Article
Computer Science, Artificial Intelligence
Erwan Giry Fouquet, Mathieu Fauvel, Clement Mallet
Summary: This paper introduces a robust classification model that is able to utilize both labelled and unlabelled samples to improve processing time and accuracy, especially in the presence of label noise.
PATTERN RECOGNITION LETTERS
(2021)
Article
Computer Science, Information Systems
Zhiwen Xiao, Xin Xu, Huanlai Xing, Shouxi Luo, Penglin Dai, Dawei Zhan
Summary: Time series data contains both local and global patterns, but existing feature networks focus on local features and neglect the relationships among them. Therefore, a novel RTFN method is proposed for feature extraction in time series, consisting of TFN and LSTMaN. Experimental results show that the RTFN-based structures achieve excellent performance on multiple datasets.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Interdisciplinary Applications
Euan T. McGonigle, Haeran Cho
Summary: This study investigates the canonical change point problem of detecting multiple mean shifts and proposes a robust estimator that uses a scale-dependent time-average variance constant to measure the noise level. The research demonstrates the consistency of the proposed estimator under general assumptions allowing heavy-tailedness and discusses its application with two widely adopted data segmentation algorithms.
COMPUTATIONAL STATISTICS & DATA ANALYSIS
(2023)
Article
Economics
Francesco Bravo, Degui Li, Dag Tjostheim
Summary: This article studies parametric robust estimation in nonlinear regression models with regressors generated by a class of non-stationary and null recurrent Markov processes. Consistency and limit distribution results for general robust estimators (including nonlinear least squares, least absolute deviation, and Huber's M-estimators) are derived under regularity conditions. Monte-Carlo simulation studies are conducted to examine numerical performance and verify established asymptotic properties, with empirical applications illustrating the usefulness of the proposed method.
JOURNAL OF ECONOMETRICS
(2021)
Article
Computer Science, Artificial Intelligence
Huanlai Xing, Zhiwen Xiao, Dawei Zhan, Shouxi Luo, Penglin Dai, Ke Li
Summary: This study introduces a powerful semisupervised deep learning model SelfMatch, which combines supervised learning, unsupervised learning, and self-distillation techniques. Experimental results demonstrate that SelfMatch performs exceptionally well on 35 widely used UCR2018 datasets compared to various semisupervised and supervised algorithms.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Sergio Freitas, Eduardo Laber, Pedro Lazera, Marco Molinaro
Summary: We propose TCT, an algorithm for training learners using a limited time and a huge labeled dataset, which consistently outperforms alternative teaching/training methods according to an experimental study.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Hasan A. Bedel, Irmak Sivgin, Onat Dalmaz, Salman U. H. Dar, Tolga Cukur
Summary: BolT is a blood-oxygen-level-dependent Transformer model for analyzing multi-variate fMRI time series. It captures contextual representations across diverse time scales using a cascade of Transformer encoders equipped with a novel fused window attention mechanism. The comprehensive experiments on large-scale public datasets demonstrate the superior performance of BolT against state-of-the-art methods.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Geochemistry & Geophysics
Jing Xu, Mi Jiang, Vagner G. Ferreira, Zhou Wu
Summary: This study introduces a robust sequential adjustment method for improving the accuracy of near-real-time InSAR deformation monitoring and reducing the impact of anomalous errors. Experimental results demonstrate that this method performs well in mitigating errors in SAR data stack and shows potential applications for geohazard monitoring.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Mathematics, Applied
Baoyan Sun, Jun Hu, Yan Gao
Summary: This paper focuses on the robust H-infinity state estimation problem for a class of discrete time-varying uncertain neural networks with uniform quantization and time-delay under variance constraints. A time-varying finite-horizon state estimator is designed to satisfy the error variance boundedness and the H-infinity performance constraint. By using stochastic analysis technique, a new H-infinity SE algorithm without resorting to the augmentation method is proposed for DTVUNNs with uniform quantization.
Article
Engineering, Mechanical
Ignacio Ramirez-Parietti, Javier E. Contreras-Reyes, Byron J. Idrovo-Aguirre
Summary: In this study, a new nonparametric approach was proposed to estimate CSE, with redefined criteria for generalization purposes. The research showed that the synchrony level between CAD/USD and SGD/USD foreign exchange rate time series was higher after the 1999 Asian financial crisis, with a slightly different estimated CSE compared to previous studies.
NONLINEAR DYNAMICS
(2021)
Article
Computer Science, Artificial Intelligence
Hussein El Amouri, Thomas Lampert, Pierre Gancarski, Clement Mallet
Summary: The analysis of time series is increasingly popular due to sensor proliferation. Dynamic Time Warping (DTW) is a widely used similarity measure for time series, but it violates metric axioms. Learning DTW-Preserving Shapelets (LDPS) reintroduces these axioms. This article extends LDPS to constrained DTW-preserving shapelets (CDPS) that consider user knowledge in the form of must-link and cannot-link constraints, allowing for clustering that respects the constraints.
PATTERN RECOGNITION
(2023)
Article
Geochemistry & Geophysics
Corne Kreemer, Geoffrey Blewitt
Summary: The CMC Imaging method proposed in this study is fully automated and utilizes robust statistics, showing superior noise reduction compared to other approaches. By defining the spatial extent of the CMC as local as possible and avoiding subjective assignment of filter stations, the method achieves significant RMS reduction in vertical, east, and north components of GPS position time-series.
JOURNAL OF GEODESY
(2021)
Article
Computer Science, Artificial Intelligence
Yitong Li, Kai Wu, Jing Liu
Summary: This paper proposes a robust time series prediction framework called spARIMA, which reduces noise interference by designing a sequential training scheme in batches based on the degree of noise. spARIMA relies on the differential prediction model in ARIMA and absorbs the advantages of the gradual training scheme in self-paced learning (SPL) to effectively address the instability caused by noise. Furthermore, spARIMA introduces diversity selection to avoid selecting similar samples, using a weighted local complexity-similarity distance expression to represent the diversity of noisy data. Comparative tests with existing ARIMA models on two gradient descent algorithms show that spARIMA not only works well with noisy data, but also performs efficiently with normal data, indicating its generalization ability.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Multidisciplinary Sciences
Hiroshi Okamura, Yutaka Osada, Shota Nishijima, Shinto Eguchi
Summary: Nonlinear phenomena in ecology pose challenges for inference and prediction due to autocorrelation and outliers. Traditional least squares and least absolute deviations methods have limitations, leading to the development of a new robust regression approach that accurately estimates autocorrelation while reducing the influence of outliers. Simulations and real data analysis demonstrate that the new method outperforms existing methods in long-term and short-term prediction of nonlinear estimation problems in spawner-recruitment data.
SCIENTIFIC REPORTS
(2021)
Article
Computer Science, Artificial Intelligence
Xinyue Wang, Liping Jing, Yilin Lyu, Mingzhe Guo, Jiaqi Wang, Huafeng Liu, Jian Yu, Tieyong Zeng
Summary: Discovering hidden patterns from imbalanced data is a critical issue in various real-world applications. To tackle this problem, this paper proposes a deep generative classifier that combines model perturbation and data perturbation. The generative classifier is derived from a deep latent variable model and captures the essential information of the original data through latent codes represented by a probability distribution. It also enforces uncertainty in the model and implements data perturbation by restricting the latent codes to lie on components in a Gaussian Mixture Model. Experimental results on real imbalanced image datasets demonstrate the superiority of the proposed model over popular imbalanced classification baselines.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Multidisciplinary Sciences
Christoph Kuppe, Ricardo O. Ramirez Flores, Zhijian Li, Sikander Hayat, Rebecca T. Levinson, Xian Liao, Monica T. Hannani, Jovan Tanevski, Florian Wuennemann, James S. Nagai, Maurice Halder, David Schumacher, Sylvia Menzel, Gideon Schaefer, Konrad Hoeft, Mingbo Cheng, Susanne Ziegler, Xiaoting Zhang, Fabian Peisker, Nadine Kaesler, Turgay Saritas, Yaoxian Xu, Astrid Kassner, Jan Gummert, Michiel Morshuis, Junedh Amrute, Rogier J. A. Veltrop, Peter Boor, Karin Klingel, Linda W. Van Laake, Aryan Vink, Remco M. Hoogenboezem, Eric M. J. Bindels, Leon Schurgers, Susanne Sattler, Denis Schapiro, Rebekka K. Schneider, Kory Lavine, Hendrik Milting, Ivan G. Costa, Julio Saez-Rodriguez, Rafael Kramann
Summary: In this study, an integrative molecular map of human myocardial infarction was generated using multiple analysis methods. The results elucidated the molecular principles of cardiac tissue organization and provided an important reference for mechanistic and therapeutic studies of cardiac disease.
Article
Biochemistry & Molecular Biology
Agnieszka Bochynska, Alexander T. Stenzel, Roksaneh Sayadi Boroujeni, Chao-Chung Kuo, Mirna Barsoum, Weili Liang, Philip Bussmann, Ivan G. Costa, Juliane Luescher-Firzlaff, Bernhard Luescher
Summary: The expression of genes is regulated by the post-translational modifications of core histones. Loss of Ash2l leads to downregulation of H3K4 methylation and gene expression, inhibiting cell proliferation and cell cycle progression, and inducing senescence.
NUCLEIC ACIDS RESEARCH
(2022)
Article
Multidisciplinary Sciences
Daniel Dimitrov, Denes Tuerei, Martin Garrido-Rodriguez, Paul L. Burmedi, James S. Nagai, Charlotte Boys, Ricardo O. Ramirez Flores, Hyojin Kim, Bence Szalai, Ivan G. Costa, Alberto Valdeolivas, Aurelien Dugourd, Julio Saez-Rodriguez
Summary: Multiple methods and resources for inferring cell-cell communication are compared in this study, and an interface called LIANA is developed to facilitate the use and comparison of these approaches. The impact of choice of resource and method on the predicted intercellular interactions is shown, and the predictions are found to be coherent with other data modalities.
NATURE COMMUNICATIONS
(2022)
Article
Genetics & Heredity
Yaoxian Xu, Christoph Kuppe, Javier Perales-Paton, Sikander Hayat, Jennifer Kranz, Ali T. Abdallah, James Nagai, Zhijian Li, Fabian Peisker, Turgay Saritas, Maurice Halder, Sylvia Menzel, Konrad Hoeft, Annegien Kenter, Hyojin Kim, Claudia R. C. van Roeyen, Michael Lehrke, Julia Moellmann, Thimoteus Speer, Eva M. Buhl, Remco Hoogenboezem, Peter Boor, Jitske Jansen, Cordula Knopp, Ingo Kurth, Bart Smeets, Eric Bindels, Marlies E. J. Reinders, Carla Baan, Joost Gribnau, Ewout J. Hoorn, Joachim Steffens, Tobias B. Huber, Ivan Costa, Jurgen Floege, Rebekka K. Schneider, Julio Saez-Rodriguez, Benjamin S. Freedman, Rafael Kramann
Summary: Adult kidney organoids, known as tubuloids, are derived from a distinct CD24(+) epithelial subpopulation and can be used to model autosomal dominant polycystic kidney disease (ADPKD) and study drug response. The study found similarities in gene expression between gene-edited tubuloids and ADPKD patients' tissue, demonstrating the potential of tubuloids in ADPKD disease modeling. Additionally, the approved drug for ADPKD, tolvaptan, was found to have a significant effect on cyst size in tubuloids but not in pluripotent stem cell-derived models.
Letter
Hematology
Marlena Buetow, Fabio J. Testaquadra, Julian Baumeister, Tiago Maie, Nicolas Chatain, Timo Jaquet, Stefan Tillmann, Martina Crysandt, Ivan G. Costa, Tim H. Bruemmendorf, Mirle Schemionek
Article
Hematology
Sevgi Kostel Bal, Sarah Giuliani, Jana Block, Peter Repiscak, Christoph Hafemeister, Tala Shahin, Nurhan Kasap, Bernhard Ransmayr, Yirun Miao, Cheryl van de Wetering, Alexandra Frohne, Raul Jimenez Heredia, Michael Schuster, Samaneh Zoghi, Vanessa Hertlein, Marini Thian, Aleksandr Bykov, Royala Babayeva, Sevgi Bilgic Eltan, Elif Karakoc-Aydiner, Lisa E. Shaw, Iftekhar Chowdhury, Markku Varjosalo, Rafael J. Arguello, Matthias Farlik, Ahmet Ozen, Edgar Serfling, Loic Dupre, Christoph Bock, Florian Halbritter, J. Thomas Hannich, Irinka Castanon, Michael J. Kraakman, Safa Baris, Kaan Boztug
Summary: In this study, the researchers investigated the role of NFATC1 mutations in human immunity and found evidence of metabolic plasticity in patient T cells. The study also demonstrated that metformin and rosiglitazone can improve the effector functions of patient T cells.
Article
Biochemical Research Methods
Zhijian Li, Chao-Chung Kuo, Fabio Ticconi, Mina Shaigan, Julia Gehrmann, Eduardo Gade Gusmao, Manuel Allhoff, Martin Manolov, Martin Zenke, Ivan G. Costa
Summary: This article introduces the Regulatory Genomics Toolbox (RGT), a computational library for the integrative analysis of regulatory genomics data. RGT provides different functionalities to handle genomic signals and regions, and several tools have been developed for distinct downstream analyses. RGT facilitates the customization of computational methods to analyze specific regulatory genomics problems.
BMC BIOINFORMATICS
(2023)
Article
Biochemistry & Molecular Biology
Marcelo A. S. de Toledo, Xuhuang Fu, Tiago Maie, Eva M. Buhl, Katrin Goetz, Susanne Schmitz, Anne Kaiser, Peter Boor, Till Braunschweig, Nicolas Chatain, Ivan G. Costa, Tim H. Bruemmendorf, Steffen Koschmieder, Martin Zenke
Summary: In this study, human induced pluripotent stem cells were differentiated into mast cells, which exhibited characteristics of systemic mastocytosis disease. These cells can be used for disease modeling and drug screening.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2023)
Article
Multidisciplinary Sciences
Luca Bello, John Wiedenhoeft, Alexander Schliep
Summary: Compression is recognized as an important component in engineering fast machine learning methods for big data. Previous work showed that compression can accelerate algorithms for Hidden Markov Models (HMM) and Gibbs sampling. In this study, we extend the compressive computation approach to the classical frequentist HMM algorithms on continuous-valued observations, providing the first compressive approach for this problem. Empirical results from a large-scale simulation study demonstrate that compressed HMM algorithms outperform the classical algorithms with negligible effects on the computed probabilities and inferred state paths of maximal likelihood.
Article
Biology
Huaming Xu, Zhijian Li, Chao-Chung Kuo, Katrin Goetz, Thomas Look, Marcelo A. S. de Toledo, Kristin Sere, Ivan G. Costa, Martin Zenke
Summary: Transcription factor IRF8 plays a crucial role in the development and function of dendritic cells (DC). This study reveals that IRF8 autoactivation is initiated by enhancers and further controlled by a long non-coding RNA (lncRNA) promoter element in a specific type of DC. This negative feedback loop of IRF8 regulates DC differentiation.
Article
Medical Laboratory Technology
Wouter H. G. Hubens, Tiago Maie, Matthis Schnitker, Ledio Bocova, Deepika Puri, Martina Wessiepe, Jan Kramer, Lothar Rink, Steffen Koschmieder, Ivan G. Costa, Wolfgang Wagner
Summary: Cell-type specific DNA methylation can be used to accurately count different leukocyte subsets in blood. In this study, the authors developed targeted DNAm assays and used digital droplet PCR to measure DNAm levels in venous blood and dried blood samples. The results showed good correlations between epigenetic estimates and conventional cell counting methods, and dried blood samples facilitated self-sampling for easier testing accessibility.
CLINICAL CHEMISTRY
(2023)
Editorial Material
Biochemical Research Methods
Ivan G. Costa
Summary: Recently proposed computational approaches use single-cell multimodal sequencing to explore the casual links between chromatin and transcriptional changes, bridging the knowledge gap in transcriptional regulatory control.
Article
Biotechnology & Applied Microbiology
Tiago Maie, Marco Schmidt, Myriam Erz, Wolfgang Wagner, Ivan G. Costa
Summary: CimpleG is a computational framework for detecting small CpG methylation signatures used for cell-type classification and deconvolution. It performs as well as top performing methods for cell-type classification and is time efficient, utilizing only a single DNA methylation site per cell type. CimpleG provides a comprehensive computational framework for delineating DNAm signatures and cellular deconvolution.
Article
Chemistry, Multidisciplinary
Simon Viet Johansson, Morteza Haghir Chehreghani, Ola Engkvist, Alexander Schliep
Summary: Artificial intelligence (AI) offers new approaches to design compounds in drug discovery, such as suggesting new molecular structures or optimizing existing leads. However, the lack of high-quality data sets limits the effectiveness of AI methods. This study proposes a framework for designing combinatorial libraries using a molecular generative model and optimization algorithms. Simulation experiments show that near-optimal library designs can be achieved even without synthesizing additional building blocks.
Meeting Abstract
Pediatrics
Authors K. Ohl, S. H. Subramanyam, E. Verjans, T. Clarner, S. Boell, I. G. Costa Filho, Z. Li, L. Gan, E. Schmitt, T. Bopp, N. Wagner, S. Schulz, T. Goodarzi, M. Scheld, S. Floess, J. Huehn, B. Lambrecht, R. Beyaert, T. Look, M. Zenke, K. Tenbrock
KLINISCHE PADIATRIE
(2022)