Project

Deep Learning

Goal: Different topics related to deep learning and project codes.

Updates
0 new
20
Recommendations
0 new
0
Followers
0 new
9
Reads
1 new
247

Project log

Zikri Bayraktar
added a research item
Sedimentary geometry on borehole images usually summarizes the arrangement of bed boundaries, erosive surfaces, cross bedding, sedimentary dip, and/or deformed beds. The interpretation, very often manual, requires a good level of expertise, is time consuming, can suffer from user bias, and become very challenging when dealing with highly deviated wells. Bedform geometry interpretation from crossbed data is rarely completed from a borehole image. The purpose of this study is to develop an automated method to interpret sedimentary structures, including the bedform geometry, from borehole images. Automation is achieved in this unique interpretation methodology using deep learning. The first task comprised the creation of a training dataset of 2D borehole images. This library of images was then used to train machine learning (ML) models. Testing different architectures of convolutional neural networks (CNN) showed the ResNet architecture to give the best performance for the classification of the different sedimentary structures. The validation accuracy was very high, in the range of 93–96%. To test the developed method, additional logs of synthetic data were created as sequences of different sedimentary structures (i.e., classes) associated with different well deviations, with addition of gaps. The model was able to predict the proper class and highlight the transitions accurately.
Zikri Bayraktar
added an update
I am honored to be selected as a Distinguished Speaker of the Society of Petrophysicists and Well Log Analysts for 2019-2020 for our Machine Learning paper that I presented at the SPWLA 60th Annual Symposium in June.
 
Zikri Bayraktar
added an update
Presented this work at the 60th SPWLA Symposium in Houston, TX
 
Zikri Bayraktar
added an update
Our ML paper “Proxy Models via Neural Networks for Borehole Oil-Based Mud Imager Inversion Workflows" is accepted for presentation at the 2019 IEEE MTT-S International Conference on Numerical Electromagnetics and Multiphysics Modeling and Optimization. #machinelearning #ieee #ML
 
Zikri Bayraktar
added an update
Our ML paper "Quantitative Interpretation of Oil-Base Mud Microresistivity Imager via Artificial Neural Networks", is accepted presentation and publication at the SPWLA 60th Annual Symposium. #machinelearning #spwla
 
Zikri Bayraktar
added an update
As one of the guest editors, it is my great pleasure to announce the 2019 IEEE Antennas and Propagation Letters Special Cluster Issue on “Machine Learning Applications in Electromagnetics, Antennas, and Propagation”. ​Please find the call-for-papers below, and if you have any questions please let me know!
 
Zikri Bayraktar
added an update
Google Colab provides a very nice environment with GPU support to practice deep learning. Here is a sample Keras code of autoencoders using the Fashion MNIST data set running on Google Colab.
 
Zikri Bayraktar
added an update
I recently completed the Deeplearning.AI CNN course. Projects involved coding standard CNNs step by step, coding Residual CNNs (ResNet), car detection with YoloV2, art generation with Neural Style Transfer and Face Recognition. It was quite fun and very informative. I highly recommend it.
 
Zikri Bayraktar
added an update
Project 11 - Vehicle detection (SVM) and tracking (computer vision).
 
Zikri Bayraktar
added an update
Project 10 - Behavioral cloning, i.e. car steering angle prediction via deep convolutional neural networks.
 
Zikri Bayraktar
added an update
Project 9 - German traffic sign classification via deep convolutional neural networks (Tensorflow version).
Project 9 - German traffic sign classification via deep convolutional neural networks (Keras version).
 
Zikri Bayraktar
added an update
Project 5 - Celebrity image generation via generative adversarial networks (GAN) in Tensorflow.
 
Zikri Bayraktar
added an update
Zikri Bayraktar
added an update
Project 3 - Simpsons TV script generation via recurrent neural networks (RNN) in Tensorflow.
 
Zikri Bayraktar
added an update
Project 2 - CiFAR10 image classification via convolutional neural networks (CNN) in Tensorflow.
 
Zikri Bayraktar
added an update
Project 1 - Simple feed-forward neural network in Tensorflow for nonlinear regression problem to predict bike usage over time.
 
Zikri Bayraktar
added a project goal
Different topics related to deep learning and project codes.