BUILDING MANAGEMENT SYSTEMS WITH DYNAMIC EDGE COMPUTING ARCHITECTURES
A method for enhancing performance of a machine learning model executing on an edge building device of a building includes executing, by one or more processors of the edge building device, the machine learning model, generating, by the one or more processors of the edge building device, an assessment of the performance of the machine learning model on the edge building device, and responsive to the assessment indicating the performance of the machine learning model is below a first level: retraining, by the one or more processors of the edge building device, the machine learning model at the edge building device, and causing, by the one or more processors of the edge building device, a device other than the edge building device to retrain the machine learning model.
1 . A method for enhancing performance of a machine learning model executing on an edge building device of a building, the method comprising:
executing, by one or more processors of the edge building device, the machine learning model;
generating, by the one or more processors of the edge building device, an assessment of the performance of the machine learning model on the edge building device; and
responsive to the assessment indicating the performance of the machine learning model is below a first level:
retraining, by the one or more processors of the edge building device, the machine learning model at the edge building device; and/or
causing, by the one or more processors of the edge building device, a device other than the edge building device to retrain the machine learning model.
2 . The method of claim 1 , wherein generating the assessment of the performance of the machine learning model comprises:
generating a score indicating the performance of the machine learning model; and
the assessment indicates the performance of the machine learning model is below the first level responsive to the score being below a threshold level.
3 . The method of claim 1 , further comprising receiving operational data of the building generated by at least one of the edge building device or one or more other building devices, wherein the machine learning model is retrained using the operational data of the building.
4 . The method of claim 3 , wherein executing the machine learning model comprises initially executing the machine learning model without the machine learning model having been trained on data specific to the building or an entity associated with the building.
5 . The method of claim 3 , wherein the method comprises retraining the machine learning model at the edge device using the operational data.
6 . The method of claim 1 , wherein the method comprises causing a cloud computing system or other off-premises computing system to retrain the machine learning model.
7 . The method of claim 1 , further comprising selecting the machine learning model for executing by the edge building device from among a plurality of machine learning models based on at least one of design data or operational data for the building.
8 . The method of claim 1 , wherein generating the assessment of the performance of the machine learning model comprises monitoring the performance of the machine learning model over a timeframe, wherein the machine learning model is retrained and caused to be retrained responsive to detecting that the performance of the machine learning model has degraded below the first level.
9 . The method of claim 1 , wherein, responsive to the assessment indicating the performance of the machine learning model is below the first level, the method further comprises generating an alert to an analyst to review the performance of the machine learning model, wherein generating the alert to the analyst comprises generating a request to the analyst to validate whether the performance of the machine learning model is attributable to the machine learning model itself or an external factor.
10 . The method of claim 1 , wherein, responsive to the assessment indicating the performance of the machine learning model is below the first level, the method further comprises:
generating a request to one or more occupants of the building to provide feedback relating to the machine learning model;
receiving the feedback from the one or more occupants; and
validating the performance of the machine learning model using the feedback.
11 . A system for enhancing performance of a machine learning model executing on an edge building device of a building, the system comprising:
one or more non-transitory computer-readable media storing instructions thereon that, when executed by one or more processors, cause the one or more processors to:
execute the machine learning model;
generate an assessment of the performance of the machine learning model on the edge building device; and
responsive to the assessment indicating the performance of the machine learning model is below a first level:
retrain the machine learning model at the edge building device; and/or
cause a device other than the edge building device to retrain the machine learning model.
12 . The system of claim 11 , wherein:
generating the assessment of the performance of the machine learning model comprises generating a score indicating the performance of the machine learning model to monitor and/or control a sustainability performance of at least a portion of the building; and
the assessment indicates the performance of the machine learning model is below the first level responsive to the score being below a threshold level.
13 . The system of claim 11 , wherein the instructions further cause the one or more processors to: receive operational data of the building generated by at least one of the edge building device or one or more other building devices, wherein the machine learning model is retrained using the operational data of the building, wherein executing the machine learning model comprises initially executing the machine learning model without the machine learning model having been trained on data specific to the building or an entity associated with the building.
14 . The system of claim 11 , wherein generating the assessment of the performance of the machine learning model comprises monitoring the performance of the machine learning model over a timeframe, wherein the machine learning model is retrained and caused to be retrained responsive to detecting that the performance of the machine learning model has degraded below the first level.
15 . The system of claim 11 , wherein, responsive to the assessment indicating the performance of the machine learning model is below the first level, the instructions further cause the one or more processors to: generate an alert to an analyst to review the performance of the machine learning model, wherein generating the alert to the analyst comprises generating a request to the analyst to validate whether the performance of the machine learning model is attributable to the machine learning model itself or an external factor.
16 . The system of claim 11 , wherein, responsive to the assessment indicating the performance of the machine learning model is below the first level, the instructions further cause the one or more processors to:
generate a request to one or more occupants of the building to provide feedback relating to the machine learning model;
receive the feedback from the one or more occupants; and
validate the performance of the machine learning model using the feedback.
17 . One or more non-transitory storage media storing instructions thereon for enhancing performance of a machine learning model executing on an edge building device of a building that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
executing the machine learning model;
generating an assessment of the performance of the machine learning model on the edge building device, wherein the machine learning model is configured to monitor and/or control a sustainability performance of at least a portion of the building; and
responsive to the assessment indicating the performance of the machine learning model to monitor and/or control the sustainability performance of the at least a portion of the building is below a first level:
retraining the machine learning model at the edge building device; and/or
causing a device other than the edge building device to retrain the machine learning model.
18 . The one or more non-transitory storage media of claim 17 , wherein generating the assessment of the performance of the machine learning model comprises generating a score indicating the performance of the machine learning model to monitor and/or control carbon emissions of at least a portion of the building; and
the assessment indicates the performance of the machine learning model to monitor and/or control the sustainability performance of the at least a portion of the building is below the first level responsive to the score being below a threshold level.
19 . The one or more non-transitory storage media of claim 17 , wherein, responsive to the assessment indicating the performance of the machine learning model to monitor the sustainability performance of the at least a portion of the building is below the first level, the instructions further cause the one or more processors to perform operations comprising:
generating an alert to an analyst to review the performance of the machine learning model, wherein generating the alert to the analyst comprises generating a request to the analyst to validate whether the performance of the machine learning model is attributable to the machine learning model itself or an external factor.
20 . The one or more non-transitory storage media of claim 17 , wherein, responsive to the assessment indicating the performance of the machine learning model is below the first level, the instructions further cause the one or more processors to perform operations comprising:
generating a request to one or more occupants of the building to provide feedback relating to the machine learning model;
receiving the feedback from the one or more occupants; and
validating the performance of the machine learning model using the feedback.