Vehicle resiliency, driving feedback and risk assessment using machine learning-based vehicle wear scoring
A machine learning model is manufactured by a process including retrieving training data, minimizing a loss function, wherein the training data may include labeled or unlabeled data, the machine learning model generating a prediction. A machine learning training/operation server includes a processor and a memory storing instructions that, when executed by the processor, cause the server to retrieve training data, input a training input, analyze the training input to generate a prediction, generate a loss score, and store the trained machine learning model. A method for training a machine learning model includes receiving training data, inputting a training input, analyzing the training input, generating a loss score, and storing the trained machine learning model.
1. A non-transitory computer readable medium storing instructions thereon, when executed by one or more processors, causing the one or more processors perform operations comprising:
retrieving a machine learning model;
wherein the machine learning model is trained until a loss score meets a predetermined criteria by at least:
inputting a training input and corresponding label,
analyzing the training input using the machine learning model to generate a training prediction,
generating the loss score by comparing the training prediction to the corresponding label using a loss function, and
modifying at least one of one or more weights and one or more biases of the machine learning model based at least in part upon the loss score;
receiving an operational input; and
analyzing the operational input using the trained machine learning model to obtain a prediction;
wherein:
the operational input includes a set of component wear scores of a vehicle, and
the prediction includes an aggregated wear score of the vehicle generated by analyzing the set of component wear scores.
2. The non-transitory computer readable medium of claim 1 , wherein the operations further comprise:
analyzing the aggregated wear score and type information of the vehicle to predict one or more general wear and tear characteristics of the vehicle.
3. The non-transitory computer readable medium of claim 1 , wherein:
the operational input includes telematics data corresponding to one or more vehicles operated by an operator of the vehicle, vehicle type data, or vehicle age; and
the prediction includes an estimated wear of the vehicle.
4. The non-transitory computer readable medium of claim 1 , wherein the operational input is adjusted programmatically or by a user to simulate a partial effect.
5. The non-transitory computer readable medium of claim 1 ,
wherein the operational input includes telematics data, and
wherein the analyzing the operational input using the trained machine learning model to obtain a prediction includes analyzing the telematics data to categorize collision risk of a vehicle operator of the vehicle.
6. The non-transitory computer readable medium of claim 5 , wherein the operations further comprise:
transmitting a feedback notification to the vehicle operator, the feedback notification including a warning in regard to the collision risk and a recommendation for reducing the collision risk.
7. The non-transitory computer readable medium of claim 1 , wherein the trained machine learning model includes a first machine learning model to determine a vehicle wear score and a second machine learning model to generate a risk assessment.
8. A machine learning training and operation server comprising: one or more processors; and a memory storing instructions that, when executed by the one or more processors, cause the server to:
retrieve a machine learning model;
wherein the machine learning model is trained until a loss score meets a predetermined criteria by at least:
inputting a training input and corresponding label;
analyzing the training input to generate a training prediction;
generating the loss score by comparing the training prediction to the label using a loss function; and
modifying at least one of one or more weights and one or more biases of the machine learning model based at least in part upon the loss score;
instantiate the trained machine learning model;
receive an operational input; and
analyze the operational input using the trained machine learning model to obtain a prediction
wherein:
the operational input includes a set of component wear scores of a vehicle, and
the prediction includes an aggregated wear score of the vehicle generated by analyzing the set of component wear scores.
9. The server of claim 8 , wherein the instructions further cause the server to:
analyze the aggregated wear score and type information of the vehicle to predict the one or more general wear and tear characteristics of the vehicle.
10. The server of claim 8 , wherein the instructions further cause the server to:
receive telematics data corresponding to one or more vehicles operated by an operator of the vehicle, vehicle type data, or vehicle age; and
generate an estimated wear of the vehicle.
11. The server of claim 8 , wherein the trained machine learning model includes a first machine learning model to determine a vehicle wear score and a second machine learning model to generate a risk assessment.
12. A computer-implemented method, the method including:
retrieving, via one or more processors, a machine learning model;
wherein the machine learning model is trained until a loss score meets a predetermined criteria by at least:
inputting to the machine learning model a training input and corresponding label;
analyzing the training input using the machine learning model to generate a training prediction;
generating the loss score by comparing the training prediction to the label using a loss function; and
modifying at least one of one or more weights and one or more biases of the machine learning model based at least in part upon the loss score;
instantiating the trained machine learning model;
receiving an operational input; and
analyzing the operational input using the trained machine learning model to obtain a prediction;
wherein:
the operational input includes a set of component wear scores of a vehicle, and
the prediction includes an aggregated wear score of the vehicle generated by analyzing the set of component wear scores.
13. The computer-implemented method of claim 12 , further comprising:
analyzing the aggregated wear score and type information of the vehicle to predict one or more general wear and tear characteristics of the vehicle.
14. The computer-implemented method of claim 12 , wherein:
the operational input includes telematics data corresponding to one or more vehicles operated by an operator of the vehicle, vehicle type data, and or vehicle age; and
the prediction corresponds to an estimated wear of the vehicle.
15. The computer-implemented method of claim 12 , wherein the operational input includes telematics data,
wherein the analyzing the operational input to obtain a prediction includes analyzing the telematics data to categorize collision risk of a vehicle operator of the vehicle.
16. The computer-implemented method of claim 15 , further comprising:
transmitting a feedback notification to the vehicle operator, the feedback notification including a warning in regard to collision risk and a recommendation for reducing the collision risk.
17. The computer-implemented method of claim 12 , wherein the trained machine learning model includes a first machine learning model to determine a vehicle wear score and a second machine learning model to generate a risk assessment.