Degradation assessment for a 5-dimensional twin framework
One example method includes receiving input from each degradation module in a group of degradation modules, and each of the degradation modules operates to track degradation of a respective dimension of a digital twin, determining, using the input, the individual degradation of each of the dimensions of the digital twin, determining, using the individual degradations for the dimensions, an overall degradation for the digital twin as a whole, and identifying, and implementing in an entity represented by the digital twin, a remedial action corresponding to one of the individual degradations.
1. A method, comprising: receiving input from each degradation module in a group of degradation modules, and each of the degradation modules operates to track degradation of a respective dimension of a digital twin; determining, using the input, an individual degradation of each of the dimensions of the digital twin; determining, using the individual degradations for the dimensions, an overall degradation for the digital twin as a whole; and identifying, and implementing in an entity represented by the digital twin, a remedial action corresponding to one of the individual degradations.
2. The method as recited in claim 1 , wherein the dimensions of the digital twin comprise any one or more of, a physical entity, a virtual entity, services, connections, and digital data.
3. The method as recited in claim 1 , wherein determining the overall degradation comprises weighting the individual degradations, and calculating the overall degradation using the weighting of the individual degradations.
4. The method as recited in claim 1 , wherein the digital twin comprises an inference model.
5. The method as recited in claim 4 , wherein a degradation of the inference model is determined using a trained shadow model.
6. The method as recited in claim 5 , wherein the shadow model was trained with output of an operational model, and with data generated using a quantified adversarial attack.
7. The method as recited in claim 1 , wherein a reliability metric for the digital twin is calculated that is based on the overall degradation of the digital twin.
8. The method as recited in claim 1 , wherein the digital twin comprises an inference model as one of the dimensions of the digital twin, and a degradation of the inference model is determined using one or both of (1) a distance between a reference time series and a time series for a virtual entity of the digital twin, and (2) a distance between a reference time series and a time series for a physical entity to which the digital twin corresponds.
9. The method as recited in claim 1 , wherein one or more of the degradation modules comprises a respective tuned degradation module.
10. The method as recited in claim 1 , wherein the input comprises data received from a sensor and data received from an inference model.
11. A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising: receiving input from each degradation module in a group of degradation modules, and each of the degradation modules operates to track degradation of a respective dimension of a digital twin; determining, using the input, an individual degradation of each of the dimensions of the digital twin; determining, using the individual degradations for the dimensions, an overall degradation for the digital twin as a whole; and identifying, and implementing in an entity represented by the digital twin, a remedial action corresponding to one of the individual degradations.
12. The non-transitory storage medium as recited in claim 11 , wherein the dimensions of the digital twin comprise any one or more of, a physical entity, a virtual entity, services, connections, and digital data.
13. The non-transitory storage medium as recited in claim 11 , wherein determining the overall degradation comprises weighting the individual degradations, and calculating the overall degradation using the weighting of the individual degradations.
14. The non-transitory storage medium as recited in claim 11 , wherein the digital twin comprises an inference model.
15. The non-transitory storage medium as recited in claim 14 , wherein a degradation of the inference model is determined using a trained shadow model.
16. The non-transitory storage medium as recited in claim 15 , wherein the shadow model was trained with output of an operational model, and with data generated using a quantified adversarial attack.
17. The non-transitory storage medium as recited in claim 11 , wherein a reliability metric for the digital twin is calculated that is based on the overall degradation of the digital twin.
18. The non-transitory storage medium as recited in claim 11 , wherein the digital twin comprises an inference model as one of the dimensions of the digital twin, and a degradation of the inference model is determined using one or both of (1) a distance between a reference time series and a time series for a virtual entity of the digital twin, and (2) a distance between a reference time series and a time series for a physical entity to which the digital twin corresponds.
19. The non-transitory storage medium as recited in claim 11 , wherein one or more of the degradation modules comprises a respective tuned degradation module.
20. The non-transitory storage medium as recited in claim 11 , wherein the input comprises data received from a sensor and data received from an inference model.