004 Datenverarbeitung; Informatik
Refine
Year of publication
Document Type
- Article (29)
- Preprint (15)
- Doctoral Thesis (4)
- Part of a Book (2)
- Contribution to a Periodical (1)
Has Fulltext
- yes (51)
Is part of the Bibliography
- no (51)
Keywords
- computer vision (2)
- deep learning (2)
- intrinsic plasticity (2)
- sparse coding (2)
- 3D image analysis (1)
- Amblyopia (1)
- Autonomous Learning (1)
- Bilderkennung (1)
- Binocular Rivalry (1)
- CBM detector (1)
Institute
- Frankfurt Institute for Advanced Studies (FIAS) (51) (remove)
PolarCAP – A deep learning approach for first motion polarity classification of earthquake waveforms
(2022)
Highlights
• We present PolarCAP, a deep learning model that can classify the polarity of a waveform with a 98% accuracy.
• The first-motion polarity of seismograms is a useful parameter, but its manual determination can be laborious and imprecise.
• We demonstrate that in several cases the model can assign trace polar-ity more accurately than a human analyst.
Abstract
The polarity of first P-wave arrivals plays a significant role in the effective determination of focal mechanisms specially for smaller earthquakes. Manual estimation of polarities is not only time-consuming but also prone to human errors. This warrants a need for an automated algorithm for first motion polarity determination. We present a deep learning model - PolarCAP that uses an autoencoder architecture to identify first-motion polarities of earth-quake waveforms. PolarCAP is trained in a supervised fashion using more than 130,000 labelled traces from the Italian seismic dataset (INSTANCE) and is cross-validated on 22,000 traces to choose the most optimal set of hyperparameters. We obtain an accuracy of 0.98 on a completely unseen test dataset of almost 33,000 traces. Furthermore, we check the model generalizability by testing it on the datasets provided by previous works and show that our model achieves a higher recall on both positive and negative polarities.
Structural rearrangements play a central role in the organization and function of complex biomolecular systems. In principle, Molecular Dynamics (MD) simulations enable us to investigate these thermally activated processes with an atomic level of resolution. In practice, an exponentially large fraction of computational resources must be invested to simulate thermal fluctuations in metastable states. Path sampling methods focus the computational power on sampling the rare transitions between states. One of their outstanding limitations is to efficiently generate paths that visit significantly different regions of the conformational space. To overcome this issue, we introduce a new algorithm for MD simulations that integrates machine learning and quantum computing. First, using functional integral methods, we derive a rigorous low-resolution spatially coarse-grained representation of the system’s dynamics, based on a small set of molecular configurations explored with machine learning. Then, we use a quantum annealer to sample the transition paths of this low-resolution theory. We provide a proof-of-concept application by simulating a benchmark conformational transition with all-atom resolution on the D-Wave quantum computer. By exploiting the unique features of quantum annealing, we generate uncorrelated trajectories at every iteration, thus addressing one of the challenges of path sampling. Once larger quantum machines will be available, the interplay between quantum and classical resources may emerge as a new paradigm of high-performance scientific computing. In this work, we provide a platform to implement this integrated scheme in the field of molecular simulations.
Residual connections have been proposed as an architecture-based inductive bias to mitigate the problem of exploding and vanishing gradients and increased task performance in both feed-forward and recurrent networks (RNNs) when trained with the backpropagation algorithm. Yet, little is known about how residual connections in RNNs influence their dynamics and fading memory properties. Here, we introduce weakly coupled residual recurrent networks (WCRNNs) in which residual connections result in well-defined Lyapunov exponents and allow for studying properties of fading memory. We investigate how the residual connections of WCRNNs influence their performance, network dynamics, and memory properties on a set of benchmark tasks. We show that several distinct forms of residual connections yield effective inductive biases that result in increased network expressivity. In particular, those are residual connections that (i) result in network dynamics at the proximity of the edge of chaos, (ii) allow networks to capitalize on characteristic spectral properties of the data, and (iii) result in heterogeneous memory properties. In addition, we demonstrate how our results can be extended to non-linear residuals and introduce a weakly coupled residual initialization scheme that can be used for Elman RNNs.
Background: The technical development of imaging techniques in life sciences has enabled the three-dimensional recording of living samples at increasing temporal resolutions. Dynamic 3D data sets of developing organisms allow for time-resolved quantitative analyses of morphogenetic changes in three dimensions, but require efficient and automatable analysis pipelines to tackle the resulting Terabytes of image data. Particle image velocimetry (PIV) is a robust and segmentation-free technique that is suitable for quantifying collective cellular migration on data sets with different labeling schemes. This paper presents the implementation of an efficient 3D PIV package using the Julia programming language—quickPIV. Our software is focused on optimizing CPU performance and ensuring the robustness of the PIV analyses on biological data.
Results: QuickPIV is three times faster than the Python implementation hosted in openPIV, both in 2D and 3D. Our software is also faster than the fastest 2D PIV package in openPIV, written in C++. The accuracy evaluation of our software on synthetic data agrees with the expected accuracies described in the literature. Additionally, by applying quickPIV to three data sets of the embryogenesis of Tribolium castaneum, we obtained vector fields that recapitulate the migration movements of gastrulation, both in nuclear and actin-labeled embryos. We show normalized squared error cross-correlation to be especially accurate in detecting translations in non-segmentable biological image data.
Conclusions: The presented software addresses the need for a fast and open-source 3D PIV package in biological research. Currently, quickPIV offers efficient 2D and 3D PIV analyses featuring zero-normalized and normalized squared error cross-correlations, sub-pixel/voxel approximation, and multi-pass. Post-processing options include filtering and averaging of the resulting vector fields, extraction of velocity, divergence and collectiveness maps, simulation of pseudo-trajectories, and unit conversion. In addition, our software includes functions to visualize the 3D vector fields in Paraview.
For medicine to fulfill its promise of personalized treatments based on a better understanding of disease biology, computational and statistical tools must exist to analyze the increasing amount of patient data that becomes available. A particular challenge is that several types of data are being measured to cope with the complexity of the underlying systems, enhance predictive modeling and enrich molecular understanding.
Here we review a number of recent approaches that specialize in the analysis of multimodal data in the context of predictive biomedicine. We focus on methods that combine different OMIC measurements with image or genome variation data. Our overview shows the diversity of methods that address analysis challenges and reveals new avenues for novel developments.
A deep convolutional neural network (CNN) is developed to study symmetry energy (Esym(ρ)) effects by learning the mapping between the symmetry energy and the two-dimensional (transverse momentum and rapidity) distributions of protons and neutrons in heavy-ion collisions. Supervised training is performed with labeled data-set from the ultrarelativistic quantum molecular dynamics (UrQMD) model simulation. It is found that, by using proton spectra on event-by-event basis as input, the accuracy for classifying the soft and stiff Esym(ρ) is about 60% due to large event-by-event fluctuations, while by setting event-summed proton spectra as input, the classification accuracy increases to 98%. The accuracies for 5-label (5 different Esym(ρ)) classification task are about 58% and 72% by using proton and neutron spectra, respectively. For the regression task, the mean absolute errors (MAE) which measure the average magnitude of the absolute differences between the predicted and actual L (the slope parameter of Esym(ρ)) are about 20.4 and 14.8 MeV by using proton and neutron spectra, respectively. Fingerprints of the density-dependent nuclear symmetry energy on the transverse momentum and rapidity distributions of protons and neutrons can be identified by convolutional neural network algorithm.
The state-of-the-art pattern recognition method in machine learning (deep convolution neural network) is used to identify the equation of state (EoS) employed in the relativistic hydrodynamic simulations of heavy ion collisions. High-level correlations of particle spectra in transverse momentum and azimuthal angle learned by the network act as an effective EoS-meter in deciphering the nature of the phase transition in QCD. The EoS-meter is model independent and insensitive to other simulation inputs including the initial conditions and shear viscosity for hydrodynamic simulations. Through this study we demonstrate that there is a traceable encoder of the dynamical information from the phase structure that survives the evolution and exists in the final snapshot of heavy ion collisions and one can exclusively and effectively decode these information from the highly complex final output with machine learning when traditional methods fail. Besides the deep neural network, the performance of traditional machine learning classifiers are also provided.
In this proceeding, we review our recent work using deep convolutional neural network (CNN) to identify the nature of the QCD transition in a hybrid modeling of heavy-ion collisions. Within this hybrid model, a viscous hydrodynamic model is coupled with a hadronic cascade “after-burner”. As a binary classification setup, we employ two different types of equations of state (EoS) of the hot medium in the hydrodynamic evolution. The resulting final-state pion spectra in the transverse momentum and azimuthal angle plane are fed to the neural network as the input data in order to distinguish different EoS. To probe the effects of the fluctuations in the event-by-event spectra, we explore different scenarios for the input data and make a comparison in a systematic way. We observe a clear hierarchy in the predictive power when the network is fed with the event-by-event, cascade-coarse-grained and event-fine-averaged spectra. The carefully-trained neural network can extract high-level features from pion spectra to identify the nature of the QCD transition in a realistic simulation scenario.
Motivation: Partial differential equations (PDEs) is a well-established and powerful tool to simulate multi-cellular biological systems. However, available free tools for validation against data are not established. The PDEparams module provides flexible functionality in Python for parameter estimation in PDE models.
Results: The PDEparams module provides a flexible interface and readily accommodates different parameter analysis tools in PDE models such as computation of likelihood profiles, and parametric boot-strapping, along with direct visualisation of the results. To our knowledge, it is the first open, freely available tool for parameter fitting of PDE models.
Availability and implementation: The PDEparams module is distributed under the MIT license. The source code, usage instructions and step-by-step examples are freely available on GitHub at github.com/systemsmedicine/PDE_params.
A key competence for open-ended learning is the formation of increasingly abstract representations useful for driving complex behavior. Abstract representations ignore specific details and facilitate generalization. Here we consider the learning of abstract representations in a multi-modal setting with two or more input modalities. We treat the problem as a lossy compression problem and show that generic lossy compression of multimodal sensory input naturally extracts abstract representations that tend to strip away modalitiy specific details and preferentially retain information that is shared across the different modalities. Furthermore, we propose an architecture to learn abstract representations by identifying and retaining only the information that is shared across multiple modalities while discarding any modality specific information.