Artificial Intelligence and Deep Learning of Seismic Data, March 14-15
Deep Learning – What’s In It For Me?
Deep learning (DL) is an exploding area of artificial intelligence (AI) research, which is driving remarkable advances in speech and image recognition. DL software learns, in a very real sense, to recognize patterns in digital data, and will almost certainly transform the oil and gas industry’s approach to seismic data interpretation. If you’ve got huge datasets, deep learning can deliver efficiencies and insights that humans may never have thought to seek.
Sharp Reflections are pleased to invite you to a special Deep Learning seminar and workshop led by Dr. Janis Keuper, a senior scientist at our R&D partner, Fraunhofer ITWM. Janis will visit Houston to speak at the 2018 RICE OIL & GAS HPC CONFERENCE, to be held in mid-March. Janis has a PhD in Computer Science, and over 15 years of research and application experience in the fields of Machine Learning, Pattern Recognition and Computer Vision. This is a fantastic opportunity to learn about the AI revolution, and how it’s likely to impact the future of seismic data analysis.
PROGRAM
The short course will be divided into a one-day seminar, followed by an optional “hands-on” workshop with tutorials and eamples.
Seminar participants will gain an overview of DL concepts, and be introduced to the most important algorithms in current use today. Manually-computed data attributes are already being replaced by these learning algorithms, which selectively adapt to changing attributes of each dataset to improve feature detection. You’ll learn what deep learning can (and cannot) do, and understand the basic mechanisms that are driving advances in face recognition, autonomous (self-driving) cars and human speech recognition.
The optional day 2 Workshop is designed to get you started with deep learning tools and methods on seismic data. You’ll be introduced to key open source software (TensorFlow), and work on simple tutorials that illustrate application of unsupervised and supervised neural network methods.
AGENDA
Day 1: Practical Introduction to Deep Learning for Seismic Analysis
Introduction to Deep Learning
- Basic concepts
- Success stories
- Requirements and Tools
- Typical Workflows
Deep Learning in Seismic
- Overview of current applications
- Preparing seismic data for DL
- Pitfalls and hurdles of DL in seismic
- likely DL applications
Inside Deep Learnining
- Convolutional Neural Netwoks (CNN)
- Generative models
- Unsupervised learning
Day 2: Workshop with Hands-On Programming
Introduction to Tensorflow
- Introduction to the Python prgramming framework Jupyter
- Basic concepts of Tensorflow (TF)
- Training Models in TF
- Debugging and Monitoring with Tensorboard
- Short hands-on exercise with CNNs
- Reading seismic data in TF
Hands on examples:
- Deep Learning on well data
- Deep Learning for Image analysis
Deep Learning in Pre-Stack Pro
Big Data and Deep Learning are inseparably linked, as scalable compute solutions are required to execute compute-intensive DL algorithms. Fraunhofer ITWM and Sharp Reflections are actively exploring ways to implement interactive deep learning tools in Pre-Stack Pro. These are expected to improve Pre-Stack Pro’s existing QI workflows, and allow users to develop and execute their own solutions for specific datasets and problems.
FAQs
How can I contact the organizer with any questions?
Please contact Grant MacRae (grant.macrae@sharpreflections.com) in our Houston office. Let us know if you prefer that we send an invoice for the registration fees.
What are my transportation/parking options for getting to and from the event?
There is ample public parking at the Lost Lake Visitor Center in Buffalo Bayou Park (3422 Allen Pkwy).
Deep Learning – What’s In It For Me?
Deep learning (DL) is an exploding area of artificial intelligence (AI) research, which is driving remarkable advances in speech and image recognition. DL software learns, in a very real sense, to recognize patterns in digital data, and will almost certainly transform the oil and gas industry’s approach to seismic data interpretation. If you’ve got huge datasets, deep learning can deliver efficiencies and insights that humans may never have thought to seek.
Sharp Reflections are pleased to invite you to a special Deep Learning seminar and workshop led by Dr. Janis Keuper, a senior scientist at our R&D partner, Fraunhofer ITWM. Janis will visit Houston to speak at the 2018 RICE OIL & GAS HPC CONFERENCE, to be held in mid-March. Janis has a PhD in Computer Science, and over 15 years of research and application experience in the fields of Machine Learning, Pattern Recognition and Computer Vision. This is a fantastic opportunity to learn about the AI revolution, and how it’s likely to impact the future of seismic data analysis.
PROGRAM
The short course will be divided into a one-day seminar, followed by an optional “hands-on” workshop with tutorials and eamples.
Seminar participants will gain an overview of DL concepts, and be introduced to the most important algorithms in current use today. Manually-computed data attributes are already being replaced by these learning algorithms, which selectively adapt to changing attributes of each dataset to improve feature detection. You’ll learn what deep learning can (and cannot) do, and understand the basic mechanisms that are driving advances in face recognition, autonomous (self-driving) cars and human speech recognition.
The optional day 2 Workshop is designed to get you started with deep learning tools and methods on seismic data. You’ll be introduced to key open source software (TensorFlow), and work on simple tutorials that illustrate application of unsupervised and supervised neural network methods.
AGENDA
Day 1: Practical Introduction to Deep Learning for Seismic Analysis
Introduction to Deep Learning
- Basic concepts
- Success stories
- Requirements and Tools
- Typical Workflows
Deep Learning in Seismic
- Overview of current applications
- Preparing seismic data for DL
- Pitfalls and hurdles of DL in seismic
- likely DL applications
Inside Deep Learnining
- Convolutional Neural Netwoks (CNN)
- Generative models
- Unsupervised learning
Day 2: Workshop with Hands-On Programming
Introduction to Tensorflow
- Introduction to the Python prgramming framework Jupyter
- Basic concepts of Tensorflow (TF)
- Training Models in TF
- Debugging and Monitoring with Tensorboard
- Short hands-on exercise with CNNs
- Reading seismic data in TF
Hands on examples:
- Deep Learning on well data
- Deep Learning for Image analysis
Deep Learning in Pre-Stack Pro
Big Data and Deep Learning are inseparably linked, as scalable compute solutions are required to execute compute-intensive DL algorithms. Fraunhofer ITWM and Sharp Reflections are actively exploring ways to implement interactive deep learning tools in Pre-Stack Pro. These are expected to improve Pre-Stack Pro’s existing QI workflows, and allow users to develop and execute their own solutions for specific datasets and problems.
FAQs
How can I contact the organizer with any questions?
Please contact Grant MacRae (grant.macrae@sharpreflections.com) in our Houston office. Let us know if you prefer that we send an invoice for the registration fees.
What are my transportation/parking options for getting to and from the event?
There is ample public parking at the Lost Lake Visitor Center in Buffalo Bayou Park (3422 Allen Pkwy).