Application of Geocognitive Technologies to Basin & Petroleum System Analyses
Day 1 Mon, November 11, 2019（2019）
Eni
Abstract
Abstract Objectives/Scope When dealing with new exploration areas, basin geologists face the challenge of collecting relevant information from all available sources. This include a number of structured commercial databases, but also large corpora of technical documents in which an invaluable amount of information is scattered across. Even if assisted by search tools to filter the documents of interest, extracting information requires a human effort in reading and understanding the documents. Methods, Procedures, Process Eni and IBM developed a cognitive engine exploiting a deep learning approach to scan documents searching for basin geology concepts, extracting information about petroleum system elements (e.g. formation name, geological age and lithology of source rocks, reservoirs and seals) and enabling basin geologists to perform automated queries to collect all the information related to a basin of interest. The collected information is fully referenced to the original paragraphs, tables or pictures of the document in which it was discovered, therefore enabling to validate the robustness of the results. Results, Observations, Conclusions The cognitive engine has been integrated within an application which enables to build a graphical representation of the Petroleum System Event Charts of the basin, integrating the information extracted from commercial databases, the results from the cognitive engine and the manual input from the geologist. The quality of the results from the cognitive engine has been evaluated using a commercial database which provides both tabular data about basins and detailed pdf reports. The cognitive engine has been trained on the pdf reports alone, and the results have been compared with the tabular content of the database, representing the ground truth. The cognitive engine succeeded in identifying the right formations, lithologies and geological ages of the petroleum systems with an accuracy in the range 75% – 90%. Novel/Additive Information The cognitive engine is built with highly innovative technologies, combining the data driven capabilities of deep neural networks with more traditional natural language processing methods based on ontologies. Documents are processed with a three-step approach. In the first step, convolutional neural networks (CNN) are used to recognize the structural elements within a technical paper (e.g. title, authors, paragraphs, figures, tables, references) and to convert a complex pdf structure into a clean sequence of text, which can be analyzed. In the second step, concepts are extracted from these processed documents using extractors, NLP annotators (based on recurrent neural networks) and aggregators. Finally, the joint use of the results from the deep learning tools and the provided ontologies are used to build a knowledge graph, which links together all the discovered entities and their relationships. A fit-for-purpose high efficient graph database has been developed so that the graph can be traversed with full flexibility, collecting all the concepts needed for basin geology studies.
MoreTranslated text
求助PDF
上传PDF
View via Publisher
AI Read Science
AI Summary
AI Summary is the key point extracted automatically understanding the full text of the paper, including the background, methods, results, conclusions, icons and other key content, so that you can get the outline of the paper at a glance.
Example
Background
Key content
Introduction
Methods
Results
Related work
Fund
Key content
- Pretraining has recently greatly promoted the development of natural language processing (NLP)
- We show that M6 outperforms the baselines in multimodal downstream tasks, and the large M6 with 10 parameters can reach a better performance
- We propose a method called M6 that is able to process information of multiple modalities and perform both single-modal and cross-modal understanding and generation
- The model is scaled to large model with 10 billion parameters with sophisticated deployment, and the 10 -parameter M6-large is the largest pretrained model in Chinese
- Experimental results show that our proposed M6 outperforms the baseline in a number of downstream tasks concerning both single modality and multiple modalities We will continue the pretraining of extremely large models by increasing data to explore the limit of its performance
Upload PDF to Generate Summary
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Related Papers
Data Disclaimer
The page data are from open Internet sources, cooperative publishers and automatic analysis results through AI technology. We do not make any commitments and guarantees for the validity, accuracy, correctness, reliability, completeness and timeliness of the page data. If you have any questions, please contact us by email: report@aminer.cn
Chat Paper