# US 12,762,373 B2 · System and method for adapting to changing constraints > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,762,373 B2 - Document ID: US12762373B2 - Status: Granted - Kind: B2 - Application number: 17/911,362 - Filed: 2021-03-12 - Published: 2023-03-23 - Granted: 2026-10-13 - Earliest claimed: 2020-03-17 - Assignees: InterDigital CE Patent Holdings, SAS - Applicants: InterDigital CE Patent Holdings - Inventors: Francois Schnitzler; Anne Lambert; Francoise Le Bolzer - Web page: https://patentgrep.com/patent/US12762373B2 ## CPC - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/045: Combinations of networks - G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections - G06N 3/084: Backpropagation, e.g. using gradient descent - G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract In general, at least one example of an embodiment can involve selecting a neural network from a plurality of neural networks based on an indication of resource availability and processing data using the selected neural network in accordance with the resource availability. ## Claims ### Claim 1 A method performed by a wireless transmit/receive unit (WTRU), the method comprising: receiving an input data sequence; receiving a first indication of a first constraint for processing a first portion of the input data sequence at a first time by a first neural network, wherein the first indication indicates a relationship between the first constraint and a characteristic of the first neural network for processing the first portion of the input data sequence; receiving a second indication of a second constraint corresponding to a change in the first constraint for processing a second portion of the input data sequence at a second time by a second neural network that differs from the first neural network, wherein the second indication indicates a relationship between the second constraint and a characteristic of the second neural network for processing the second portion of the input data sequence; determining, via the first neural network, a state for a first input element included in the first portion of the input data sequence based on the first input element and based on a previous state for a previous input element included in the first portion of the input data sequence; determining, via the second neural network, a state for a second input element included in the second portion of the input data sequence based on the second input element and based on the state for the first input element; and processing, via the second neural network based on the state for the second input element, the second portion of the input data sequence based on the second constraint. ### Claim 2 (depends on 1) The method of claim 1, wherein: the characteristic of the first neural network comprises at least one of a first computation cost or a first accuracy associated with the first neural network; and the characteristic of the second neural network comprises at least one of a second computation cost or a second accuracy associated with the second neural network. ### Claim 3 (depends on 1) The method of claim 1, wherein the first constraint comprises at least one of a computational resource availability or a data processing accuracy. ### Claim 4 (depends on 1) The method of claim 1, wherein the first neural network has a greater computational load than the second neural network, and wherein the first indication indicates a greater computational resource availability than the second indication. ### Claim 5 (depends on 4) The method of claim 4, further comprising: transmitting, to a device other than the WTRU, at least one value indicating a difference in accuracy associated with the first neural network and the second neural network; or transmitting, to the device other than the WTRU, an expected delay associated with switching between using the first neural network and the second neural network for processing the input data sequence. ### Claim 6 (depends on 5) The method of claim 5, wherein the first neural network and the second neural network are included in a family of neural networks comprising at least one additional neural network, wherein each neural network in the family of neural networks is associated with a different computational load and a different accuracy, and wherein at least one of the computational load or accuracy associated with each neural network in the family of neural networks is transmitted to the device other than the WTRU. ### Claim 7 (depends on 6) The method of claim 6, wherein the family of neural networks is communicated in a package indicating available neural networks at the WTRU for processing the input data sequence, and wherein the package includes metadata that indicates the at least one of the computational load or the accuracy associated with each neural network in the family of neural networks. ### Claim 8 (depends on 4) The method of claim 4, wherein the first neural network comprises a first skip recurrent neural network (RNN) model, wherein the second neural network comprises a second skip RNN model, wherein the second skip RNN model has a lower computational load when processing the second portion of the input data sequence than a computational load of the first skip RNN model when processing the first portion of the input data sequence. ### Claim 9 (depends on 4) The method of claim 4, wherein the second neural network is adapted from the first neural network to enable processing of the second portion of the input data sequence with a lower computational load, and wherein the second neural network is configured to minimize a loss in accuracy from the first neural network. ### Claim 10 (depends on 1) The method of claim 1, wherein the input data sequence comprises video data or audio data, and wherein the processing of the first portion or the processing of the second portion is performed using an encoder or a decoder on the WTRU. ### Claim 11 (depends on 1) The method of claim 1, further comprising, prior to receiving the input data sequence: receiving a training sequence of input elements; determining, via the first neural network, a first sequence of respective states for the training sequence of input elements; determining, via the second neural network, a second sequence of respective states for the training sequence of input elements; and training the second neural network by reducing a global loss, wherein the global loss is based at least on a state difference loss between the second sequence of respective states and the first sequence of respective states. ### Claim 12 A wireless transmit receive unit (WTRU) comprising a processor, the processor configured to: receive an input data sequence; receive a first indication of a first constraint to process a first portion of the input data sequence at a first time by a first neural network, wherein the first indication indicates a relationship between the current constraint and a characteristic of the first neural network configured to process the first portion of the input data sequence; while processing the first portion of the input data sequence by the first neural network, receive a second indication of a second constraint corresponding to a change in the first constraint to process a second portion of the input data sequence at a second time by a second neural network that differs from the first neural network, wherein the second indication indicates a relationship between the second constraint and a characteristic of the second neural network configured to process the second portion of the input data sequence; determine, via the first neural network, a state for a first input element included in the first portion of the input data sequence based on the first input element and based on a previous state for a previous input element included in the first portion of the input data sequence; determine, via the second neural network, a state for a second input element included in the second portion of the input data sequence based on the second input element and based on the state for the first input element; and process, via the second neural network based on the state for the second input element, the second portion of the input data sequence based on the second constraint. ### Claim 13 (depends on 12) The WTRU of claim 12, wherein: the characteristic of the first neural network comprises at least one of a first computation cost or a first accuracy associated with the first neural network; and the characteristic of the second neural network comprises at least one of a second computation cost or a second accuracy associated with the second neural network. ### Claim 14 (depends on 12) The WTRU of claim 12, wherein the first constraint comprises at least one of a computational resource availability or a data processing accuracy. ### Claim 15 (depends on 12) The WTRU of claim 12, wherein the first neural network has a greater computational load than the second neural network, and wherein the first indication indicates a greater computational resource availability than the second indication. ### Claim 16 (depends on 15) The WTRU of claim 15, further comprising a transceiver, and wherein the processor is further configured to: transmit, via the transceiver to a device other than the WTRU, at least one value indicating a difference in accuracy associated with the first neural network and the second neural network; or transmit, via the transceiver to the device other than the WTRU, an expected delay associated with switching between using the first neural network and the second neural network for processing the input data sequence. ### Claim 17 (depends on 16) The WTRU of claim 16, wherein the first neural network and the second neural network are included in a family of neural networks comprising at least one additional neural network, wherein each neural network in the family of neural networks is associated with a different computational load and a different accuracy, and wherein the processor is further configured to transmit, via the transceiver, at least one of the computational load or accuracy associated with each neural network in the family of neural networks to the device other than the WTRU. ### Claim 18 (depends on 17) The WTRU of claim 17, wherein the processor is configured to communicate, via the transceiver, the family of neural networks in a package indicating available neural networks at the WTRU for processing the input data sequence, and wherein the package includes metadata that indicates the at least one of the computational load or the accuracy associated with each neural network in the family of neural networks. ### Claim 19 (depends on 15) The WTRU of claim 15, wherein the first neural network comprises a first skip recurrent neural network (RNN) model, wherein the second neural network comprises a second skip RNN model, wherein the second skip RNN model has a lower computational load when processing the second portion of the input data sequence than a computational load of the first skip RNN model when processing the first portion of the input data sequence. ### Claim 20 (depends on 15) The WTRU of claim 15, wherein the second neural network is adapted from the first neural network to enable the processor to process the second portion of the input data sequence with a lower computational load, and wherein the second neural network is configured to minimize a loss in accuracy from the first neural network. ### Claim 21 (depends on 12) The WTRU of claim 12, wherein the input data sequence comprises video data or audio data, and wherein the processor is configured to process the first portion or the second portion using an encoder or a decoder on the WTRU. ### Claim 22 (depends on 12) The WTRU of claim 12, wherein the processor is further configured to, prior to receiving the input data sequence: receive a training sequence of input elements; determine, via the first neural network, a first sequence of respective states for the training sequence of input elements; determine, via the second neural network, a second sequence of respective states for the training sequence of input elements; and train the second neural network by reducing a global loss, wherein the global loss is based at least on a state difference loss between the second sequence of respective states and the first sequence of respective states.