# US 2020/0401944 A1 · Mechanism for machine learning in distributed computing > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2020/0401944 A1 - Document ID: US20200401944A1 - Status: Published application - Kind: A1 - Application number: 16/970,479 - Filed: 2019-04-01 - Published: 2020-12-24 - Earliest claimed: 2018-04-27 - Assignees: Sony Corporation - Applicants: Sony Corporation - Inventors: Henrik SUNDSTRÖM; Basuki PRIYANTO; Andrej PETEF; Lars NORD; Anders ISBERG - Web page: https://patentgrep.com/patent/US20200401944A1 ## CPC - G06F 9/4881: {Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues} - G06N 20/00: Machine learning - G06F 9/505: {considering the load} - G06F 9/5072: {Grid computing} - G06F 9/5088: {involving task migration} - G06N 5/04: Inference or reasoning models - Y02D 10/00: Energy efficient computing, e.g. low power processors, power management or thermal management - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE ## Abstract A method for distributed computation in a hierarchical system having a compute deployment including a plurality of compute nodes, comprising providing a control function communicatively connected to said compute nodes; determining a cost function for the system, which cost function includes at least one first parameter associated with carrying out a compute task and at least one second parameter associated with escalating a compute task; employing a machine learning mechanism in the control function to optimize said cost function; and configuring said compute deployment based on the optimization of the cost function by the machine learning mechanism. ## Claims ### Claim 1 A method for distributed computation in a hierarchical system having a compute deployment including a plurality of compute nodes, wherein each node is configured to execute a respective estimation model to obtain a confidence level of an estimation model output for carrying out a compute task, the method comprising providing a control function communicatively connected to said compute nodes; determining a cost function for the system, which cost function includes at least one first parameter associated with power consumption for carrying out said compute task in a first node of said nodes, and at least one second parameter associated with bandwidth utilization for escalating the compute task from the first node to a second node in the hierarchical system; employing a machine learning mechanism in the control function to optimize said cost function on one or more overall system cost parameters; and configuring said compute deployment based on the optimization of the cost function by the machine learning mechanism, including adjusting a confidence level threshold to be used by the estimation model in one or more of said nodes. ### Claim 2 (depends on 1) The method of claim 1, comprising receiving first metrics from one or more of said nodes associated with a compute task; and determining one or more of said first and/or second parameters based on said metrics. ### Claim 3 (depends on 1) The method of claim 1, wherein configuring said compute deployment includes providing compute deployment data to at least one of said nodes. ### Claim 4 (depends on 1) The method of claim 1, wherein configuring said compute deployment includes adjusting a confidence level threshold in one or more of said nodes. ### Claim 5 (depends on 1) The method of claim 1, wherein configuring said compute deployment includes updating a computation model in one or more of said nodes. ### Claim 6 (depends on 1) The method of claim 1, wherein said cost function includes a weight associated to one or more of the first and/or second parameters. ### Claim 7 (depends on 1) The method of claim 1, wherein said first parameter is associated with carrying out a compute task in a node of the system and depends on at least one metric of the group: confidence threshold values, confidence level of an estimation model output, power consumption, bandwidth utilization, latency, sensor data. ### Claim 8 (depends on 1) The method of claim 1, wherein said second parameter is associated with escalating a compute task between nodes in the system and depends on at least one metric of the group: of latency, bandwidth utilization, power consumption, autonomy, privacy protection, security. ### Claim 9 (depends on 2) The method of claim 2, wherein said cost function comprises a weighted sum of said metrics. ### Claim 10 (depends on 1) The method of claim 1, wherein said machine learning mechanism includes a reinforcement algorithm, the method further comprising, based on the reinforcement algorithm, configured to optimize control function decisions over time to take action to improve a current compute deployment state based on an observed environment including metrics received from said plurality of nodes. ### Claim 11 (depends on 1) The method of claim 1, comprising receiving a compute task; controlling a compute node to carry out the received compute task in accordance with the configured compute deployment. ### Claim 12 (depends on 11) The method of claim 11, wherein controlling a compute node to carry out the received compute task includes one of carrying out the compute task in the compute node in which the compute task was received; or escalating the compute task from a compute node in which the compute task was received to the compute node controlled to carry out the compute task. ### Claim 13 A non-transitory computer readable medium storing a computer program product in the form of executable instructions for managing distributed computation in a hierarchical system having a compute deployment including a plurality of compute nodes, wherein each node is configured to execute a respective estimation model to obtain a confidence level of an estimation model output for carrying out a compute task, wherein the executable instructions are configured to determine a cost function for the system, which cost function includes at least one first parameter associated with power consumption for carrying out said compute task in a first node of said nodes, and at least one second parameter associated with bandwidth utilization for escalating said compute task from the first node to a second node in the hierarchical system; employ a machine learning mechanism in the control function to optimize said cost function on one or more overall system cost parameters; and configure said compute deployment based on the optimization of the cost function by the machine learning mechanism, including adjusting a confidence level threshold to be used by the estimation model in one or more of said nodes. ### Claim 14 (depends on 13) A computer system comprising control circuitry, which control circuitry includes a processing device and the non-transitory computer readable medium of claim 13 inclusive of the executable instructions. ### Claim 15 (canceled) ### Claim 16 A hierarchical system comprising a compute deployment including a plurality of compute nodes, wherein each node is configured to execute a respective estimation model to obtain a confidence level of an estimation model output for carrying out a compute task, and a control function communicatively connected to said compute nodes, wherein said control function comprises a computer program product for managing distributed computation in the hierarchical system, configured to determine a cost function for the system, which cost function includes at least one first parameter associated with power consumption for carrying out said compute task in a first node of said nodes, and at least one second parameter associated with power consumption for escalating said compute task from the first node to a second node in the hierarchical system; employ a machine learning mechanism in the control function to optimize said cost function on one or more overall system cost parameters; and configure said compute deployment based on the optimization of the cost function by the machine learning mechanism, including adjusting a confidence level threshold to be used by the estimation model in one or more of said nodes. ### Claim 17 (canceled) ### Claim 18 (canceled) ### Claim 19 (depends on 7) The method of claim 7, wherein said cost function comprises a weighted sum of said metrics. ### Claim 20 (depends on 8) The method of claim 8, wherein said cost function comprises a weighted sum of said metrics.