This work is concerned with discovering the governing partial differential equation (PDE) of a physical system. Existing methods have demonstrated the PDE identification from finite observations but failed to maintain satisfying results against noisy data, partly owing to suboptimal estimated derivatives and found PDE coefficients. We address the issues by introducing a noise-aware physics-informed machine learning framework to discover the governing PDE from data following arbitrary distributions. We propose training a couple of neural networks, namely solver and preselector, in a multi-task learning paradigm, which yields important scores of basis candidates that constitute the hidden physical constraint. After they are jointly trained, the solver network estimates potential candidates, e.g. partial derivatives, for the sparse regression to initially unveil the most likely parsimonious PDE, decided according to information criterion. Denoising physics-informed neural networks, based on discrete Fourier transform, is proposed to deliver the optimal PDE coefficients respecting the noise-reduced variables. Extensive experiments on five canonical PDEs affirm that the proposed framework presents a robust and interpretable approach for PDE discovery, leading to a new automatic PDE selection algorithm established on minimization of the information criterion decay rate.
11 December 2014 in Journal on Multimodal User Interfaces
Two of the biggest challenges in building models for detecting emotions from electroencephalography (EEG) devices are the relatively small amount of labeled samples and the strong variability of signal feature distributions between different subjects. In this study, we propose a context-generalized model that tackles the data constraints and subject variability simultaneously using a deep neural network architecture optimized for normally distributed subject-independent feature embeddings. Variational autoencoders (VAEs) at the input level allow the lower feature layers of the model to be trained on both labeled and unlabeled samples, maximizing the use of the limited data resources. Meanwhile, variational regularization encourages the model to learn Gaussian-distributed feature embeddings, resulting in robustness to small dataset imbalances. Subject-adversarial regularization applied to the bi-lateral features further enforces subject-independence on the final feature embedding used for emotion classification. The results from subject-independent performance experiments on the SEED and DEAP EEG-emotion datasets show that our model generalizes better across subjects than other state-of-the-art feature embeddings when paired with deep learning classifiers. Furthermore, qualitative analysis of the embedding space reveals that our proposed subject-invariant bi-lateral variational domain adversarial neural network (BiVDANN) architecture may improve the subject-independent performance by discovering normally distributed features.
Numerical modeling for geothermal reservoir engineering is a crucial process to evaluate the performance of the reservoir and to develop strategies for the future development. The governing equations in the geothermal reservoir models consist of several constitutive parameters, and each parameter is given to a large number of simulation grids. Thus, the combinations of parameters we need to estimate are almost limitless. Although several inverse analysis algorithms have been developed, determining the constitutive parameters in the reservoir model is still a matter of trial-and-error estimation in actual practice, and is largely based on the experience of the analyst. There are several parameters which control the hydrothermal processes in the geothermal reservoir modeling. In this study, as an initial challenge, we focus on permeability, which is one of the most important parameters for the modeling. We propose a machine-learning-based method to estimate permeability distributions using measurable data. A large number of learning data were prepared by a geothermal reservoir simulator capable of calculating pressure and temperature distributions in the natural state with different permeability distributions. Several machine learning algorithms (i.e., linear regression, ridge regression, Lasso regression, support vector regression (SVR), multilayer perceptron (MLP), random forest, gradient boosting, and the k-nearest neighbor algorithm) were applied to learn the relationship between the permeability and the pressure and temperature distributions. By comparing the feature importance and the scores of estimations, random forest using pressure differences as feature variables provided the best estimation (the training score of 0.979 and the test score of 0.789). Since it was learned independently of the grids and locations, this model is expected to be generalized. It was also found that estimation is possible to some extent, even for different heat source conditions. This study is a successful demonstration of the first step in achieving the goal of new data-driven geothermal reservoir engineering, which will be developed and enhanced with the knowledge of information science.
We consider a framework that learns human habitual behaviors from the data obtained by various kinds of sensors installed in an environment, so that the environment can interact with us based on those patterns. In this paper, we achieved extracting human behaviors by infrared sensor network as an initial step for the framework. Infrared sensor network is able to track us without putting an extra burden on us. Moreover it is able to collect long-term data. However, tracking with it has two problems, that is, link miss and incorrect link. In order to mitigate these problems, we propose the tracking method utilizing estimated "time distances" between sensors from movements' records. We have installed infrared sensor network in our laboratory, and validated the proposed tracking method by test courses. Afterwards, we confirmed that human behaviors can be extracted from longterm data.
In order to improve Remaining Useful Life (RUL) prediction accuracy for rolling bearings under defect progressing, the robustness for individual differences and the fluctuation of vibration features are challenging issues. In this research, we propose a novel RUL prediction framework based on a Convolutional Neural Network (CNN) and Hierarchical Bayesian Regression (HBR) for considering the degradation conditions and individual differences of RUL to improve the prediction accuracy. The characteristics of the proposed framework are: (1) In order to reduce the effect of the fluctuation of vibration features, the proposed framework uses an intermediate variable indicating the degradation condition instead of predicting RUL from vibration features. (2) The proposed framework considers not only present but also past degradation conditions in CNN. We conducted the experiment on rolling bearings under defect progression and evaluated the RUL prediction accuracy of the proposed framework. The proposed framework can generate a monotonous RUL prediction curve with a probability distribution and improve the RUL prediction accuracy under defect progression.