Automated vehicle repair estimation by voting ensembling of multiple artificial intelligence functions
Automated vehicle repair estimation by voting ensembling of multiple artificial intelligence functions is provided. A method comprises receiving a plurality of vehicle repair recommendation sets, each identifying (i) at least one component of a damaged vehicle, (ii) a recommended vehicle repair operation for each identified component, and (iii) a score and/or confidence percentage for each operation; when a plurality of the sets identify recommended operations for one of the components, selecting the operation having the highest score, and unselecting the other operations for the component; generating a composite vehicle repair recommendation set, wherein the composite vehicle repair recommendation set identifies the selected recommended vehicle repair operation, and wherein the composite vehicle repair recommendation set does not identify the unselected recommended vehicle repair operation; and providing the composite vehicle repair recommendation set to one or more claims management systems.
1. A system, comprising:
a hardware processor; and
a non-transitory machine-readable storage medium encoded with instructions executable by the hardware processor to perform a method comprising:
receiving a plurality of vehicle repair recommendation sets for a damaged vehicle, wherein each of the vehicle repair recommendation sets is generated by a respective artificial intelligence function, wherein at least two of the artificial intelligence functions differ from each other, and wherein each of the vehicle repair recommendation sets identifies (i) at least one component of the damaged vehicle, (ii) a recommended vehicle repair operation for each identified component, and (iii) a score and/or confidence percentage for each recommended vehicle repair operation;
wherein each artificial intelligence function is implemented as a trained computer vision machine learning model, wherein each trained computer vision machine learning model is trained with historical images of other damaged vehicles and corresponding vehicle repair operations applied to repair the other damaged vehicles;
when a plurality of the vehicle repair recommendation sets identify recommended vehicle repair operations for one of the components of the damaged vehicle, selecting the recommended vehicle repair operation for the one of the components having the highest score, and unselecting the other recommended vehicle repair operations for the one of the components;
generating a composite vehicle repair recommendation set, wherein the composite vehicle repair recommendation set identifies the selected recommended vehicle repair operation, and wherein the composite vehicle repair recommendation set does not identify the unselected recommended vehicle repair operation; and
providing the composite vehicle repair recommendation set to one or more claims management systems.
2. The system of claim 1 , wherein:
each of the vehicle repair recommendation sets identifies one or more images of the damaged vehicle; and
the method further comprises:
selecting the images identified by the selected recommended vehicle repair operation, and
identifying the selected images in the generated composite vehicle repair recommendation set.
3. The system of claim 1 , the method further comprising:
identifying the score of the selected recommended vehicle repair operation in the generated composite vehicle repair recommendation set.
4. The system of claim 1 , the method further comprising:
generating a composite score based on the score of the selected recommended vehicle repair operation and the score of each unselected recommended vehicle repair operation.
5. The system of claim 1 , wherein generating a composite score comprises one of:
taking the highest score; and
averaging the scores.
6. The system of claim 1 , wherein:
each of the artificial intelligence functions is trained; and
the method further comprises: requesting one or more of the artificial intelligence functions be re-trained when a predetermined event occurs.
7. The system of claim 1 , wherein selecting the recommended vehicle repair operation for the one of the components having the highest score comprises:
selecting the recommended vehicle repair operation for the one of the components having the highest score based on statistical aggregation methodologies comprising at least one of highest voting, maximal scores, counting, and plurality of maximal score votes.
8. A non-transitory machine-readable storage medium encoded with instructions executable by a hardware processor of a computing component, the machine-readable storage medium comprising instructions to cause the hardware processor to perform a method comprising:
receiving a plurality of vehicle repair recommendation sets for a damaged vehicle, wherein each of the vehicle repair recommendation sets is generated by a respective artificial intelligence function, wherein at least two of the artificial intelligence functions differ from each other, and wherein each of the vehicle repair recommendation sets identifies (i) at least one component of the damaged vehicle, (ii) a recommended vehicle repair operation for each identified component, and (iii) a score and/or confidence percentage for each recommended vehicle repair operation;
wherein each artificial intelligence function is implemented as a trained computer vision machine learning model, wherein each trained computer vision machine learning model is trained with historical images of other damaged vehicles and corresponding vehicle repair operations applied to repair the other damaged vehicles;
when a plurality of the vehicle repair recommendation sets identify recommended vehicle repair operations for one of the components of the damaged vehicle, selecting the recommended vehicle repair operation for the one of the components having the highest score, and unselecting the other recommended vehicle repair operations for the one of the components;
generating a composite vehicle repair recommendation set, wherein the composite vehicle repair recommendation set identifies the selected recommended vehicle repair operation, and wherein the composite vehicle repair recommendation set does not identify the unselected recommended vehicle repair operation; and
providing the composite vehicle repair recommendation set to one or more claims management systems.
9. The non-transitory machine-readable storage medium of claim 8 , wherein:
each of the vehicle repair recommendation sets identifies one or more images of the damaged vehicle; and
the method further comprises:
selecting the images identified by the selected recommended vehicle repair operation, and
identifying the selected images in the generated composite vehicle repair recommendation set.
10. The non-transitory machine-readable storage medium of claim 8 , the method further comprising:
identifying the score of the selected recommended vehicle repair operation in the generated composite vehicle repair recommendation set.
11. The non-transitory machine-readable storage medium of claim 8 , the method further comprising:
generating a composite score based on the score of the selected recommended vehicle repair operation and the score of each unselected recommended vehicle repair operation.
12. The non-transitory machine-readable storage medium of claim 8 , wherein generating a composite score comprises one of:
taking the highest score; and
averaging the scores.
13. The non-transitory machine-readable storage medium of claim 8 , wherein:
each of the artificial intelligence functions is trained; and
the method further comprises: requesting one or more of the artificial intelligence functions be re-trained when a predetermined event occurs.
14. The non-transitory machine-readable storage medium of claim 8 , wherein selecting the recommended vehicle repair operation for the one of the components having the highest score comprises:
selecting the recommended vehicle repair operation for the one of the components having the highest score based on statistical aggregation methodologies comprising at least one of highest voting, maximal scores, counting, and plurality of maximal score votes.
15. A method comprising:
receiving a plurality of vehicle repair recommendation sets for a damaged vehicle, wherein each of the vehicle repair recommendation sets is generated by a respective artificial intelligence function, wherein at least two of the artificial intelligence functions differ from each other, and wherein each of the vehicle repair recommendation sets identifies (i) at least one component of the damaged vehicle, (ii) a recommended vehicle repair operation for each identified component, and (iii) a score and/or confidence percentage for each recommended vehicle repair operation;
wherein each artificial intelligence function is implemented as a trained computer vision machine learning model, wherein each trained computer vision machine learning model is trained with historical images of other damaged vehicles and corresponding vehicle repair operations applied to repair the other damaged vehicles;
when a plurality of the vehicle repair recommendation sets identify recommended vehicle repair operations for one of the components of the damaged vehicle, selecting the recommended vehicle repair operation for the one of the components having the highest score, and unselecting the other recommended vehicle repair operations for the one of the components;
generating a composite vehicle repair recommendation set, wherein the composite vehicle repair recommendation set identifies the selected recommended vehicle repair operation, and wherein the composite vehicle repair recommendation set does not identify the unselected recommended vehicle repair operation; and
providing the composite vehicle repair recommendation set to one or more claims management systems.
16. The method of claim 15 , wherein:
each of the vehicle repair recommendation sets identifies one or more images of the damaged vehicle; and
the method further comprises:
selecting the images identified by the selected recommended vehicle repair operation, and
identifying the selected images in the generated composite vehicle repair recommendation set.
17. The method of claim 15 , further comprising:
identifying the score of the selected recommended vehicle repair operation in the generated composite vehicle repair recommendation set.
18. The method of claim 15 , further comprising:
generating a composite score based on the score of the selected recommended vehicle repair operation and the score of each unselected recommended vehicle repair operation.
19. The method of claim 15 , wherein generating a composite score comprises one of:
taking the highest score; and
averaging the scores.
20. The method of claim 15 , wherein:
each of the artificial intelligence functions is trained; and
the method further comprises: requesting one or more of the artificial intelligence functions be re-trained when a predetermined event occurs.
21. The method of claim 15 , wherein selecting the recommended vehicle repair operation for the one of the components having the highest score comprises:
selecting the recommended vehicle repair operation for the one of the components having the highest score based on statistical aggregation methodologies comprising at least one of highest voting, maximal scores, counting, and plurality of maximal score votes.