Enhancing fiber-optic DAS microseismic event detection in imbalanced data using embedding space optimization

Min Jun Park, Hassan Almomin, Bob Clapp

International Meeting for Applied Geoscience and Energy (IMAGE)

... networks are trained, followed by the extraction of embeddings to define class centers in the embedding space. The embedding model is then fine-tuned...

Deep learning in salt interpretation from R&D to deployment: Challenges and lessons learned

Pandu Devarakota, Apurva Gala, Zhenggang Li, Engin Alkan, Yihua Cai, John Kimbro, Dean Knott, Jeff Moore, Gislain Madiba

International Meeting for Applied Geoscience and Energy (IMAGE)

... a critical role in velocity model building in both exploration and development fields. It is a time-consuming effort that requires key domain expertise...

Seismic interpolation based on quadratic denoising neural network

Yuhan Sui, Xiaojing Wang, Jianwei Ma

International Meeting for Applied Geoscience and Energy (IMAGE)

.... Recently, a deep learning-based method is developed to achieve seismic interpolation by integrating the well-trained denoising convolutional neural...

Introducing stochasticity into CNN-based property estimation from angle-stack seismic

Haibin Di, Tao Zhao, Aria Abubakar

International Meeting for Applied Geoscience and Energy (IMAGE)

... and perturbing with Gaussian noises ℕ(0,1) per prior rock property model. convolutional layer for reconstructing the fullstack seismic, and (iii) one...

Convolutional Neural Networks for Semantic Segmentation of Micro-Pores in SEM Based Images of Shales

Search and Discovery.com

SaltCrawler: AI solution for accelerating velocity model building

Engin Alkan, Yihua Cai, Pandu Devarakota, Apurva Gala, John Kimbro, Dean Knott, Gislain Madiba, Jeff Moore

International Meeting for Applied Geoscience and Energy (IMAGE)

...SaltCrawler: AI solution for accelerating velocity model building Engin Alkan, Yihua Cai, Pandu Devarakota, Apurva Gala, John Kimbro, Dean Knott...

Using Deep Learning and Distributed Machine Learning Algorithms to Forecast Missing Well Log Data; #42234 (2018)

Chijioke Ejimuda, Emenike Ejimuda

Search and Discovery.com

... model. The model accuracy was very low (about 10%). However, currently we are using auto encoder and convolutional neural network ResNet deep...

Deep learning based automatic marker separation

Atul Laxman Katole, Aria Abubakar, Edo Hoekstra, Srikanth Ryali, Tao Zhao

International Meeting for Applied Geoscience and Energy (IMAGE)

... entirely dispenses with convolutional and recurrence-based approaches, and instead rely on the attention mechanism to model the sequential nature...

Identifying geologic facies through seismic dataset-to-dataset transfer learning using convolutional neural networks

Joseph Stitt, Adam Shugar, Rachael Wang

International Meeting for Applied Geoscience and Energy (IMAGE)

... with high and relevant levels of accuracy using deep convolutional neural networks to create the pretrained model. The high degree of similarity between...

A strategy for acoustic impedance direct inversion in depth domain

Ruiqian Cai, Chengyu Sun, Shizhong Li

International Meeting for Applied Geoscience and Energy (IMAGE)

... of the depth-domain seismic data, the traditional convolutional model cannot be used to calculate the synthetic seismogram in depth domain. Therefore...

High-resolution prestack seismic inversion of reservoir parameters using an arch network

Ting Chen, Yaojun Wang, Yuan Yuan, Gang Yu, Guangmin Hu

International Meeting for Applied Geoscience and Energy (IMAGE)

...) construct a convolutional neural network (CNN) that is trained to perform mapping of relevant seismic data cubes to respective velocity logs. Meanwhile...

Magnetotelluric inversion using supervised learning trained with random smooth geoelectric models

Lian Liu, Bo Yang, Yixian Xu, Dikun Yang

International Meeting for Applied Geoscience and Energy (IMAGE)

... frequencies and observation stations, noise, and the model equivalence regarding its resolution (Backus & Gilbert 1967; Parker 1983). Geophysicists...

Strike-slip fault skeletonization based on deep learning cascade ant tracking method

Zhipeng Gui, Junhua Zhang, Rujun Wang, Yintao Zhang, Chong Sun, Mei Yang

International Meeting for Applied Geoscience and Energy (IMAGE)

.... Then, utilizing advantages of ant tracking, fault skeletonization also called as fault thinning is operated. Model test and real data application show: (1...

Evaluation of AI-enhanced processing for automated passive seismic detection and location

Aaron Booterbaugh, Evgeny Naumov

International Meeting for Applied Geoscience and Energy (IMAGE)

... model and provide access to the rich datasets collected by both public arrays and Nanometrics’s private installations. CONVOLUTIONAL NEURAL NETWORK...

ABSTRACT: Deep forest cover classification of consecutive landsat imageries over Borneo

Azalea Kamellia Abdullah, Mohd Nadzri Md Reba, Nur Efarina Jali, Sikula Magupin, Diana Anthony

Geological Society of Malaysia (GSM)

... learning image classification algorithms such as Convolutional Neural Networks (CNN) attains higher accuracy mapping with low human interruption. Deep...

Abstract: Automated Fault Detection from 3-D Seismic Using Artificial Intelligence „ Practical Application and Examples from the Gulf of Mexico and North Slope Alaska;

Andrew Pomroy, Zachary Wolfe

Search and Discovery.com

... in the realm of seismic attributes given its well established strengths in image pattern analysis and recognition. With this in mind, a Convolutional...

What samples must seismic interpreters label for efficient machine learning?

Ryan Benkert, Mohit Prabhushankar, Ghassan AlRegib

International Meeting for Applied Geoscience and Energy (IMAGE)

... resources AlRegib et al. (2018). At the core of successful machine learning algorithms, stands the mathematical model representation of data points...

Machine-learning based arrival-picking in continuous DAS recordings „ Application to the Utah FORGE EGS project

Nepomuk Boitz, William Tegtow, Serge Shapiro

International Meeting for Applied Geoscience and Energy (IMAGE)

... a Convolutional Neural Network (CNN). In contrast to geophone-based picking methods, we suggest to train CNNs only on data from a specific dataset to best...

Seismic imaging uncertainty using deep learning predicted Greens functions

Han Liu, Anar Yusifov, Muhong Zhou, Linda Hodgson

International Meeting for Applied Geoscience and Energy (IMAGE)

... velocity model. We used U-net architecture (Ronneberger et al., 2015) with 5 layers, and each layer has 2 convolutional layers. Ensemble models were...

Estimating CO2 saturation maps from seismic data using deep convolutional neural networks

Zi Xian Leong, Tieyuan Zhu, Alexander Y. Sun

International Meeting for Applied Geoscience and Energy (IMAGE)

... deep convolutional neural networks interpolated velocity and density conform with the seismic structure. We select a 2D slice (Fig. 1) from the 3D model...

A two-stage deep learning workflow for automated seismic inversion

Haibin Di, Wenyi Hu, Aria Abubakar

International Meeting for Applied Geoscience and Energy (IMAGE)

... of four major steps, (i) large-scale structural model construction, (ii) initial property model estimation via a multi-task convolutional neural...

Coloured Seismic Inversion, a Simple, Fast and Cost Effective Way of Inverting Seismic Data: Examples from Clastic and Carbonate Reservoirs, Indonesia

Keith Maynard, Paulus Allo, Phill Houghton

Indonesian Petroleum Association

..., and although an interpretive low frequency model is not used, the technique provides a robust inversion that honours the impedance trend of available well data...

Deep Convolutional Neural Networks for Seismic Salt-Body Delineation

Search and Discovery.com

3D Seismic Facies Classification on CPU and GPU HPC Clusters

Sergio Botelho, Vishal Das, Davide Vanzo, Pandu Devarakota, Vinay Rao, Santi Adavani

Unconventional Resources Technology Conference (URTEC)

...; second, neural network design becomes increasingly challenging due to the higher number of parameters in the model and its larger training time. We...

Comparison of Machine Learning and Statistical Predictive Models for Production Time Series Forecasting in Tight Oil Reservoirs

Hamid Rahmanifard, Ian Gates, Abdolmohsen Shabib-Asl

Unconventional Resources Technology Conference (URTEC)

... by a pooling layer and a fully connected layer. A 1D CNN with a convolutional layer to extract features from the input sequences, followed by an LSTM model...