• Geophysics and Geophysical Exploration
• /
• v.15 no.3
• /
• pp.129-135
• /
• 2012

The simultaneous-source full waveform inversion improves the applicability of full waveform inversion by reducing the computational cost. Since this technique adopts simultaneous multi-source for forward modeling, unwanted events remain in the residual seismograms when the receiver geometry of field acquisition is different from that of numerical modeling. As a result, these events impede the convergence of the full waveform inversion. In particular, the streamer-type data with limited offsets is the most difficult data to apply the simultaneous-source technique. To overcome this problem, the global-correlation-based objective function was suggested and it was successfully applied to the simultaneous-source full waveform inversion in time domain. However, this method distorts residual wavefields due to the modified objective function and has a negative influence on the inversion result. In addition, this method has not been applied to the frequency-domain simultaneous-source full waveform inversion. In this paper, we apply a timedamping function to the observed and modeled data, which are used to compute global correlation, to minimize the distortion of residual wavefields. Since the damped wavefields optimize the performance of the global correlation, it mitigates the distortion of the residual wavefields and improves the inversion result. Our algorithm incorporates the globalcorrelation-based full waveform inversion into the frequency domain by back-propagating the time-domain residual wavefields in the frequency domain. Through the numerical examples using the streamer-type data, we show that our inversion algorithm better describes the velocity structure than the conventional global correlation approach does.

• Journal of Intelligence and Information Systems
• /
• v.23 no.2
• /
• pp.1-17
• /
• 2017

The deep learning framework is software designed to help develop deep learning models. Some of its important functions include "automatic differentiation" and "utilization of GPU". The list of popular deep learning framework includes Caffe (BVLC) and Theano (University of Montreal). And recently, Microsoft's deep learning framework, Microsoft Cognitive Toolkit, was released as open-source license, following Google's Tensorflow a year earlier. The early deep learning frameworks have been developed mainly for research at universities. Beginning with the inception of Tensorflow, however, it seems that companies such as Microsoft and Facebook have started to join the competition of framework development. Given the trend, Google and other companies are expected to continue investing in the deep learning framework to bring forward the initiative in the artificial intelligence business. From this point of view, we think it is a good time to compare some of deep learning frameworks. So we compare three deep learning frameworks which can be used as a Python library. Those are Google's Tensorflow, Microsoft's CNTK, and Theano which is sort of a predecessor of the preceding two. The most common and important function of deep learning frameworks is the ability to perform automatic differentiation. Basically all the mathematical expressions of deep learning models can be represented as computational graphs, which consist of nodes and edges. Partial derivatives on each edge of a computational graph can then be obtained. With the partial derivatives, we can let software compute differentiation of any node with respect to any variable by utilizing chain rule of Calculus. First of all, the convenience of coding is in the order of CNTK, Tensorflow, and Theano. The criterion is simply based on the lengths of the codes and the learning curve and the ease of coding are not the main concern. According to the criteria, Theano was the most difficult to implement with, and CNTK and Tensorflow were somewhat easier. With Tensorflow, we need to define weight variables and biases explicitly. The reason that CNTK and Tensorflow are easier to implement with is that those frameworks provide us with more abstraction than Theano. We, however, need to mention that low-level coding is not always bad. It gives us flexibility of coding. With the low-level coding such as in Theano, we can implement and test any new deep learning models or any new search methods that we can think of. The assessment of the execution speed of each framework is that there is not meaningful difference. According to the experiment, execution speeds of Theano and Tensorflow are very similar, although the experiment was limited to a CNN model. In the case of CNTK, the experimental environment was not maintained as the same. The code written in CNTK has to be run in PC environment without GPU where codes execute as much as 50 times slower than with GPU. But we concluded that the difference of execution speed was within the range of variation caused by the different hardware setup. In this study, we compared three types of deep learning framework: Theano, Tensorflow, and CNTK. According to Wikipedia, there are 12 available deep learning frameworks. And 15 different attributes differentiate each framework. Some of the important attributes would include interface language (Python, C ++, Java, etc.) and the availability of libraries on various deep learning models such as CNN, RNN, DBN, and etc. And if a user implements a large scale deep learning model, it will also be important to support multiple GPU or multiple servers. Also, if you are learning the deep learning model, it would also be important if there are enough examples and references.

• Geophysics and Geophysical Exploration
• /
• v.8 no.1
• /
• pp.7-17
• /
• 2005

The determination of seismic velocities in refractors for near-surface seismic refraction investigations is an ill-posed problem. Small variations in the computed time parameters can result in quite large lateral variations in the derived velocities, which are often artefacts of the inversion algorithms. Such artefacts are usually not recognized or corrected with forward modelling. Therefore, if detailed refractor models are sought with model based inversion, then detailed starting models are required. The usual source of artefacts in seismic velocities is irregular refractors. Under most circumstances, the variable migration of the generalized reciprocal method (GRM) is able to accommodate irregular interfaces and generate detailed starting models of the refractor. However, where the very-near-surface environment of the Earth is also irregular, the efficacy of the GRM is reduced, and weathering corrections can be necessary. Standard methods for correcting for surface irregularities are usually not practical where the very-near-surface irregularities are of limited lateral extent. In such circumstances, the GRM smoothing statics method (SSM) is a simple and robust approach, which can facilitate more-accurate estimates of refractor velocities. The GRM SSM generates a smoothing 'statics' correction by subtracting an average of the time-depths computed with a range of XY values from the time-depths computed with a zero XY value (where the XY value is the separation between the receivers used to compute the time-depth). The time-depths to the deeper target refractors do not vary greatly with varying XY values, and therefore an average is much the same as the optimum value. However, the time-depths for the very-near-surface irregularities migrate laterally with increasing XY values and they are substantially reduced with the averaging process. As a result, the time-depth profile averaged over a range of XY values is effectively corrected for the near-surface irregularities. In addition, the time-depths computed with a Bero XY value are the sum of both the near-surface effects and the time-depths to the target refractor. Therefore, their subtraction generates an approximate 'statics' correction, which in turn, is subtracted from the traveltimes The GRM SSM is essentially a smoothing procedure, rather than a deterministic weathering correction approach, and it is most effective with near-surface irregularities of quite limited lateral extent. Model and case studies demonstrate that the GRM SSM substantially improves the reliability in determining detailed seismic velocities in irregular refractors.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.