# US 2023/0066080 A1 · Data sequence prediction and resource allocation

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0066080 A1
- Document ID: US20230066080A1
- Status: Published application
- Kind: A1
- Application number: 17/463,601
- Filed: 2021-09-01
- Published: 2023-03-02
- Earliest claimed: 2021-09-01
- Assignees: Micron Technology, Inc.
- Applicants: Micron Technology, Inc.
- Inventors: Reshmi Basu; David A. Palmer; Jonathan S. Parry
- Web page: https://patentgrep.com/patent/US20230066080A1

## CPC

- G06F 9/5016: {the resource being the memory}
- G06N 3/0499: Feedforward networks
- G06F 9/5005: {to service a request}
- G06F 9/5033: {considering data affinity}
- G06F 9/5083: {Techniques for rebalancing the load in a distributed system}
- G06N 3/04: Architecture, e.g. interconnection topology
- G06F 2209/5019: Workload prediction
- G06N 3/048: Activation functions
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method for data sequence prediction and resource allocation includes determining, by a memory system, a plurality of resource parameters associated with operation of the memory system and determining respective time intervals associated with usage patterns corresponding to the memory system, the respective time intervals being associated with one or more sets of the plurality of resource parameters. The method further includes determining, using the plurality of resource parameters, one or more weights for hidden layers of a neural network for the respective time intervals associated with the usage patterns and allocating computing resources within the memory system for use in execution of workloads based on the determined one or more weights for hidden layers of the neural network.

## Claims

### Claim 1

A method, comprising:
determining, by a memory system, a plurality of resource parameters associated with operation of the memory system;
determining respective time intervals associated with usage patterns corresponding to the memory system, the respective time intervals being associated with one or more sets of the plurality of resource parameters;
determining, using the plurality of resource parameters, one or more weights for hidden layers of a neural network for the respective time intervals associated with the usage patterns; and
allocating computing resources within the memory system for use in execution of workloads based on the determined one or more weights for hidden layers of the neural network.

### Claim 2 (depends on 1)

The method of claim 1, further comprising determining respective autocorrelation time intervals for data traffic processed by the memory system as part of determining the respective time intervals.

### Claim 3 (depends on 1)

The method of claim 1, further comprising allocating the computing resources amongst a plurality of volatile memory resources or a plurality of non-volatile memory resources, or both, associated with the memory system.

### Claim 4 (depends on 1)

The method of claim 1, further comprising determining the plurality of resource parameters associated with operation of the memory system using received individual user workload parameters and received aggregate workload parameters corresponding to other users of a cloud network to which the memory system is communicatively coupled.

### Claim 5 (depends on 1)

The method of claim 1, further comprising determining the plurality of resource parameters associated with operation of the memory system based on characteristics corresponding to execution of workloads by memory devices resident on the memory system.

### Claim 6 (depends on 1)

The method of claim 1, further comprising determining a command latency associated with commands received to a front-end portion of the memory system as part of determining the respective time intervals associated with the usage patterns corresponding to the memory system.

### Claim 7 (depends on 1)

The method of claim 1, further comprising proactively allocating the computing resources within the memory system for use in execution of future workloads based on the determined one or more weights for hidden layers of the neural network.

### Claim 8 (depends on 1)

The method of claim 1, wherein the respective time intervals comprise four distinct time intervals, and wherein the method further comprises determining the one or more weights for the hidden layers of the neural network for each of the four distinct time intervals.

### Claim 9

An apparatus, comprising:
a memory system comprising a plurality of memory devices; and
a processing unit resident on the memory system and communicatively coupled to each of the memory devices, wherein the processing unit is configured to:
monitor a plurality of resource parameters corresponding to execution of applications involving the memory system;
determine usage patterns associated with the memory system based on the plurality of monitored resource parameters corresponding to execution of the applications;
determine respective time intervals associated with the determined usage patterns, the respective time intervals being associated with particular resource parameters among the plurality of monitored resource parameters;
determine, using at least one monitored resource parameter among the plurality of monitored resource parameters, or the determined usage patterns, or both, one or more weights for hidden layers of a neural network for the respective time intervals associated with the determined usage patterns; and
allocate computing resources within the memory system for use in execution of subsequent applications based on the determined one or more weights corresponding to the hidden layers of the neural network.

### Claim 10 (depends on 9)

The apparatus of claim 9, wherein the plurality of memory devices includes at least one volatile memory device and at least one non-volatile memory device, and wherein the processing unit is further configured to allocate computing resources amongst the at least one volatile memory device and the at least one non-volatile memory device as part of allocation of the computing resources for use in execution of the subsequent applications.

### Claim 11 (depends on 9)

The apparatus of claim 9, wherein the memory system and the processing unit are resident on a mobile computing device.

### Claim 12 (depends on 9)

The apparatus of claim 9, wherein the processing unit is further configured to:
determine individual user application parameters corresponding to applications executed by a user of the memory system;
receive aggregate workload parameters corresponding to applications executed by other users connected to a network to which the memory system is connected; and
determine characteristics of the plurality of resource parameters using the determined individual user application parameters and the received aggregate workload parameters.

### Claim 13 (depends on 9)

The apparatus of claim 9, wherein the processing unit is configured to determine a command latency associated with commands corresponding to execution of the applications to determine the respective time intervals associated with the determined usage patterns.

### Claim 14 (depends on 9)

The apparatus of claim 9, wherein the processing unit is configured to determine at least three distinct time intervals to determine usage patterns associated with the memory system based on the plurality of monitored resource parameters corresponding to execution of the applications, and wherein:
a first distinct time interval corresponds to a current term,
a second distinct term corresponds to a short term, and
a third distinct term corresponds to a medium term or a long term.

### Claim 15 (depends on 9)

The apparatus of claim 9, wherein the neural network is a recurrent neural network.

### Claim 16

A system, comprising:
a telecommunications network comprising a base station and a data center; and
a mobile computing device communicatively coupled to the base station, the mobile computing device comprising a memory system and a processing unit resident thereon, wherein the processing unit is configured to:
monitor a plurality of resource parameters corresponding to execution of workloads running on the mobile computing device;
determine usage patterns associated with the mobile computing device based on the plurality of monitored resource parameters corresponding to execution of the applications;
determine respective autocorrelation time intervals associated with the determined usage patterns, the respective autocorrelation time intervals being determined by the processing unit based on characteristics of particular usage patterns among the determined usage patterns;
determine individual user application parameters corresponding to applications executed by a user of the mobile computing device;
receive, from the data center, aggregate workload parameters corresponding to applications executed by other users connected to the telecommunications network to which the mobile computing device is connected;
determine, using the particular usage patterns among the determined usage patterns, the determined individual user application parameters, or the received aggregate workload parameters, one or more weights for hidden layers of a neural network for the respective autocorrelation time intervals associated with the characteristics of the particular usage pattern; and
allocate computing resources within the mobile computing device for use in execution of subsequent applications based on the determined one or more weights for hidden layers of the neural network.

### Claim 17 (depends on 16)

The system of claim 16, wherein the processing unit is configured to determine a latency associated with initiation commands corresponding to applications executed by a user of the mobile computing device or applications executed by other users connected to the telecommunications network to which the mobile device is connected, or both, to determine the respective time intervals associated with the particular usage patterns.

### Claim 18 (depends on 16)

The system of claim 16, wherein the processing unit is configured to set at least four distinct time intervals to determine usage patterns associated with the mobile computing device based on the plurality of monitored resource parameters corresponding to execution of the applications, and wherein:
a first distinct time interval corresponds to a current term time interval,
a second distinct time interval term corresponds to a short term time interval,
a third distinct time interval corresponds to a medium term time interval, and
a fourth distinct time interval corresponds to a long term time interval.

### Claim 19 (depends on 18)

The system of claim 18, wherein:
the first distinct time interval corresponds to applications executed within a first time period from receipt of an application initiation indicator,
the second distinct time interval corresponds to applications executed within a second time period from receipt of an application initiation indicator,
the third distinct time interval corresponds to applications executed within a third time period from receipt of an application initiation indicator, and
the fourth distinct time interval corresponds to applications executed within a fourth time period from receipt of an application initiation indicator.

### Claim 20 (depends on 16)

The system of claim 16, wherein the processing unit is configured to allocate the computing resources within the mobile computing device prior to execution of a subsequent application based on the determined one or more weights for hidden layers of the neural network.
