IP Library › Granted Patent US 12,626,103
Granted Patent B2
US 12,626,103 · App. 18/546,802 · Granted May 12, 2026

Geologic learning framework

Inventors: Jagrit Sheoran (Pune, IN); Preetika Shedde (London, GB); Sunil Manikani (Pune, IN)
Assignee: SCHLUMBERGER TECHNOLOGY CORPORATION
G06N3/0455G06N3/088G01V20/00
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,626,103
App. No.
18/546,802
Filed
Aug 17, 2023
Granted
May 12, 2026
Kind
B2
Art Unit
2495
USPC
706/25
Abstract

A method can include receiving data files, where the data files include different types of content; training an encoder using the data files to generate a trained encoder; compressing each of the data files using the trained encoder to generate a compressed representation of each of the data files; and processing the compressed representations of the data files to generate groups, where each of the groups represents one of the different types of content, where each of the groups includes members, and where each of the members is associated with a corresponding one of the data files.

Claims (28)

1 . A method comprising:

receiving data files, wherein the data files comprise different types of content collected from a workspace framework tailored to a geologic environment and processed using a client layer, an application layer, a source or source of site information including offset well information, a storage layer, and an artificial intelligence layer, wherein the different types of content comprise text and images, wherein the images comprise at least one downhole log image, and/or wherein the images comprise at least one micrograph of a geologic sample;

training an encoder using the data files to generate a trained encoder;

compressing each of the data files using the trained encoder to generate a compressed representation of each of the data files; and

processing the compressed representations of the data files to generate groups, wherein each of the groups represents one of the different types of content, wherein each of the groups comprises members, and wherein each of the members is associated with a corresponding one of the data files.

2 . The method of claim 1 , wherein the data files comprise digitized document files.

3 . The method of claim 1 , wherein at least a portion of the data files comprise corresponding metadata.

4 . The method of claim 1 , comprising rendering a visual representation of the groups to a graphical user interface on a display.

5 . The method of claim 4 , comprising receiving a search command that instructs a computing system to perform a search on the members of the groups.

6 . The method of claim 1 , comprising storing a data structure of the groups to a storage medium, wherein the data structure associates each of the members of the groups with a corresponding one of the data files.

7 . The method of claim 1 , wherein the processing comprises performing an orthogonal linear transformation of the compressed representations of the data files to generate a transformed representation of the compressed representations of the data files.

8 . The method of claim 7 , wherein the performing the orthogonal linear transformation comprises performing a principal component analysis (PCA).

9 . The method of claim 7 , wherein the processing further comprises performing a nonlinear dimensionality reduction process on the transformed representation of the compressed representations of the data files to generate a dimensionality reduced representation of the transformed representation.

10 . The method of claim 9 , wherein the performing the nonlinear dimensionality reduction process comprises performing a t-distributed stochastic neighbor embedding (t-SNE) process.

11 . The method of claim 9 , wherein the processing further comprises performing a clustering process on the dimensionality reduced representation to generate clusters, optionally wherein the performing a clustering process comprises performing a k-means clustering process, and optionally comprising automatically determining a value for a k parameter of the k-means clustering process.

12 . A system comprising:

one or more processors;

memory accessible to at least one of the one or more processors;

processor-executable instructions stored in the memory and executable to instruct the system to:

receive data files, wherein the data files comprise different types of content collected from a workspace framework tailored to a geologic environment and processed using a client layer, an application layer, a source or source of site information including offset well information, a storage layer, and an artificial intelligence layer, wherein the different types of content comprise text and images, wherein the images comprise at least one downhole log image, and/or wherein the images comprise at least one micrograph of a geologic sample;

train an encoder using the data files to generate a trained encoder;

compress each of the data files using the trained encoder to generate a compressed representation of each of the data files; and

process the compressed representations of the data files to generate groups, wherein each of the groups represents one of the different types of content, wherein each of the groups comprises members, and wherein each of the members is associated with a corresponding one of the data files.

13 . A non-transitory computer readable medium storing computer-executable instructions to instruct a computing system:

receive data files, wherein the data files comprise different types of content collected from a workspace framework tailored to a geologic environment and processed using a client layer, an application layer, a source or source of site information including offset well information, a storage layer, and an artificial intelligence layer, wherein the different types of content comprise text and images, wherein the images comprise at least one downhole log image, and/or wherein the images comprise at least one micrograph of a geologic sample;

train an encoder using the data files to generate a trained encoder;

compress each of the data files using the trained encoder to generate a compressed representation of each of the data files; and

process the compressed representations of the data files to generate groups, wherein each of the groups represents one of the different types of content, wherein each of the groups comprises members, and wherein each of the members is associated with a corresponding one of the data files.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 22, 2023
From: SHEORAN, JAGRIT; SHEDDE, PREETIKA; MANIKANI, SUNIL
To: SCHLUMBERGER TECHNOLOGY CORPORATION
Reel/Frame 064656/0969 →
Priority Claims (1)
IN 202121006902 · Feb 18, 2021 · national
Continuity (1)
Related Publication 20240127039A1 · Apr 18, 2024
References Cited (16)
US 10158859B2 · Nowozin · 2018 [cited by examiner]
US 11042809B1 · Breckenridge · 2021 [cited by examiner]
US 20170270407A1 · Alberti · 2017 [cited by examiner]
US 20190087939A1 · Hakimuddin · 2019 [cited by examiner]
US 20190169962A1 · Aqrawi · 2019 [cited by examiner]
US 20200210768A1 · Turkelson · 2020 [cited by examiner]
Bangoria, B. et al., “A survey on Efficient Enhanced K-Means Clustering Algorithm”, International Journal for Scientific Research Development, 2013, 1(9), 3 pages. [cited by applicant]
Khayatkhoei, M. et al., “Joint Image-Text Clustering using Deep Neural Networks”, downloaded from the internet on Dec. 6, 2024 at [https://api.semanticscholar.org/Corpus], 2017, 6 pages. [cited by applicant]
Lee, et al., “Spatial Assessment of Urban Flood Susceptibility Using Data Mining and Geographic Information System (GIS) Tools”, Sustainability, 2018, 10(3), 648, 19 pages. [cited by applicant]
Mahmoodi et al., “Supervised classification of down-hole physical properties measurements using neural network to predict the lithology”, Journal of Applied Geophysics, 2016, 24, pp. 17-26. [cited by applicant]
Sang, X. et al., “Intelligent High-Resolution Geological Mapping Based on SLIC-CNN”, Journal of Geo-Information, 2020, 9(99), 23 pages. [cited by applicant]
Worden, R. H. et al., “Lower Cretaceous Rodby and Palaeocene Lista Shales: Characterisation and Comparison of Top-Seal Mudstones at Two Planned CCS Sites, Offshore UK”, Journal of Minerals, 2020, 10(8), 691, 3 pages. [cited by applicant]
Yang, S. T. et al., “JECL: Joint Embedding and Cluster Learning for Image-Text Pairs”, 25th Internationl Conference on Pattern Recognition, 2021, Milan Italy, 8 pages. [cited by applicant]
“Zhou et C. al., “A Stratigraphic Prediction Method Based on Machine Learning”, Journal of Applied Sciences, 2019, 9(17), 3553, 29 pages.” [cited by applicant]
Search Report and Written Opinion of International Patent Application No. PCT/US2022/070674 dated on May 25, 2022, 9 Pages. [cited by applicant]
Extended Search Report issued in European Patent Application No. 22757160.1 dated Dec. 16, 2024, 6 pages. [cited by applicant]