Differential multi model training for multiple interpretation options
Computing systems, computer-readable media, and methods for providing multiple computer-generated seismic data interpretation options, of which the method includes receiving a training input, sorting the training input into a first group and a second group, subgrouping the second group into a plurality of subgroups, generating a plurality of trained models based on the plurality of subgroups and the first group, receiving a prediction input having a set of data to be interpreted, generating a plurality of interpretation options for the prediction input by applying the plurality of training models to the prediction input, and outputting the plurality of interpretation options.
1. A method, comprising:
receiving a training input comprising a plurality of interpretations for a set of training seismic data, each of the plurality of interpretations corresponding to a respective data interpreter;
sorting the training input into: a first group comprising portions of the plurality of interpretations that are in agreement for all of the plurality of interpretations; and
a second group comprising portions of the plurality of interpretations that are in disagreement for at least two of the plurality of interpretations;
subgrouping the second group into a plurality of subgroups, each of the plurality of subgroups corresponding to one of the plurality of interpretations;
generating a plurality of machine-learning trained models, each of the plurality of machine- learning trained models being based on a respective one of the plurality of subgroups and the first group;
receiving a prediction input having a set of data to be interpreted;
generating a plurality of interpretation options for the prediction input by applying each of the plurality of machine-learning trained models to the prediction input, each of the plurality of interpretation options corresponding to a respective data interpreter;
outputting the plurality of interpretation options;
assigning a respective score to each of the plurality of interpretation options, each respective score being assigned according to at least one scoring rule, the at least one scoring rule including:
assigning a better score to an interpretation corresponding to a data interpreter having a greater experience level than to an interpretation corresponding to another data interpreter having a lesser experience level;
assigning a better score to an interpretation corresponding to a data interpreter who uses one or more particular techniques; or
assigning a better score to an interpretation corresponding to a data interpreter specializing in a particular geographic location corresponding to the prediction input;
sequentially sorting the plurality of interpretation options based on the assigned scores;
selecting at least one of the outputted plurality of interpretation options, the selecting being in sequential order according to the sorting and including first selecting a best score among the scores; and
sending a signal to a controller to cause a drilling tool in a drilling operation to deviate from an original drilling plan based on the selected at least one of the outputted plurality of interpretation options to perform at least one mitigating action comprising controlling at least one of: drilling, weight on a bit, a pump rate, or a physical parameter of the drilling operation,
wherein the at least one mitigating action improves an operating condition of the drilling operation based on the selected at least one of the outputted plurality of interpretation options or avoids a problem identified by the selected at least one of the outputted plurality of interpretation options.
2. The method of claim 1 , wherein the plurality of interpretations includes interpretations of seismic data.
3. The method of claim 2 , wherein:
the first group comprises a portion of the training input in which interpretations of the seismic data match; and
the second group comprises a portion of the training input in which the interpretations of the seismic data do not match.
4. The method of claim 3 , wherein individual subgroups of the plurality of subgroups are associated with interpretation attributes.
5. The method of claim 1 , wherein the prediction input includes target seismic data to be interpreted.
6. The method of claim 1 , wherein the machine-learning trained models are further generated based on a supervised machine learning technique.
7. A computing system, comprising:
one or more processors; and
a memory system comprising one or more non-transitory computer-readable media storing instructions that, when executed by at least one of the one or more processors, cause the computing system to perform operations, the operations comprising:
receiving a training input comprising a plurality of interpretations for a set of training seismic data, each of the plurality of interpretations corresponding to a respective data interpreter;
sorting the training input into:
a first group comprising portions of the plurality of interpretations that are in agreement for all of the plurality of interpretations; and
a second group comprising portions of the plurality of interpretations that are in disagreement for at least two of the plurality of interpretations;
subgrouping the second group into a plurality of subgroups, each of the plurality of subgroups corresponding to one of the plurality of interpretations;
generating a plurality of machine-learning trained models; each of the plurality of machine-learning trained models being based on a respective one of the plurality of subgroups and the first group;
receiving a prediction input having a set of data to be interpreted;
generating a plurality of interpretation options for the prediction input by applying each of the plurality of machine-learning trained models to the prediction input, each of the plurality of interpretation options corresponding to a respective data interpreter;
outputting the plurality of interpretation options;
assigning a respective score to each of the plurality of interpretation options, each respective score being assigned according to at least one scoring rule, the at least one scoring rule including:
assigning a better score to an interpretation corresponding to a data interpreter having a greater experience level than to an interpretation corresponding to another data interpreter having a lesser experience level;
assigning a better score to an interpretation corresponding to a data interpreter who uses one or more particular techniques; or
assigning a better score to an interpretation corresponding to a data interpreter specializing in a particular geographic location corresponding to the prediction input;
sequentially sorting the plurality of interpretation options based on the assigned scores;
selecting at least one of the outputted plurality of interpretation options, the selecting being in sequential order according to the sorting and including first selecting a best score among the scores; and
the one or more processors sending a signal to a controller to cause a drilling tool in a drilling operation to deviate from an original drilling plan based on the selected at least one of the outputted plurality of interpretation options to perform at least one mitigating action comprising controlling at least one of: drilling, weight on a bit, a pump rate, or a physical parameter of the drilling operation,
wherein the at least one mitigating action improves an operating condition of the drilling operation based on the selected at least one of the outputted plurality of interpretation options or avoids a problem identified by the selected at least one of the outputted plurality of interpretation options.
8. The system of claim 7 , wherein the plurality of interpretations includes interpretations of seismic data.
9. The system of claim 8 , wherein:
the first group comprises a portion of the training input in which interpretations of the seismic data match; and
the second group comprises a portion of the training input in which the interpretations of the seismic data do not match.
10. The system of claim 9 , wherein individual subgroups of the plurality of subgroups are associated with interpretation attributes.
11. The system of claim 7 , wherein the prediction input includes target seismic data to be interpreted.
12. The system of claim 7 , wherein the machine-learning trained models are further generated based on a supervised machine learning technique.
13. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations, the operations comprising:
receiving a training input comprising a plurality of interpretations for a set of training seismic data, each of the plurality of interpretations corresponding to a respective data interpreter;
sorting the training input into:
a first group comprising portions of the plurality of interpretations that are in agreement for all of the plurality of interpretations; and
a second group comprising portions of the plurality of interpretations that are in disagreement for at least two of the plurality of interpretations;
subgrouping the second group into a plurality of subgroups, each of the plurality of subgroups corresponding to one of the plurality of interpretations;
generating a plurality of machine-learning trained models, each of the plurality of machine-learning trained models being based on a respective one of the plurality of subgroups and the first group;
receiving a prediction input having a set of data to be interpreted;
generating a plurality of interpretation options for the prediction input by applying each of the plurality of machine-learning trained models to the prediction input, each of the plurality of interpretation options corresponding to a respective data interpreter;
outputting the plurality of interpretation options;
assigning a respective score to each of the plurality of interpretation options, each respective score being assigned according to at least one scoring rule, the at least one scoring rule including:
assigning a better score to an interpretation corresponding to a data interpreter having a greater experience level than to an interpretation corresponding to another data interpreter having a lesser experience level;
assigning a better score to an interpretation corresponding to a data interpreter who uses one or more particular techniques; or
assigning a better score to an interpretation corresponding to a data interpreter specializing in a particular geographic location corresponding to the prediction input;
sequentially sorting the plurality of interpretation options based on the assigned scores;
selecting at least one of the outputted plurality of interpretation options, the selecting being in sequential order according to the sorting and including first selecting a best score among the scores; and
the one or more processors sending a signal to a controller to cause a drilling tool in a drilling operation to deviate from an original drilling plan based on the selected at least one of the outputted plurality of interpretation options to perform at least one mitigating action comprising controlling at least one of: drilling, weight on a bit, a pump rate, or a physical parameter of the drilling operation,
wherein the at least one mitigating action improves an operating condition of the drilling operation based on the selected at least one of the outputted plurality of interpretation options or avoids a problem identified by the selected at least one of the outputted plurality of interpretation options.
14. The computer-readable medium of claim 13 , wherein the plurality of interpretations includes interpretations of seismic data.
15. The computer-readable medium of claim 13 , wherein:
the first group comprises a portion of the training input in which interpretations of the seismic data match; and
the second group comprises a portion of the training input in which the interpretations of the seismic data do not match.
16. The computer-readable medium of claim 13 , wherein individual subgroups of the plurality of subgroups are associated with interpretation attributes.
17. The computer-readable medium of claim 13 , wherein the prediction input includes target seismic data to be interpreted.
18. The computer-readable medium of claim 13 , wherein the machine-learning trained models are further generated based on a supervised machine learning technique.