PyTorchKR Papers

최신 논문

총 261편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2109.01928공식 코드 · Papers with Code

Modelling Arbitrary Complex Dielectric Properties -- an automated implementation for gprMax

There is a need to accurately simulate materials with complex electromagnetic properties when modelling Ground Penetrating Radar (GPR), as many objects encountered with GPR contain water, e.g. soils, curing concrete, and water-filled pipes. One of widely-used open-source software that simulates electromagnetic wave propagation is gprMax. It uses Yee's algorithm to solve Maxwell's equations with the Finite-Difference Time-Domain (FDTD) method. A significant drawback of the FDTD method is the limited ability to model materials with dispersive properties, currently narrowed to specific set of relaxation mechanisms, namely multi-Debye, Drude and Lorentz media. Consequently, modelling any arbitrary complex material should be done by approximating it as a combination of these functions. This paper describes work carried out as part of the Google Summer of Code (GSoC) programme 2021 to develop a new module within gprMax that can be used to simulate complex dispersive materials using multi-Debye expansions in an automatic manner. The module is capable of modelling Havriliak-Negami, Cole-Cole, Cole-Davidson, Jonscher, Complex-Refractive Index Models, and indeed any arbitrary dispersive material with real and imaginary permittivity specified by the user.

arXiv:2108.11141

Moving average options: Machine Learning and Gauss-Hermite quadrature for a double non-Markovian problem

Evaluating moving average options is a tough computational challenge for the energy and commodity market as the payoff of the option depends on the prices of a certain underlying observed on a moving window so, when a long window is considered, the pricing problem becomes high dimensional. We present an efficient method for pricing Bermudan style moving average options, based on Gaussian Process Regression and Gauss-Hermite quadrature, thus named GPR-GHQ. Specifically, the proposed algorithm proceeds backward in time and, at each time-step, the continuation value is computed only in a few points by using Gauss-Hermite quadrature, and then it is learned through Gaussian Process Regression. We test the proposed approach in the Black-Scholes model, where the GPR-GHQ method is made even more efficient by exploiting the positive homogeneity of the continuation value, which allows one to reduce the problem size. Positive homogeneity is also exploited to develop a binomial Markov chain, which is able to deal efficiently with medium-long windows. Secondly, we test GPR-GHQ in the Clewlow-Strickland model, the reference framework for modeling prices of energy commodities. Finally, we consider a challenging problem which involves double non-Markovian feature, that is the rough-Bergomi model. In this case, the pricing problem is even harder since the whole history of the volatility process impacts the future distribution of the process. The manuscript includes a numerical investigation, which displays that GPR-GHQ is very accurate and it is able to handle options with a very long window, thus overcoming the problem of high dimensionality.

arXiv:2108.06907공식 코드 · Papers with Code

Select Wisely and Explain: Active Learning and Probabilistic Local Post-hoc Explainability

Albeit the tremendous performance improvements in designing complex artificial intelligence (AI) systems in data-intensive domains, the black-box nature of these systems leads to the lack of trustworthiness. Post-hoc interpretability methods explain the prediction of a black-box ML model for a single instance, and such explanations are being leveraged by domain experts to diagnose the underlying biases of these models. Despite their efficacy in providing valuable insights, existing approaches fail to deliver consistent and reliable explanations. In this paper, we propose an active learning-based technique called UnRAvEL (Uncertainty driven Robust Active Learning Based Locally Faithful Explanations), which consists of a novel acquisition function that is locally faithful and uses uncertainty-driven sampling based on the posterior distribution on the probabilistic locality using Gaussian process regression(GPR). We present a theoretical analysis of UnRAvEL by treating it as a local optimizer and analyzing its regret in terms of instantaneous regrets over a global optimizer. We demonstrate the efficacy of the local samples generated by UnRAvEL by incorporating different kernels such as the Matern and linear kernels in GPR. Through a series of experiments, we show that UnRAvEL outperforms the baselines with respect to stability and local fidelity on several real-world models and datasets. We show that UnRAvEL is an efficient surrogate dataset generator by deriving importance scores on this surrogate dataset using sparse linear models. We also showcase the sample efficiency and flexibility of the developed framework on the Imagenet dataset using a pre-trained ResNet model.

arXiv:2108.06618

Adaptive Selection of Informative Path Planning Strategies via Reinforcement Learning

In our previous work, we designed a systematic policy to prioritize sampling locations to lead significant accuracy improvement in spatial interpolation by using the prediction uncertainty of Gaussian Process Regression (GPR) as "attraction force" to deployed robots in path planning. Although the integration with Traveling Salesman Problem (TSP) solvers was also shown to produce relatively short travel distance, we here hypothesise several factors that could decrease the overall prediction precision as well because sub-optimal locations may eventually be included in their paths. To address this issue, in this paper, we first explore "local planning" approaches adopting various spatial ranges within which next sampling locations are prioritized to investigate their effects on the prediction performance as well as incurred travel distance. Also, Reinforcement Learning (RL)-based high-level controllers are trained to adaptively produce blended plans from a particular set of local planners to inherit unique strengths from that selection depending on latest prediction states. Our experiments on use cases of temperature monitoring robots demonstrate that the dynamic mixtures of planners can not only generate sophisticated, informative plans that a single planner could not create alone but also ensure significantly reduced travel distances at no cost of prediction reliability without any assist of additional modules for shortest path calculation.

arXiv:2108.04698

Active Learning for Saddle Point Calculation

The saddle point (SP) calculation is a grand challenge for computationally intensive energy function in computational chemistry area, where the saddle point may represent the transition state (TS). The traditional methods need to evaluate the gradients of the energy function at a very large number of locations. To reduce the number of expensive computations of the true gradients, we propose an active learning framework consisting of a statistical surrogate model, Gaussian process regression (GPR) for the energy function, and a single-walker dynamics method, gentle accent dynamics (GAD), for the saddle-type transition states. SP is detected by the GAD applied to the GPR surrogate for the gradient vector and the Hessian matrix. Our key ingredient for efficiency improvements is an active learning method which sequentially designs the most informative locations and takes evaluations of the original model at these locations to train GPR. We formulate this active learning task as the optimal experimental design problem and propose a very efficient sample-based sub-optimal criterion to construct the optimal locations. We show that the new method significantly decreases the required number of energy or force evaluations of the original model.

arXiv:2107.05707

Computational modelling and data-driven homogenisation of knitted membranes

Knitting is an effective technique for producing complex three-dimensional surfaces owing to the inherent flexibility of interlooped yarns and recent advances in manufacturing providing better control of local stitch patterns. Fully yarn-level modelling of large-scale knitted membranes is not feasible. Therefore, we use a two-scale homogenisation approach and model the membrane as a Kirchhoff-Love shell on the macroscale and as Euler-Bernoulli rods on the microscale. The governing equations for both the shell and the rod are discretised with cubic B-spline basis functions. For homogenisation we consider only the in-plane response of the membrane. The solution of the nonlinear microscale problem requires a significant amount of time due to the large deformations and the enforcement of contact constraints, rendering conventional online computational homogenisation approaches infeasible. To sidestep this problem, we use a pre-trained statistical Gaussian Process Regression (GPR) model to map the macroscale deformations to macroscale stresses. During the offline learning phase, the GPR model is trained by solving the microscale problem for a sufficiently rich set of deformation states obtained by either uniform or Sobol sampling. The trained GPR model encodes the nonlinearities and anisotropies present in the microscale and serves as a material model for the membrane response of the macroscale shell. The bending response can be chosen in dependence of the mesh size to penalise the fine out-of-plane wrinkling of the membrane. After verifying and validating the different components of the proposed approach, we introduce several examples involving membranes subjected to tension and shear to demonstrate its versatility and good performance.

arXiv:2107.02463공식 코드 · Papers with Code

EVARS-GPR: EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data

Time series forecasting is a growing domain with diverse applications. However, changes of the system behavior over time due to internal or external influences are challenging. Therefore, predictions of a previously learned fore-casting model might not be useful anymore. In this paper, we present EVent-triggered Augmented Refitting of Gaussian Process Regression for Seasonal Data (EVARS-GPR), a novel online algorithm that is able to handle sudden shifts in the target variable scale of seasonal data. For this purpose, EVARS-GPR com-bines online change point detection with a refitting of the prediction model using data augmentation for samples prior to a change point. Our experiments on sim-ulated data show that EVARS-GPR is applicable for a wide range of output scale changes. EVARS-GPR has on average a 20.8 % lower RMSE on different real-world datasets compared to methods with a similar computational resource con-sumption. Furthermore, we show that our algorithm leads to a six-fold reduction of the averaged runtime in relation to all comparison partners with a periodical refitting strategy. In summary, we present a computationally efficient online fore-casting algorithm for seasonal time series with changes of the target variable scale and demonstrate its functionality on simulated as well as real-world data. All code is publicly available on GitHub: https://github.com/grimmlab/evars-gpr.

arXiv:2106.10994공식 코드 · Papers with Code

BernNet: Learning Arbitrary Graph Spectral Filters via Bernstein Approximation

Many representative graph neural networks, e.g., GPR-GNN and ChebNet, approximate graph convolutions with graph spectral filters. However, existing work either applies predefined filter weights or learns them without necessary constraints, which may lead to oversimplified or ill-posed filters. To overcome these issues, we propose BernNet, a novel graph neural network with theoretical support that provides a simple but effective scheme for designing and learning arbitrary graph spectral filters. In particular, for any filter over the normalized Laplacian spectrum of a graph, our BernNet estimates it by an order-$K$ Bernstein polynomial approximation and designs its spectral property by setting the coefficients of the Bernstein basis. Moreover, we can learn the coefficients (and the corresponding filter weights) based on observed graphs and their associated signals and thus achieve the BernNet specialized for the data. Our experiments demonstrate that BernNet can learn arbitrary spectral filters, including complicated band-rejection and comb filters, and it achieves superior performance in real-world graph modeling tasks. Code is available at https://github.com/ivam-he/BernNet.

arXiv:2106.01907

Robotic Inspection of Underground Utilities for Construction Survey Using a Ground Penetrating Radar

Ground Penetrating Radar (GPR) is a very useful non-destructive evaluation (NDE) device for locating and mapping underground assets prior to digging and trenching efforts in construction. This paper presents a novel robotic system to automate the GPR data collection process, localize the underground utilities, interpret and reconstruct the underground objects for better visualization allowing regular non-professional users to understand the survey results. This system is composed of three modules: 1) an Omni-directional robotic data collection platform, that carries an RGB-D camera with an Inertial Measurement Unit (IMU) and a GPR antenna to perform automatic GPR data collection, and tag each GPR measurement with visual positioning information at every sampling step; 2) a learning-based migration module to interpret the raw GPR B-scan image into a 2D cross-section model of objects; 3) a 3D reconstruction module, i.e., GPRNet, to generate underground utility model represented as fine 3D point cloud. Comparative studies are performed on synthetic data and field GPR raw data with various incompleteness and noise. Experimental results demonstrate that our proposed method achieves a $30.0\%$ higher GPR imaging accuracy in mean Intersection Over Union (IoU) than the conventional back projection (BP) migration approach and $6.9\%$-$7.2\%$ less loss in Chamfer Distance (CD) than baseline methods regarding point cloud model reconstruction. The GPR-based robotic inspection provides an effective tool for civil engineers to detect and survey underground utilities before construction.

arXiv:2105.14645공식 코드 · Papers with Code

Empirical Models for Multidimensional Regression of Fission Systems

The development of next-generation autonomous control of fission systems, such as nuclear power plants, will require leveraging advancements in machine learning. For fission systems, accurate prediction of nuclear transport is important to quantify the safety margin and optimize performance. The state-of-the-art approach to this problem is costly Monte Carlo (MC) simulations to approximate solutions of the neutron transport equation. Such an approach is feasible for offline calculations e.g., for design or licensing, but is precluded from use as a model-based controller. In this work, we explore the use of Artificial Neural Networks (ANN), Gradient Boosting Regression (GBR), Gaussian Process Regression (GPR) and Support Vector Regression (SVR) to generate empirical models. The empirical model can then be deployed, e.g., in a model predictive controller. Two fission systems are explored: the subcritical MIT Graphite Exponential Pile (MGEP), and the critical MIT Research Reactor (MITR). Findings from this work establish guidelines for developing empirical models for multidimensional regression of neutron transport. An assessment of the accuracy and precision finds that the SVR, followed closely by ANN, performs the best. For both MGEP and MITR, the optimized SVR model exhibited a domain-averaged, test, mean absolute percentage error of 0.17 %. A spatial distribution of performance metrics indicates that physical regions of poor performance coincide with locations of largest neutron flux perturbation -- this outcome is mitigated by ANN and SVR. Even at local maxima, ANN and SVR bias is within experimental uncertainty bounds. A comparison of the performance vs. training dataset size found that SVR is more data-efficient than ANN. Both ANN and SVR achieve a greater than 7 order reduction in evaluation time vs. a MC simulation.

arXiv:2105.11492

Adaptive Local Kernels Formulation of Mutual Information with Application to Active Post-Seismic Building Damage Inference

The abundance of training data is not guaranteed in various supervised learning applications. One of these situations is the post-earthquake regional damage assessment of buildings. Querying the damage label of each building requires a thorough inspection by experts, and thus, is an expensive task. A practical approach is to sample the most informative buildings in a sequential learning scheme. Active learning methods recommend the most informative cases that are able to maximally reduce the generalization error. The information theoretic measure of mutual information (MI) is one of the most effective criteria to evaluate the effectiveness of the samples in a pool-based sample selection scenario. However, the computational complexity of the standard MI algorithm prevents the utilization of this method on large datasets. A local kernels strategy was proposed to reduce the computational costs, but the adaptability of the kernels to the observed labels was not considered in the original formulation of this strategy. In this article, an adaptive local kernels methodology is developed that allows for the conformability of the kernels to the observed output data while enhancing the computational complexity of the standard MI algorithm. The proposed algorithm is developed to work on a Gaussian process regression (GPR) framework, where the kernel hyperparameters are updated after each label query using the maximum likelihood estimation. In the sequential learning procedure, the updated hyperparameters can be used in the MI kernel matrices to improve the sample suggestion performance. The advantages are demonstrated on a simulation of the 2018 Anchorage, AK, earthquake. It is shown that while the proposed algorithm enables GPR to reach acceptable performance with fewer training data, the computational demands remain lower than the standard local kernels strategy.

arXiv:2105.04738

Lightweight Distributed Gaussian Process Regression for Online Machine Learning

In this paper, we study the problem where a group of agents aim to collaboratively learn a common static latent function through streaming data. We propose a lightweight distributed Gaussian process regression (GPR) algorithm that is cognizant of agents' limited capabilities in communication, computation and memory. Each agent independently runs agent-based GPR using local streaming data to predict test points of interest; then the agents collaboratively execute distributed GPR to obtain global predictions over a common sparse set of test points; finally, each agent fuses results from distributed GPR with agent-based GPR to refine its predictions. By quantifying the transient and steady-state performances in predictive variance and error, we show that limited inter-agent communication improves learning performances in the sense of Pareto. Monte Carlo simulation is conducted to evaluate the developed algorithm.

arXiv:2104.14364

Making decisions in the dark basement of the brain: A look back at the GPR model of action selection and the basal ganglia

How does your brain decide what you will do next? Over the past few decades compelling evidence has emerged that the basal ganglia, a collection of nuclei in the fore- and mid-brain of all vertebrates, are vital to action selection. Gurney, Prescott, and Redgrave published an influential computational account of this idea in Biological Cybernetics in 2001. Here we take a look back at this pair of papers, outlining the "GPR" model contained therein, the context of that model's development, and the influence it has had over the past twenty years. Tracing its lineage into models and theories still emerging now, we are encouraged that the GPR model is that rare thing, a computational model of a brain circuit whose advances were directly built on by others.

arXiv:2104.10722

Towards 3D Metric GPR Imaging Based on DNN Noise Removal and Dielectric Estimation

Ground Penetrating Radar (GPR) is one of the most important non-destructive evaluation (NDE) devices to detect subsurface objects (i.e., rebars, utility pipes) and reveal the underground scene. The two biggest challenges in GPR-based inspection are the GPR data collection and subsurface target imaging. To address these challenges, we propose a robotic solution that automates the GPR data collection process with a free motion pattern. It facilitates the 3D metric GPR imaging by tagging the pose information with GPR measurement in real-time. We also introduce a deep neural network (DNN) based GPR data analysis method which includes a noise removal segmentation module to clear the noise in GPR raw data and a DielectricNet to estimate the dielectric value of subsurface media in each GPR B-scan data. We use both the field and synthetic data to verify the proposed method. Experimental results demonstrate that our proposed method can achieve better performance and faster processing speed in GPR data collection and 3D GPR imaging than other methods.

arXiv:2011.10177

Flight Sensor Data and Beamforming based Integrated UAV Tracking with Channel Estimation using Gaussian Process Regression

With explosively increasing demands for unmanned aerial vehicle (UAV) applications, reliable link acquisition for serving UAVs is required. Considering the dynamic characteristics of UAV, it is hugely challenging to persist a reliable link without beam misalignment. In this paper, we propose a flight sensor data and beamforming signal based integrated UAV tracking scheme to deal with this problem. The proposed scheme provides a compatible integrated system considering the practical specification of the flight sensor data and the beamforming pilot signal. The UAV position tracking is comprised of two steps: 1) UAV position prediction by the flight sensor data and 2) position update with the beamforming signal using Gaussian process regression (GPR) method, which is a nonparametric machine learning. The flight sensor data can assist ground station (GS) or UAV nodes in designing the precoding and the receive beamforming matrix with drastically reduced overheads. The beamforming signal can accomplish high beamforming gain to be maintained even when the flight sensor data is absent. Therefore, the proposed scheme can support the moving target continuously by utilizing these two signals. The simulation results are provided to confirm that the proposed scheme outperforms other conventional beam tracking schemes. We also derive 3-dimensional (3D) beamforming gain and spectral efficiency (SE) from the mean absolute error (MAE) of the angular value estimation, which can be used as beamforming performance metrics of the data transmission link in advance.

arXiv:2012.04596

Mapping Leaf Area Index with a Smartphone and Gaussian Processes

Leaf area index (LAI) is a key biophysical parameter used to determine foliage cover and crop growth in environmental studies. Smartphones are nowadays ubiquitous sensor devices with high computational power, moderate cost, and high-quality sensors. A smartphone app, called PocketLAI, was recently presented and tested for acquiring ground LAI estimates. In this letter, we explore the use of state-of-the-art nonlinear Gaussian process regression (GPR) to derive spatially explicit LAI estimates over rice using ground data from PocketLAI and Landsat 8 imagery. GPR has gained popularity in recent years because of their solid Bayesian foundations that offers not only high accuracy but also confidence intervals for the retrievals. We show the first LAI maps obtained with ground data from a smartphone combined with advanced machine learning. This work compares LAI predictions and confidence intervals of the retrievals obtained with PocketLAI to those obtained with classical instruments, such as digital hemispheric photography (DHP) and LI-COR LAI-2000. This letter shows that all three instruments got comparable result but the PocketLAI is far cheaper. The proposed methodology hence opens a wide range of possible applications at moderate cost.

arXiv:2011.14918

Compact Dual-Polarized Vivaldi Antenna for Ground Penetrating Radar (GPR) Application

In this paper, a compact dual-polarized Vivaldi antenna is presented. Four Vivaldi elements are used as radiators, and are positioned obliquely and connected in a horn shape. By exciting two sets of elements, two orthogonally polarized radiations can be achieved. The dual-polarized antenna features a low operating frequency band with a wide bandwidth, high port isolation, good directional radiation performance, and a very compact size, making it highly suitable for the ground penetrating radar (GPR) application.

arXiv:2104.06575공식 코드 · Papers with Code

Five Degree-of-Freedom Property Interpolation of Arbitrary Grain Boundaries via Voronoi Fundamental Zone Octonion Framework

We introduce the Voronoi fundamental zone octonion interpolation framework for grain boundary (GB) structure-property models and surrogates. The VFZO framework offers an advantage over other five degree-of-freedom based property interpolation methods because it is constructed as a point set in a manifold. This means that directly computed Euclidean distances approximate the original octonion distance with significantly reduced computation runtime (~7 CPU minutes vs. 153 CPU days for a 50000x50000 pairwise-distance matrix). This increased efficiency facilitates lower interpolation error through the use of significantly more input data. We demonstrate grain boundary energy interpolation results for a non-smooth validation function and simulated bi-crystal datasets for Fe and Ni using four interpolation methods: barycentric interpolation, Gaussian process regression (GPR), inverse-distance weighting, and nearest-neighbor interpolation. These are evaluated for 50000 random input GBs and 10 000 random prediction GBs. The best performance was achieved with GPR, which resulted in a reduction of the root mean square error (RMSE) by 83.0% relative to RMSE of a constant, average model. Likewise, interpolation on a large, noisy, molecular statics Fe simulation dataset improves performance by 34.4% compared to 21.2% in prior work. Interpolation on a small, low-noise MS Ni simulation dataset is similar to interpolation results for the original octonion metric (57.6% vs. 56.4%). A vectorized, parallelized, MATLAB interpolation function (interp5DOF.m) and related routines are available in our VFZO repository (github.com/sgbaird-5dof/interp) which can be applied to other crystallographic point groups. The VFZO framework offers advantages for computing distances between GBs, estimating property values for arbitrary GBs, and modeling surrogates of computationally expensive 5DOF functions and simulations.

arXiv:2103.11352

Detecting Label Noise via Leave-One-Out Cross-Validation

We present a simple algorithm for identifying and correcting real-valued noisy labels from a mixture of clean and corrupted sample points using Gaussian process regression. A heteroscedastic noise model is employed, in which additive Gaussian noise terms with independent variances are associated with each and all of the observed labels. Optimizing the noise model using maximum likelihood estimation leads to the containment of the GPR model's predictive error by the posterior standard deviation in leave-one-out cross-validation. A multiplicative update scheme is proposed for solving the maximum likelihood estimation problem under non-negative constraints. While we provide proof of convergence for certain special cases, the multiplicative scheme has empirically demonstrated monotonic convergence behavior in virtually all our numerical experiments. We show that the presented method can pinpoint corrupted sample points and lead to better regression models when trained on synthetic and real-world scientific data sets.

arXiv:2103.08349

Data-driven method for real-time prediction and uncertainty quantification of fatigue failure under stochastic loading using artificial neural networks and Gaussian process regression

Various engineering systems such as naval and aerial vehicles, offshore structures, and mechanical components of motorized systems, are exposed to fatigue failures due to stochastic loadings. Methods for early failure prediction are essential for engineering, military, and civil applications. In addition to the prediction of time to failure (TtF), uncertainty quantification (UQ) is of major importance for real-time decision-making purposes. Usually, time domain or frequency domain methods are used for fatigue prediction, such as rainflow counting and Miner's rule or Dirlik's method. However, those methods suffer from over-simplistic modeling and inaccurate failure predictions under stochastic loadings. During the last years, several data-driven models were suggested for offline fatigue failure. However, most of them are not capable of both accurate real-time fatigue prediction and UQ. In the current work, a probabilistic data-driven model is introduced. A hybrid architecture of a fully-connected artificial neural network (FC-ANN) and Gaussian process regression (GPR) is proposed to ensure enhanced predictive abilities and simultaneous UQ of the predicted TtF. The real-time prediction and UQ performances of the suggested model are validated using both synthetic and experimental data. This novel hybrid method is fully data-driven and extends the forecasting capabilities of existing time-domain and machine learning-based methods for fatigue prediction. It paves the way towards the development of a preventive system that provides real-time safety and operational instructions and insights for structural health monitoring (SHM) purposes, allowing prevention of environmental damage, and loss of human lives.