Hybrid quantized decision model framework

1. A system, comprising:

at least one computing device; and

program instructions stored in at least one memory of the at least one computing device, wherein the instructions, when executed by at least one processor, cause the at least one computing device to at least:

receive, by an edge device, device layer data from at least one device;

generate, by the edge device, an edge layer prediction by processing the device layer data using a quantized decision model comprising a decreased computational requirement relative to a cloud layer decision model, wherein the quantized decision model is stored on the edge device as a quantized version of the cloud layer decision model stored in a cloud computing environment;

perform, by the edge device, a confidence check based on an edge layer decision threshold, and a confidence probability for the edge layer prediction; and

identify, by the edge device, an action to perform based on the confidence check.

2. The system of claim 1 , wherein the action corresponds to the edge layer prediction based on the confidence check indicating that the confidence probability for the edge layer prediction is greater than or equal to the edge layer decision threshold.

3. The system of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

transmit a cloud layer prediction request from the edge device to a cloud computing environment based on the confidence check indicating that the confidence probability for the edge layer prediction is less than the edge layer decision threshold; and

receive, by the edge device, a cloud layer prediction based on the cloud layer decision model stored in the cloud computing environment, wherein the action corresponds to the cloud layer prediction.

4. The system of claim 3 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

receive, by the edge device, the quantized decision model from the cloud computing environment.

5. The system of claim 1 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

perform, by the edge device, a latency check based on a cloud consultation threshold, and a round trip latency to contact a cloud computing environment, wherein the latency check is performed in response to the confidence check indicating that the confidence probability for the edge layer prediction is less than the edge layer decision threshold.

6. The system of claim 5 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

transmit a cloud layer prediction request from the edge device to the cloud computing environment based on the latency check indicating that the round trip latency is less than or equal to a cloud consultation threshold latency.

7. The system of claim 5 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

receive, by the edge device, a notification to perform the action corresponding to the edge layer prediction based on the latency check indicating that the round trip latency is greater than a cloud consultation threshold latency; and

schedule a correction time checking process that determines whether to transmit a cloud layer prediction request based on: a current remaining correction time to correct the edge layer prediction, and a maximum affordable waiting time to correct the edge layer prediction.

8. A method, comprising:

receiving, by an edge device, device layer data from at least one device;

generating, by the edge device, an edge layer prediction by processing the device layer data using a quantized decision model comprising a decreased computational requirement relative to a cloud layer decision model, wherein the quantized decision model is stored on the edge device as a quantized version of the cloud layer decision model stored in a cloud computing environment;

performing, by the edge device, a confidence check based on an edge layer decision threshold, and a confidence probability for the edge layer prediction; and

identifying, by the edge device, an action to perform based on the confidence check.

9. The method of claim 8 , wherein the action corresponds to the edge layer prediction based on the confidence check indicating that the confidence probability for the edge layer prediction is greater than or equal to the edge layer decision threshold.

10. The method of claim 8 , further comprising:

transmitting a cloud layer prediction request from the edge device to a cloud computing environment based on the confidence check indicating that the confidence probability for the edge layer prediction is less than the edge layer decision threshold; and

receiving, by the edge device, a cloud layer prediction based on the cloud layer decision model stored in the cloud computing environment, wherein the action corresponds to the cloud layer prediction.

11. The method of claim 10 , further comprising:

receiving, by the edge device, the quantized decision model from the cloud computing environment.

12. The method of claim 8 , further comprising:

performing, by the edge device, a latency check based on a cloud consultation threshold, and a round trip latency to contact a cloud computing environment, wherein the latency check is performed in response to the confidence check indicating that the confidence probability for the edge layer prediction is less than the edge layer decision threshold.

13. The method of claim 12 , further comprising:

transmitting a cloud layer prediction request from the edge device to the cloud computing environment based on the latency check indicating that the round trip latency is less than or equal to a cloud consultation threshold latency.

14. The method of claim 12 , further comprising:

receiving, by the edge device, a notification to perform the action corresponding to the edge layer prediction based on the latency check indicating that the round trip latency is greater than a cloud consultation threshold latency; and

scheduling a correction time checking process that determines whether to transmit a cloud layer prediction request based on: a current remaining correction time to correct the edge layer prediction, and a maximum affordable waiting time to correct the edge layer prediction.

15. A non-transitory computer-readable medium comprising program instructions that when executed by at least one processor, cause at least one computing device to at least:

receive, by an edge device, device layer data from at least one device;

generate, by the edge device, an edge layer prediction by processing the device layer data using a quantized decision model comprising a decreased computational requirement relative to a cloud layer decision model, wherein the quantized decision model is stored on the edge device as a quantized version of the cloud layer decision model stored in a cloud computing environment;

perform, by the edge device, a confidence check based on an edge layer decision threshold, and a confidence probability for the edge layer prediction; and

identify, by the edge device, an action to perform based on the confidence check.

16. The non-transitory computer-readable medium of claim 15 , wherein

the action corresponds to the edge layer prediction based on the confidence check indicating that the confidence probability for the edge layer prediction is greater than or equal to the edge layer decision threshold.

17. The non-transitory computer-readable medium of claim 15 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

transmit a cloud layer prediction request from the edge device to a cloud computing environment based on the confidence check indicating that the confidence probability for the edge layer prediction is less than the edge layer decision threshold; and

receive, by the edge device, a cloud layer prediction based on the cloud layer decision model stored in the cloud computing environment, wherein the action corresponds to the cloud layer prediction.

18. The non-transitory computer-readable medium of claim 17 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

receive, by the edge device, the quantized decision model from the cloud computing environment.

19. The non-transitory computer-readable medium of claim 15 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

perform, by the edge device, a latency check based on a cloud consultation threshold, and a round trip latency to contact a cloud computing environment, wherein the latency check is performed in response to the confidence check indicating that the confidence probability for the edge layer prediction is less than the edge layer decision threshold.

20. The non-transitory computer-readable medium of claim 19 , wherein the instructions, when executed by the at least one processor, cause the at least one computing device to at least:

transmit a cloud layer prediction request from the edge device to the cloud computing environment based on the latency check indicating that the round trip latency is less than or equal to a cloud consultation threshold latency.