a complex karyotype without der(5;17). No fusion genes resulted from the unbalanced translocation. This study demonstrates that the unbalanced translocation der(5;17) is associated with a biallelic inactivation of TP53 due to a deletion of TP53 in one allele and a pathogenic variant of the second TP53 allele. Since the breakpoints are located within (near-) heterochromatic regions, alterations of DNA methylation or histone modifications may be involved in the generation of der(5;17). - AbstractURLBibTeXEndNoteBibSonomy1.Booth, A., Reed, A. B., Ponzo, S., Yassaee, A., Aral, M., Plans, D., Labrique, A., and Mohan, D. (2021) Population risk factors for severe disease and mortality in COVID-19: A global systematic review and meta-analysis, PLOS ONE, Public Library of Science 16, 1–30.Aim COVID-19 clinical presentation is heterogeneous, ranging from asymptomatic to severe cases. While there are a number of early publications relating to risk factors for COVID-19 infection, low sample size and heterogeneity in study design impacted consolidation of early findings. There is a pressing need to identify the factors which predispose patients to severe cases of COVID-19. For rapid and widespread risk stratification, these factors should be easily obtainable, inexpensive, and avoid invasive clinical procedures. The aim of our study is to fill this knowledge gap by systematically mapping all the available evidence on the association of various clinical, demographic, and lifestyle variables with the risk of specific adverse outcomes in patients with COVID-19. Methods The systematic review was conducted using standardized methodology, searching two electronic databases (PubMed and SCOPUS) for relevant literature published between 1st January 2020 and 9th July 2020. Included studies reported characteristics of patients with COVID-19 while reporting outcomes relating to disease severity. In the case of sufficient comparable data, meta-analyses were conducted to estimate risk of each variable. Results Seventy-six studies were identified, with a total of 17,860,001 patients across 14 countries. The studies were highly heterogeneous in terms of the sample under study, outcomes, and risk measures reported. A large number of risk factors were presented for COVID-19. Commonly reported variables for adverse outcome from COVID-19 comprised patient characteristics, including age >75 (OR: 2.65, 95% CI: 1.81–3.90), male sex (OR: 2.05, 95% CI: 1.39–3.04) and severe obesity (OR: 2.57, 95% CI: 1.31–5.05). Active cancer (OR: 1.46, 95% CI: 1.04–2.04) was associated with increased risk of severe outcome. A number of common symptoms and vital measures (respiratory rate and SpO2) also suggested elevated risk profiles. Conclusions Based on the findings of this study, a range of easily assessed parameters are valuable to predict elevated risk of severe illness and mortality as a result of COVID-19, including patient characteristics and detailed comorbidities, alongside the novel inclusion of real-time symptoms and vital measurements.
@article{10.1371/journal.pone.0247461,
abstract = {Aim COVID-19 clinical presentation is heterogeneous, ranging from asymptomatic to severe cases. While there are a number of early publications relating to risk factors for COVID-19 infection, low sample size and heterogeneity in study design impacted consolidation of early findings. There is a pressing need to identify the factors which predispose patients to severe cases of COVID-19. For rapid and widespread risk stratification, these factors should be easily obtainable, inexpensive, and avoid invasive clinical procedures. The aim of our study is to fill this knowledge gap by systematically mapping all the available evidence on the association of various clinical, demographic, and lifestyle variables with the risk of specific adverse outcomes in patients with COVID-19. Methods The systematic review was conducted using standardized methodology, searching two electronic databases (PubMed and SCOPUS) for relevant literature published between 1st January 2020 and 9th July 2020. Included studies reported characteristics of patients with COVID-19 while reporting outcomes relating to disease severity. In the case of sufficient comparable data, meta-analyses were conducted to estimate risk of each variable. Results Seventy-six studies were identified, with a total of 17,860,001 patients across 14 countries. The studies were highly heterogeneous in terms of the sample under study, outcomes, and risk measures reported. A large number of risk factors were presented for COVID-19. Commonly reported variables for adverse outcome from COVID-19 comprised patient characteristics, including age >75 (OR: 2.65, 95% CI: 1.81–3.90), male sex (OR: 2.05, 95% CI: 1.39–3.04) and severe obesity (OR: 2.57, 95% CI: 1.31–5.05). Active cancer (OR: 1.46, 95% CI: 1.04–2.04) was associated with increased risk of severe outcome. A number of common symptoms and vital measures (respiratory rate and SpO2) also suggested elevated risk profiles. Conclusions Based on the findings of this study, a range of easily assessed parameters are valuable to predict elevated risk of severe illness and mortality as a result of COVID-19, including patient characteristics and detailed comorbidities, alongside the novel inclusion of real-time symptoms and vital measurements.},
author = {Booth, Adam and Reed, Angus Bruno and Ponzo, Sonia and Yassaee, Arrash and Aral, Mert and Plans, David and Labrique, Alain and Mohan, Diwakar},
journal = {PLOS ONE},
keywords = {l3s leibnizailab},
month = {03},
number = 3,
pages = {1-30},
publisher = {Public Library of Science},
title = {Population risk factors for severe disease and mortality in COVID-19: A global systematic review and meta-analysis},
volume = 16,
year = 2021
}%0 Journal Article
%1 10.1371/journal.pone.0247461
%A Booth, Adam
%A Reed, Angus Bruno
%A Ponzo, Sonia
%A Yassaee, Arrash
%A Aral, Mert
%A Plans, David
%A Labrique, Alain
%A Mohan, Diwakar
%D 2021
%I Public Library of Science
%J PLOS ONE
%N 3
%P 1-30
%R 10.1371/journal.pone.0247461
%T Population risk factors for severe disease and mortality in COVID-19: A global systematic review and meta-analysis
%U https://doi.org/10.1371/journal.pone.0247461
%V 16
%X Aim COVID-19 clinical presentation is heterogeneous, ranging from asymptomatic to severe cases. While there are a number of early publications relating to risk factors for COVID-19 infection, low sample size and heterogeneity in study design impacted consolidation of early findings. There is a pressing need to identify the factors which predispose patients to severe cases of COVID-19. For rapid and widespread risk stratification, these factors should be easily obtainable, inexpensive, and avoid invasive clinical procedures. The aim of our study is to fill this knowledge gap by systematically mapping all the available evidence on the association of various clinical, demographic, and lifestyle variables with the risk of specific adverse outcomes in patients with COVID-19. Methods The systematic review was conducted using standardized methodology, searching two electronic databases (PubMed and SCOPUS) for relevant literature published between 1st January 2020 and 9th July 2020. Included studies reported characteristics of patients with COVID-19 while reporting outcomes relating to disease severity. In the case of sufficient comparable data, meta-analyses were conducted to estimate risk of each variable. Results Seventy-six studies were identified, with a total of 17,860,001 patients across 14 countries. The studies were highly heterogeneous in terms of the sample under study, outcomes, and risk measures reported. A large number of risk factors were presented for COVID-19. Commonly reported variables for adverse outcome from COVID-19 comprised patient characteristics, including age >75 (OR: 2.65, 95% CI: 1.81–3.90), male sex (OR: 2.05, 95% CI: 1.39–3.04) and severe obesity (OR: 2.57, 95% CI: 1.31–5.05). Active cancer (OR: 1.46, 95% CI: 1.04–2.04) was associated with increased risk of severe outcome. A number of common symptoms and vital measures (respiratory rate and SpO2) also suggested elevated risk profiles. Conclusions Based on the findings of this study, a range of easily assessed parameters are valuable to predict elevated risk of severe illness and mortality as a result of COVID-19, including patient characteristics and detailed comorbidities, alongside the novel inclusion of real-time symptoms and vital measurements. - AbstractBibTeXEndNoteBibSonomy1.Kaushal, A., Saha, A., and Ganguly, N. (2021) tWT–WT: A Dataset to Assert the Role of Target Entities for Detecting Stance of Tweets, pp. 3879–3889.The stance detection task aims at detecting the stance of a tweet or a text for a target. These targets can be named entities or free-form sentences (claims). Though the task involves reasoning of the tweet with respect to a target, we find that it is possible to achieve high accuracy on several publicly available Twitter stance detection datasets without looking at the target sentence. Specifically, a simple tweet classification model achieved human-level performance on the WT–WT dataset and more than two-third accuracy on various other datasets. We investigate the existence of biases in such datasets to find the potential spurious correlations of sentiment-stance relations and lexical choice associated with the stance category. Furthermore, we propose a new large dataset free of such biases and demonstrate its aptness on the existing stance detection systems. Our empirical findings show much scope for research on the stance detection task and proposes several considerations for creating future stance detection datasets.
@proceedings{J,
abstract = {The stance detection task aims at detecting the stance of a tweet or a text for a target. These targets can be named entities or free-form sentences (claims). Though the task involves reasoning of the tweet with respect to a target, we find that it is possible to achieve high accuracy on several publicly available Twitter stance detection datasets without looking at the target sentence. Specifically, a simple tweet classification model achieved human-level performance on the WT–WT dataset and more than two-third accuracy on various other datasets. We investigate the existence of biases in such datasets to find the potential spurious correlations of sentiment-stance relations and lexical choice associated with the stance category. Furthermore, we propose a new large dataset free of such biases and demonstrate its aptness on the existing stance detection systems. Our empirical findings show much scope for research on the stance detection task and proposes several considerations for creating future stance detection datasets.},
author = {Kaushal, Ayush and Saha, Avirup and Ganguly, Niloy},
keywords = {l3s leibnizailab},
month = {jun},
pages = {3879-3889},
title = {tWT–WT: A Dataset to Assert the Role of Target Entities for Detecting Stance of Tweets},
year = 2021
}%0 Conference Proceedings
%1 J
%A Kaushal, Ayush
%A Saha, Avirup
%A Ganguly, Niloy
%D 2021
%P 3879-3889
%T tWT–WT: A Dataset to Assert the Role of Target Entities for Detecting Stance of Tweets
%X The stance detection task aims at detecting the stance of a tweet or a text for a target. These targets can be named entities or free-form sentences (claims). Though the task involves reasoning of the tweet with respect to a target, we find that it is possible to achieve high accuracy on several publicly available Twitter stance detection datasets without looking at the target sentence. Specifically, a simple tweet classification model achieved human-level performance on the WT–WT dataset and more than two-third accuracy on various other datasets. We investigate the existence of biases in such datasets to find the potential spurious correlations of sentiment-stance relations and lexical choice associated with the stance category. Furthermore, we propose a new large dataset free of such biases and demonstrate its aptness on the existing stance detection systems. Our empirical findings show much scope for research on the stance detection task and proposes several considerations for creating future stance detection datasets. - AbstractURLBibTeXEndNoteBibSonomy1.Mukherjee, R., Naik, A., Poddar, S., Dasgupta, S., and Ganguly, N. (2021) Understanding the Role of Affect Dimensions in Detecting Emotions from Tweets: A Multi-task Approach. In .We propose VADEC, a multi-task framework that exploits the correlation between the categorical and dimensional models of emotion representation for better subjectivity analysis. Focusing primarily on the effective detection of emotions from tweets, we jointly train multi-label emotion classification and multi-dimensional emotion regression, thereby utilizing the inter-relatedness between the tasks. Co-training especially helps in improving the performance of the classification task as we outperform the strongest baselines with 3.4%, 11%, and 3.9% gains in Jaccard Accuracy, Macro-F1, and Micro-F1 scores respectively on the AIT dataset. We also achieve state-of-the-art results with 11.3% gains averaged over six different metrics on the SenWave dataset. For the regression task, VADEC, when trained with SenWave, achieves 7.6% and 16.5% gains in Pearson Correlation scores over the current state-of-the-art on the EMOBANK dataset for the Valence (V) and Dominance (D) affect dimensions respectively. We conclude our work with a case study on COVID-19 tweets posted by Indians that further helps in establishing the efficacy of our proposed solution.
@inproceedings{mukherjee2021understanding,
abstract = {We propose VADEC, a multi-task framework that exploits the correlation between the categorical and dimensional models of emotion representation for better subjectivity analysis. Focusing primarily on the effective detection of emotions from tweets, we jointly train multi-label emotion classification and multi-dimensional emotion regression, thereby utilizing the inter-relatedness between the tasks. Co-training especially helps in improving the performance of the classification task as we outperform the strongest baselines with 3.4%, 11%, and 3.9% gains in Jaccard Accuracy, Macro-F1, and Micro-F1 scores respectively on the AIT dataset. We also achieve state-of-the-art results with 11.3% gains averaged over six different metrics on the SenWave dataset. For the regression task, VADEC, when trained with SenWave, achieves 7.6% and 16.5% gains in Pearson Correlation scores over the current state-of-the-art on the EMOBANK dataset for the Valence (V) and Dominance (D) affect dimensions respectively. We conclude our work with a case study on COVID-19 tweets posted by Indians that further helps in establishing the efficacy of our proposed solution.},
author = {Mukherjee, Rajdeep and Naik, Atharva and Poddar, Sriyash and Dasgupta, Soham and Ganguly, Niloy},
keywords = {l3s leibnizailab},
note = {cite arxiv:2105.03983Comment: 5 pages, Short Paper accepted at SIGIR 2021},
title = {Understanding the Role of Affect Dimensions in Detecting Emotions from Tweets: A Multi-task Approach},
year = 2021
}%0 Conference Paper
%1 mukherjee2021understanding
%A Mukherjee, Rajdeep
%A Naik, Atharva
%A Poddar, Sriyash
%A Dasgupta, Soham
%A Ganguly, Niloy
%D 2021
%R 10.1145/3404835.3463080
%T Understanding the Role of Affect Dimensions in Detecting Emotions from Tweets: A Multi-task Approach
%U http://arxiv.org/abs/2105.03983
%X We propose VADEC, a multi-task framework that exploits the correlation between the categorical and dimensional models of emotion representation for better subjectivity analysis. Focusing primarily on the effective detection of emotions from tweets, we jointly train multi-label emotion classification and multi-dimensional emotion regression, thereby utilizing the inter-relatedness between the tasks. Co-training especially helps in improving the performance of the classification task as we outperform the strongest baselines with 3.4%, 11%, and 3.9% gains in Jaccard Accuracy, Macro-F1, and Micro-F1 scores respectively on the AIT dataset. We also achieve state-of-the-art results with 11.3% gains averaged over six different metrics on the SenWave dataset. For the regression task, VADEC, when trained with SenWave, achieves 7.6% and 16.5% gains in Pearson Correlation scores over the current state-of-the-art on the EMOBANK dataset for the Valence (V) and Dominance (D) affect dimensions respectively. We conclude our work with a case study on COVID-19 tweets posted by Indians that further helps in establishing the efficacy of our proposed solution. - AbstractURLBibTeXEndNoteBibSonomy1.Xue, Y., Kudenko, D., and Khosla, M. (2021) Graph Learning based Generation of Abstractions for Reinforcement Learning. In Adaptive and Learning Agents Workshop at AAMAS 2021.The application of Reinforcement Learning (RL) Algorithms is often hindered by the combinatorial explosion of the state space. Previous works have leveraged abstractions which condense large state spaces to find tractable solutions, however they assumed that the abstractions are provided by a domain expert. In this work we propose a new approach to automatically construct Abstract Markov Decision Processes (AMDPs) for Potential Based Reward Shaping to improve the sample efficiency of RL algorithms. Our approach to construct abstract states is inspired by graph representation learning methods and effectively encodes topological and reward structure of the ground level MDP. We perform large scale quantitative experiments on Flag Collection domain. We show improvements of up to 6.5 times in sample efficiency and up to 3 times in run time over the baseline approach. Besides, with our qualitative analyses of the generated AMDP we demonstrate the capability of our approach to preserve topological and reward structure of the ground level MDP.
@inproceedings{yuanxue2021graph,
abstract = {The application of Reinforcement Learning (RL) Algorithms is often hindered by the combinatorial explosion of the state space. Previous works have leveraged abstractions which condense large state spaces to find tractable solutions, however they assumed that the abstractions are provided by a domain expert. In this work we propose a new approach to automatically construct Abstract Markov Decision Processes (AMDPs) for Potential Based Reward Shaping to improve the sample efficiency of RL algorithms. Our approach to construct abstract states is inspired by graph representation learning methods and effectively encodes topological and reward structure of the ground level MDP. We perform large scale quantitative experiments on Flag Collection domain. We show improvements of up to 6.5 times in sample efficiency and up to 3 times in run time over the baseline approach. Besides, with our qualitative analyses of the generated AMDP we demonstrate the capability of our approach to preserve topological and reward structure of the ground level MDP.},
author = {Xue, Yuan and Kudenko, Daniel and Khosla, Megha},
booktitle = {Adaptive and Learning Agents Workshop at AAMAS 2021},
keywords = {leibnizailab myown},
title = {Graph Learning based Generation of Abstractions for Reinforcement Learning},
year = 2021
}%0 Conference Paper
%1 yuanxue2021graph
%A Xue, Yuan
%A Kudenko, Daniel
%A Khosla, Megha
%B Adaptive and Learning Agents Workshop at AAMAS 2021
%D 2021
%T Graph Learning based Generation of Abstractions for Reinforcement Learning
%U https://ala2021.vub.ac.be/papers/ALA2021_paper_57.pdf
%X The application of Reinforcement Learning (RL) Algorithms is often hindered by the combinatorial explosion of the state space. Previous works have leveraged abstractions which condense large state spaces to find tractable solutions, however they assumed that the abstractions are provided by a domain expert. In this work we propose a new approach to automatically construct Abstract Markov Decision Processes (AMDPs) for Potential Based Reward Shaping to improve the sample efficiency of RL algorithms. Our approach to construct abstract states is inspired by graph representation learning methods and effectively encodes topological and reward structure of the ground level MDP. We perform large scale quantitative experiments on Flag Collection domain. We show improvements of up to 6.5 times in sample efficiency and up to 3 times in run time over the baseline approach. Besides, with our qualitative analyses of the generated AMDP we demonstrate the capability of our approach to preserve topological and reward structure of the ground level MDP. - AbstractURLBibTeXEndNoteBibSonomy1.Wu, W., Li, B., Luo, C., and Nejdl, W. (2021) Hashing-Accelerated Graph Neural Networks for Link Prediction. In .Networks are ubiquitous in the real world. Link prediction, as one of the key problems for network-structured data, aims to predict whether there exists a link between two nodes. The traditional approaches are based on the explicit similarity computation between the compact node representation by embedding each node into a low-dimensional space. In order to efficiently handle the intensive similarity computation in link prediction, the hashing technique has been successfully used to produce the node representation in the Hamming space. However, the hashing-based link prediction algorithms face accuracy loss from the randomized hashing techniques or inefficiency from the learning to hash techniques in the embedding process. Currently, the Graph Neural Network (GNN) framework has been widely applied to the graph-related tasks in an end-to-end manner, but it commonly requires substantial computational resources and memory costs due to massive parameter learning, which makes the GNN-based algorithms impractical without the help of a powerful workhorse. In this paper, we propose a simple and effective model called #GNN, which balances the trade-off between accuracy and efficiency. #GNN is able to efficiently acquire node representation in the Hamming space for link prediction by exploiting the randomized hashing technique to implement message passing and capture high-order proximity in the GNN framework. Furthermore, we characterize the discriminative power of #GNN in probability. The extensive experimental results demonstrate that the proposed #GNN algorithm achieves accuracy comparable to the learning-based algorithms and outperforms the randomized algorithm, while running significantly faster than the learning-based algorithms. Also, the proposed algorithm shows excellent scalability on a large-scale network with the limited resources.
@conference{wu2021hashingaccelerated,
abstract = {Networks are ubiquitous in the real world. Link prediction, as one of the key problems for network-structured data, aims to predict whether there exists a link between two nodes. The traditional approaches are based on the explicit similarity computation between the compact node representation by embedding each node into a low-dimensional space. In order to efficiently handle the intensive similarity computation in link prediction, the hashing technique has been successfully used to produce the node representation in the Hamming space. However, the hashing-based link prediction algorithms face accuracy loss from the randomized hashing techniques or inefficiency from the learning to hash techniques in the embedding process. Currently, the Graph Neural Network (GNN) framework has been widely applied to the graph-related tasks in an end-to-end manner, but it commonly requires substantial computational resources and memory costs due to massive parameter learning, which makes the GNN-based algorithms impractical without the help of a powerful workhorse. In this paper, we propose a simple and effective model called #GNN, which balances the trade-off between accuracy and efficiency. #GNN is able to efficiently acquire node representation in the Hamming space for link prediction by exploiting the randomized hashing technique to implement message passing and capture high-order proximity in the GNN framework. Furthermore, we characterize the discriminative power of #GNN in probability. The extensive experimental results demonstrate that the proposed #GNN algorithm achieves accuracy comparable to the learning-based algorithms and outperforms the randomized algorithm, while running significantly faster than the learning-based algorithms. Also, the proposed algorithm shows excellent scalability on a large-scale network with the limited resources.},
author = {Wu, Wei and Li, Bin and Luo, Chuan and Nejdl, Wolfgang},
keywords = {l3s leibnizailab machinelearning},
note = {cite arxiv:2105.14280},
title = {Hashing-Accelerated Graph Neural Networks for Link Prediction},
year = 2021
}%0 Generic
%1 wu2021hashingaccelerated
%A Wu, Wei
%A Li, Bin
%A Luo, Chuan
%A Nejdl, Wolfgang
%D 2021
%R 10.1145/3442381.3449884
%T Hashing-Accelerated Graph Neural Networks for Link Prediction
%U http://arxiv.org/abs/2105.14280
%X Networks are ubiquitous in the real world. Link prediction, as one of the key problems for network-structured data, aims to predict whether there exists a link between two nodes. The traditional approaches are based on the explicit similarity computation between the compact node representation by embedding each node into a low-dimensional space. In order to efficiently handle the intensive similarity computation in link prediction, the hashing technique has been successfully used to produce the node representation in the Hamming space. However, the hashing-based link prediction algorithms face accuracy loss from the randomized hashing techniques or inefficiency from the learning to hash techniques in the embedding process. Currently, the Graph Neural Network (GNN) framework has been widely applied to the graph-related tasks in an end-to-end manner, but it commonly requires substantial computational resources and memory costs due to massive parameter learning, which makes the GNN-based algorithms impractical without the help of a powerful workhorse. In this paper, we propose a simple and effective model called #GNN, which balances the trade-off between accuracy and efficiency. #GNN is able to efficiently acquire node representation in the Hamming space for link prediction by exploiting the randomized hashing technique to implement message passing and capture high-order proximity in the GNN framework. Furthermore, we characterize the discriminative power of #GNN in probability. The extensive experimental results demonstrate that the proposed #GNN algorithm achieves accuracy comparable to the learning-based algorithms and outperforms the randomized algorithm, while running significantly faster than the learning-based algorithms. Also, the proposed algorithm shows excellent scalability on a large-scale network with the limited resources. - AbstractURLBibTeXEndNoteBibSonomy1.Wallat, J., Singh, J., and Anand, A. (2020) BERTnesia: Investigating the capture and forgetting of knowledge in BERT.. In Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP, pp. 174–183, Association for Computational Linguistics, Online.Probing complex language models has recently revealed several insights into linguistic and semantic patterns found in the learned representations. In this paper, we probe BERT specifically to understand and measure the relational knowledge it captures. We utilize knowledge base completion tasks to probe every layer of pre-trained as well as fine-tuned BERT (ranking, question answering, NER). Our findings show that knowledge is not just contained in BERT's final layers. Intermediate layers contribute a significant amount (17-60\%) to the total knowledge found. Probing intermediate layers also reveals how different types of knowledge emerge at varying rates. When BERT is fine-tuned, relational knowledge is forgotten but the extent of forgetting is impacted by the fine-tuning objective but not the size of the dataset. We found that ranking models forget the least and retain more knowledge in their final layer.
@inproceedings{wallat-etal-2020-bertnesia,
abstract = {Probing complex language models has recently revealed several insights into linguistic and semantic patterns found in the learned representations. In this paper, we probe BERT specifically to understand and measure the relational knowledge it captures. We utilize knowledge base completion tasks to probe every layer of pre-trained as well as fine-tuned BERT (ranking, question answering, NER). Our findings show that knowledge is not just contained in BERT's final layers. Intermediate layers contribute a significant amount (17-60\%) to the total knowledge found. Probing intermediate layers also reveals how different types of knowledge emerge at varying rates. When BERT is fine-tuned, relational knowledge is forgotten but the extent of forgetting is impacted by the fine-tuning objective but not the size of the dataset. We found that ranking models forget the least and retain more knowledge in their final layer.},
address = {Online},
author = {Wallat, Jonas and Singh, Jaspreet and Anand, Avishek},
booktitle = {Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP},
keywords = {l3s leibnizailab},
month = {nov},
pages = {174--183},
publisher = {Association for Computational Linguistics},
title = {BERTnesia: Investigating the capture and forgetting of knowledge in BERT.},
volume = {abs/2010.09313},
year = 2020
}%0 Conference Paper
%1 wallat-etal-2020-bertnesia
%A Wallat, Jonas
%A Singh, Jaspreet
%A Anand, Avishek
%B Proceedings of the Third BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP
%C Online
%D 2020
%I Association for Computational Linguistics
%P 174--183
%R 10.18653/v1/2020.blackboxnlp-1.17
%T BERTnesia: Investigating the capture and forgetting of knowledge in BERT.
%U http://dblp.uni-trier.de/db/journals/corr/corr2010.html#abs-2010-09313
%V abs/2010.09313
%X Probing complex language models has recently revealed several insights into linguistic and semantic patterns found in the learned representations. In this paper, we probe BERT specifically to understand and measure the relational knowledge it captures. We utilize knowledge base completion tasks to probe every layer of pre-trained as well as fine-tuned BERT (ranking, question answering, NER). Our findings show that knowledge is not just contained in BERT's final layers. Intermediate layers contribute a significant amount (17-60\%) to the total knowledge found. Probing intermediate layers also reveals how different types of knowledge emerge at varying rates. When BERT is fine-tuned, relational knowledge is forgotten but the extent of forgetting is impacted by the fine-tuning objective but not the size of the dataset. We found that ranking models forget the least and retain more knowledge in their final layer. - URLBibTeXEndNoteBibSonomy1.Hu, T., Iosifidis, V., Liao, W., Zhang, H., Yang, M. Y., Ntoutsi, E., and Rosenhahn, B. (2020) FairNN - Conjoint Learning of Fair Representations for Fair Decisions.. In Discovery Science, pp. 581–595, Springer International Publishing.
@incollection{Hu_2020,
author = {Hu, Tongxin and Iosifidis, Vasileios and Liao, Wentong and Zhang, Hang and Yang, Michael Ying and Ntoutsi, Eirini and Rosenhahn, Bodo},
booktitle = {Discovery Science},
keywords = {l3s learning leibnizailab},
pages = {581-595},
publisher = {Springer International Publishing},
series = {Lecture Notes in Computer Science},
title = {FairNN - Conjoint Learning of Fair Representations for Fair Decisions.},
type = {Publication},
volume = 12323,
year = 2020
}%0 Book Section
%1 Hu_2020
%A Hu, Tongxin
%A Iosifidis, Vasileios
%A Liao, Wentong
%A Zhang, Hang
%A Yang, Michael Ying
%A Ntoutsi, Eirini
%A Rosenhahn, Bodo
%B Discovery Science
%D 2020
%I Springer International Publishing
%P 581-595
%R 10.1007/978-3-030-61527-7_38
%T FairNN - Conjoint Learning of Fair Representations for Fair Decisions.
%U https://doi.org/10.1007%2F978-3-030-61527-7_38
%V 12323 - AbstractURLBibTeXEndNoteBibSonomy1.Awiszus, M., Schubert, F., and Rosenhahn, B. (2020, October) TOAD-GAN: Coherent Style Level Generation from a Single Example.In this work, we present TOAD-GAN (Token-based One-shot Arbitrary Dimension Generative Adversarial Network), a novel Procedural Content Generation (PCG) algorithm that generates token-based video game levels. TOAD-GAN follows the SinGAN architecture and can be trained using only one example. We demonstrate its application for Super Mario Bros. levels and are able to generate new levels of similar style in arbitrary sizes. We achieve state-of-the-art results in modeling the patterns of the training level and provide a comparison with different baselines under several metrics. Additionally, we present an extension of the method that allows the user to control the generation process of certain token structures to ensure a coherent global level layout. We provide this tool to the community to spur further research by publishing our source code.
@misc{awiszus2020toadgan,
abstract = {In this work, we present TOAD-GAN (Token-based One-shot Arbitrary Dimension Generative Adversarial Network), a novel Procedural Content Generation (PCG) algorithm that generates token-based video game levels. TOAD-GAN follows the SinGAN architecture and can be trained using only one example. We demonstrate its application for Super Mario Bros. levels and are able to generate new levels of similar style in arbitrary sizes. We achieve state-of-the-art results in modeling the patterns of the training level and provide a comparison with different baselines under several metrics. Additionally, we present an extension of the method that allows the user to control the generation process of certain token structures to ensure a coherent global level layout. We provide this tool to the community to spur further research by publishing our source code.},
author = {Awiszus, Maren and Schubert, Frederik and Rosenhahn, Bodo},
keywords = {l3s leibnizailab},
month = {oct},
note = {cite arxiv:2008.01531Comment: 7 pages, 7 figures. AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE) 2020},
title = {TOAD-GAN: Coherent Style Level Generation from a Single Example},
year = 2020
}%0 Generic
%1 awiszus2020toadgan
%A Awiszus, Maren
%A Schubert, Frederik
%A Rosenhahn, Bodo
%D 2020
%T TOAD-GAN: Coherent Style Level Generation from a Single Example
%U http://arxiv.org/abs/2008.01531
%X In this work, we present TOAD-GAN (Token-based One-shot Arbitrary Dimension Generative Adversarial Network), a novel Procedural Content Generation (PCG) algorithm that generates token-based video game levels. TOAD-GAN follows the SinGAN architecture and can be trained using only one example. We demonstrate its application for Super Mario Bros. levels and are able to generate new levels of similar style in arbitrary sizes. We achieve state-of-the-art results in modeling the patterns of the training level and provide a comparison with different baselines under several metrics. Additionally, we present an extension of the method that allows the user to control the generation process of certain token structures to ensure a coherent global level layout. We provide this tool to the community to spur further research by publishing our source code. - BibTeXEndNoteBibSonomy1.Sen, H., Wentong, L., Rezazadegan Tavakoli, H., Ying Yang, M., Rosenhahn, B., and Pugeault, N. (2020) Image Captioning through Image Transformer. In .
@inproceedings{HeSLia2020,
author = {Sen, He and Wentong, Liao and Rezazadegan Tavakoli, Hamed and Ying Yang, Michael and Rosenhahn, Bodo and Pugeault, Nicolas},
keywords = {l3s leibnizailab},
month = {nov},
title = {Image Captioning through Image Transformer},
year = 2020
}%0 Conference Paper
%1 HeSLia2020
%A Sen, He
%A Wentong, Liao
%A Rezazadegan Tavakoli, Hamed
%A Ying Yang, Michael
%A Rosenhahn, Bodo
%A Pugeault, Nicolas
%D 2020
%T Image Captioning through Image Transformer - AbstractURLBibTeXEndNoteBibSonomy1.Rudolph, M., Wandt, B., and Rosenhahn, B. (2020, August) Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows..The detection of manufacturing errors is crucial in fabrication processes to ensure product quality and safety standards. Since many defects occur very rarely and their characteristics are mostly unknown a priori, their detection is still an open research question. To this end, we propose DifferNet: It leverages the descriptiveness of features extracted by convolutional neural networks to estimate their density using normalizing flows. Normalizing flows are well-suited to deal with low dimensional data distributions. However, they struggle with the high dimensionality of images. Therefore, we employ a multi-scale feature extractor which enables the normalizing flow to assign meaningful likelihoods to the images. Based on these likelihoods we develop a scoring function that indicates defects. Moreover, propagating the score back to the image enables pixel-wise localization. To achieve a high robustness and performance we exploit multiple transformations in training and evaluation. In contrast to most other methods, ours does not require a large number of training samples and performs well with as low as 16 images. We demonstrate the superior performance over existing approaches on the challenging and newly proposed MVTec AD and Magnetic Tile Defects datasets.
@misc{rudolph2020differnet,
abstract = {The detection of manufacturing errors is crucial in fabrication processes to ensure product quality and safety standards. Since many defects occur very rarely and their characteristics are mostly unknown a priori, their detection is still an open research question. To this end, we propose DifferNet: It leverages the descriptiveness of features extracted by convolutional neural networks to estimate their density using normalizing flows. Normalizing flows are well-suited to deal with low dimensional data distributions. However, they struggle with the high dimensionality of images. Therefore, we employ a multi-scale feature extractor which enables the normalizing flow to assign meaningful likelihoods to the images. Based on these likelihoods we develop a scoring function that indicates defects. Moreover, propagating the score back to the image enables pixel-wise localization. To achieve a high robustness and performance we exploit multiple transformations in training and evaluation. In contrast to most other methods, ours does not require a large number of training samples and performs well with as low as 16 images. We demonstrate the superior performance over existing approaches on the challenging and newly proposed MVTec AD and Magnetic Tile Defects datasets.},
author = {Rudolph, Marco and Wandt, Bastian and Rosenhahn, Bodo},
keywords = {l3s leibnizailab},
month = {aug},
note = {cite arxiv:2008.12577},
title = {Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows.},
year = 2020
}%0 Generic
%1 rudolph2020differnet
%A Rudolph, Marco
%A Wandt, Bastian
%A Rosenhahn, Bodo
%D 2020
%T Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows.
%U http://arxiv.org/abs/2008.12577
%X The detection of manufacturing errors is crucial in fabrication processes to ensure product quality and safety standards. Since many defects occur very rarely and their characteristics are mostly unknown a priori, their detection is still an open research question. To this end, we propose DifferNet: It leverages the descriptiveness of features extracted by convolutional neural networks to estimate their density using normalizing flows. Normalizing flows are well-suited to deal with low dimensional data distributions. However, they struggle with the high dimensionality of images. Therefore, we employ a multi-scale feature extractor which enables the normalizing flow to assign meaningful likelihoods to the images. Based on these likelihoods we develop a scoring function that indicates defects. Moreover, propagating the score back to the image enables pixel-wise localization. To achieve a high robustness and performance we exploit multiple transformations in training and evaluation. In contrast to most other methods, ours does not require a large number of training samples and performs well with as low as 16 images. We demonstrate the superior performance over existing approaches on the challenging and newly proposed MVTec AD and Magnetic Tile Defects datasets. - AbstractURLBibTeXEndNoteBibSonomy1.Wulff, A., Mast, M., Hassler, M., Montag, S., Marschollek, M., and Jack, T. (2020) Designing an openEHR-Based Pipeline for Extracting and Standardizing Unstructured Clinical Data Using Natural Language Processing, Methods Inf Med 59, e64-e78.BACKGROUND: Merging disparate and heterogeneous datasets from clinical routine in a standardized and semantically enriched format to enable a multiple use of data also means incorporating unstructured data such as medical free texts. Although the extraction of structured data from texts, known as natural language processing (NLP), has been researched at least for the English language extensively, it is not enough to get a structured output in any format. NLP techniques need to be used together with clinical information standards such as openEHR to be able to reuse and exchange still unstructured data sensibly. OBJECTIVES: The aim of the study is to automatically extract crucial information from medical free texts and to transform this unstructured clinical data into a standardized and structured representation by designing and implementing an exemplary pipeline for the processing of pediatric medical histories. METHODS: We constructed a pipeline that allows reusing medical free texts such as pediatric medical histories in a structured and standardized way by (1) selecting and modeling appropriate openEHR archetypes as standard clinical information models, (2) defining a German dictionary with crucial text markers serving as expert knowledge base for a NLP pipeline, and (3) creating mapping rules between the NLP output and the archetypes. The approach was evaluated in a first pilot study by using 50 manually annotated medical histories from the pediatric intensive care unit of the Hannover Medical School. RESULTS: We successfully reused 24 existing international archetypes to represent the most crucial elements of unstructured pediatric medical histories in a standardized form. The self-developed NLP pipeline was constructed by defining 3.055 text marker entries, 132 text events, 66 regular expressions, and a text corpus consisting of 776 entries for automatic correction of spelling mistakes. A total of 123 mapping rules were implemented to transform the extracted snippets to an openEHR-based representation to be able to store them together with other structured data in an existing openEHR-based data repository. In the first evaluation, the NLP pipeline yielded 97% precision and 94% recall. CONCLUSION: The use of NLP and openEHR archetypes was demonstrated as a viable approach for extracting and representing important information from pediatric medical histories in a structured and semantically enriched format. We designed a promising approach with potential to be generalized, and implemented a prototype that is extensible and reusable for other use cases concerning German medical free texts. In a long term, this will harness unstructured clinical data for further research purposes such as the design of clinical decision support systems. Together with structured data already integrated in openEHR-based representations, we aim at developing an interoperable openEHR-based application that is capable of automatically assessing a patient's risk status based on the patient's medical history at time of admission.
@article{wulff2020designing,
abstract = {BACKGROUND: Merging disparate and heterogeneous datasets from clinical routine in a standardized and semantically enriched format to enable a multiple use of data also means incorporating unstructured data such as medical free texts. Although the extraction of structured data from texts, known as natural language processing (NLP), has been researched at least for the English language extensively, it is not enough to get a structured output in any format. NLP techniques need to be used together with clinical information standards such as openEHR to be able to reuse and exchange still unstructured data sensibly. OBJECTIVES: The aim of the study is to automatically extract crucial information from medical free texts and to transform this unstructured clinical data into a standardized and structured representation by designing and implementing an exemplary pipeline for the processing of pediatric medical histories. METHODS: We constructed a pipeline that allows reusing medical free texts such as pediatric medical histories in a structured and standardized way by (1) selecting and modeling appropriate openEHR archetypes as standard clinical information models, (2) defining a German dictionary with crucial text markers serving as expert knowledge base for a NLP pipeline, and (3) creating mapping rules between the NLP output and the archetypes. The approach was evaluated in a first pilot study by using 50 manually annotated medical histories from the pediatric intensive care unit of the Hannover Medical School. RESULTS: We successfully reused 24 existing international archetypes to represent the most crucial elements of unstructured pediatric medical histories in a standardized form. The self-developed NLP pipeline was constructed by defining 3.055 text marker entries, 132 text events, 66 regular expressions, and a text corpus consisting of 776 entries for automatic correction of spelling mistakes. A total of 123 mapping rules were implemented to transform the extracted snippets to an openEHR-based representation to be able to store them together with other structured data in an existing openEHR-based data repository. In the first evaluation, the NLP pipeline yielded 97% precision and 94% recall. CONCLUSION: The use of NLP and openEHR archetypes was demonstrated as a viable approach for extracting and representing important information from pediatric medical histories in a structured and semantically enriched format. We designed a promising approach with potential to be generalized, and implemented a prototype that is extensible and reusable for other use cases concerning German medical free texts. In a long term, this will harness unstructured clinical data for further research purposes such as the design of clinical decision support systems. Together with structured data already integrated in openEHR-based representations, we aim at developing an interoperable openEHR-based application that is capable of automatically assessing a patient's risk status based on the patient's medical history at time of admission.},
author = {Wulff, A. and Mast, M. and Hassler, M. and Montag, S. and Marschollek, M. and Jack, T.},
journal = {Methods Inf Med},
keywords = {l3s leibnizailab},
month = {dec},
number = {S 02},
pages = {e64-e78},
title = {Designing an openEHR-Based Pipeline for Extracting and Standardizing Unstructured Clinical Data Using Natural Language Processing},
volume = 59,
year = 2020
}%0 Journal Article
%1 wulff2020designing
%A Wulff, A.
%A Mast, M.
%A Hassler, M.
%A Montag, S.
%A Marschollek, M.
%A Jack, T.
%D 2020
%J Methods Inf Med
%N S 02
%P e64-e78
%R doi: 10.1055/s-0040-1716403
%T Designing an openEHR-Based Pipeline for Extracting and Standardizing Unstructured Clinical Data Using Natural Language Processing
%U https://www.ncbi.nlm.nih.gov/pubmed/33058101
%V 59
%X BACKGROUND: Merging disparate and heterogeneous datasets from clinical routine in a standardized and semantically enriched format to enable a multiple use of data also means incorporating unstructured data such as medical free texts. Although the extraction of structured data from texts, known as natural language processing (NLP), has been researched at least for the English language extensively, it is not enough to get a structured output in any format. NLP techniques need to be used together with clinical information standards such as openEHR to be able to reuse and exchange still unstructured data sensibly. OBJECTIVES: The aim of the study is to automatically extract crucial information from medical free texts and to transform this unstructured clinical data into a standardized and structured representation by designing and implementing an exemplary pipeline for the processing of pediatric medical histories. METHODS: We constructed a pipeline that allows reusing medical free texts such as pediatric medical histories in a structured and standardized way by (1) selecting and modeling appropriate openEHR archetypes as standard clinical information models, (2) defining a German dictionary with crucial text markers serving as expert knowledge base for a NLP pipeline, and (3) creating mapping rules between the NLP output and the archetypes. The approach was evaluated in a first pilot study by using 50 manually annotated medical histories from the pediatric intensive care unit of the Hannover Medical School. RESULTS: We successfully reused 24 existing international archetypes to represent the most crucial elements of unstructured pediatric medical histories in a standardized form. The self-developed NLP pipeline was constructed by defining 3.055 text marker entries, 132 text events, 66 regular expressions, and a text corpus consisting of 776 entries for automatic correction of spelling mistakes. A total of 123 mapping rules were implemented to transform the extracted snippets to an openEHR-based representation to be able to store them together with other structured data in an existing openEHR-based data repository. In the first evaluation, the NLP pipeline yielded 97% precision and 94% recall. CONCLUSION: The use of NLP and openEHR archetypes was demonstrated as a viable approach for extracting and representing important information from pediatric medical histories in a structured and semantically enriched format. We designed a promising approach with potential to be generalized, and implemented a prototype that is extensible and reusable for other use cases concerning German medical free texts. In a long term, this will harness unstructured clinical data for further research purposes such as the design of clinical decision support systems. Together with structured data already integrated in openEHR-based representations, we aim at developing an interoperable openEHR-based application that is capable of automatically assessing a patient's risk status based on the patient's medical history at time of admission. - AbstractBibTeXEndNoteBibSonomy1.Scheffner, I., Gietzelt, M., Abeling, T., Marschollek, M., and Gwinner, W. (2020) Patient Survival After Kidney Transplantation: Important Role of Graft-sustaining Factors as Determined by Predictive Modeling Using Random Survival Forest Analysis, Transplantation 104, 1095–1107.Background: Identification of the relevant factors for death can improve patient's individual risk assessment and decision making. A well-documented patient cohort (n = 892) in a renal transplant program with protocol biopsies was used to establish multivariable models for risk assessment at 3 and 12 months posttransplantation by random survival forest analysis. Methods: Patients transplanted between 2000 and 2007 were observed for up to 11 years. Loss to follow-up was negligible (n = 15). A total of 2251 protocol biopsies and 1214 biopsies for cause were performed. All rejections and clinical borderline rejections in protocol biopsies were treated. Results: Ten-year patient survival was 78%, with inferior survival of patients with graft loss. Using all pre- and posttransplant variables until 3 and 12 months (n = 65), the obtained models showed good performance to predict death (concordance index: 0.77-0.78). Validation with a separate cohort of patients (n = 349) showed a concordance index of 0.76 and good discrimination of risks by the models, despite substantial differences in clinical variables. Random survival forest analysis produced robust models over a wide range of parameter settings. Besides well-established risk factors like age, cardiovascular disease, type 2 diabetes, and graft function, posttransplant urinary tract infection and rejection treatment were important factors. Urinary tract infection and rejection treatment were not specifically associated with death due to infection or malignancy but correlated strongly with inferior graft function and graft loss. Conclusions: The established models indicate the important areas that need special attention in the care of renal transplant patients, particularly modifiable factors like graft rejection and urinary tract infection.
@article{noauthororeditor,
abstract = {Background: Identification of the relevant factors for death can improve patient's individual risk assessment and decision making. A well-documented patient cohort (n = 892) in a renal transplant program with protocol biopsies was used to establish multivariable models for risk assessment at 3 and 12 months posttransplantation by random survival forest analysis. Methods: Patients transplanted between 2000 and 2007 were observed for up to 11 years. Loss to follow-up was negligible (n = 15). A total of 2251 protocol biopsies and 1214 biopsies for cause were performed. All rejections and clinical borderline rejections in protocol biopsies were treated. Results: Ten-year patient survival was 78%, with inferior survival of patients with graft loss. Using all pre- and posttransplant variables until 3 and 12 months (n = 65), the obtained models showed good performance to predict death (concordance index: 0.77-0.78). Validation with a separate cohort of patients (n = 349) showed a concordance index of 0.76 and good discrimination of risks by the models, despite substantial differences in clinical variables. Random survival forest analysis produced robust models over a wide range of parameter settings. Besides well-established risk factors like age, cardiovascular disease, type 2 diabetes, and graft function, posttransplant urinary tract infection and rejection treatment were important factors. Urinary tract infection and rejection treatment were not specifically associated with death due to infection or malignancy but correlated strongly with inferior graft function and graft loss. Conclusions: The established models indicate the important areas that need special attention in the care of renal transplant patients, particularly modifiable factors like graft rejection and urinary tract infection.},
author = {Scheffner, Irina and Gietzelt, Matthias and Abeling, Tanja and Marschollek, Michael and Gwinner, Wilfried},
journal = {Transplantation},
keywords = {l3s leibnizailab},
month = {may},
number = 5,
pages = {1095-1107},
title = {Patient Survival After Kidney Transplantation: Important Role of Graft-sustaining Factors as Determined by Predictive Modeling Using Random Survival Forest Analysis},
volume = 104,
year = 2020
}%0 Journal Article
%1 noauthororeditor
%A Scheffner, Irina
%A Gietzelt, Matthias
%A Abeling, Tanja
%A Marschollek, Michael
%A Gwinner, Wilfried
%D 2020
%J Transplantation
%N 5
%P 1095-1107
%R DOI: 10.1097/TP.0000000000002922
%T Patient Survival After Kidney Transplantation: Important Role of Graft-sustaining Factors as Determined by Predictive Modeling Using Random Survival Forest Analysis
%V 104
%X Background: Identification of the relevant factors for death can improve patient's individual risk assessment and decision making. A well-documented patient cohort (n = 892) in a renal transplant program with protocol biopsies was used to establish multivariable models for risk assessment at 3 and 12 months posttransplantation by random survival forest analysis. Methods: Patients transplanted between 2000 and 2007 were observed for up to 11 years. Loss to follow-up was negligible (n = 15). A total of 2251 protocol biopsies and 1214 biopsies for cause were performed. All rejections and clinical borderline rejections in protocol biopsies were treated. Results: Ten-year patient survival was 78%, with inferior survival of patients with graft loss. Using all pre- and posttransplant variables until 3 and 12 months (n = 65), the obtained models showed good performance to predict death (concordance index: 0.77-0.78). Validation with a separate cohort of patients (n = 349) showed a concordance index of 0.76 and good discrimination of risks by the models, despite substantial differences in clinical variables. Random survival forest analysis produced robust models over a wide range of parameter settings. Besides well-established risk factors like age, cardiovascular disease, type 2 diabetes, and graft function, posttransplant urinary tract infection and rejection treatment were important factors. Urinary tract infection and rejection treatment were not specifically associated with death due to infection or malignancy but correlated strongly with inferior graft function and graft loss. Conclusions: The established models indicate the important areas that need special attention in the care of renal transplant patients, particularly modifiable factors like graft rejection and urinary tract infection.