A processing system including at least one processor may obtain operational data from a radio access network (RAN), format the operational data into state information and reward information for a reinforcement learning agent (RLA), processing the state information and the reward information via the RLA, where the RLA comprises a plurality of sub-agents, each comprising a respective neural network, each of the neural networks encoding a respective policy for selecting at least one setting of at least one parameter of the RAN to increase a respective predicted reward in accordance with the state information, and where each neural network is updated in accordance with the reward information. The processing system may further determine settings for parameters of the RAN via the RLA, where the RLA determines the settings in accordance with selections for the settings via the plurality of sub-agents, and apply the plurality of settings to the RAN.

Systems and methods are provided for implementing a self-adaptive batch dataset partitioning control process which is utilized in conjunction with a distributed deep learning model training process to optimize load balancing among a set of accelerator resources. An iterative batch size tuning process is configured to determine an optimal job partition ratio for partitioning mini-batch datasets into sub-batch datasets for processing by a set of hybrid accelerator resources, wherein the sub-batch datasets are partitioned into optimal batch sizes for processing by respective accelerator resources to minimize a time for completing the deep learning model training process.

The disclosure describes various aspects of optical control of atomic quantum bits (qubits) for phase control operations. More specifically, the disclosure describes methods for coherently controlling quantum phases on atomic qubits mediated by optical control fields, applying to quantum logic gates, and generalized interactions between qubits. Various attributes and settings of optical/qubit interactions (e.g., atomic energy structure, laser beam geometry, polarization, spectrum, phase, background magnetic field) are identified for imprinting and storing phase in qubits. The disclosure further describes how these control attributes are best matched in order to control and stabilize qubit interactions and allow extended phase-stable quantum gate sequences.

The disclosure describes various aspects of optical control of atomic quantum bits (qubits) for phase control operations. More specifically, the disclosure describes methods for coherently controlling quantum phases on atomic qubits mediated by optical control fields, applying to quantum logic gates, and generalized interactions between qubits. Various attributes and settings of optical/qubit interactions (e.g., atomic energy structure, laser beam geometry, polarization, spectrum, phase, background magnetic field) are identified for imprinting and storing phase in qubits. The disclosure further describes how these control attributes are best matched in order to control and stabilize qubit interactions and allow extended phase-stable quantum gate sequences.

There is provided an apparatus comprising at least one processing core, at least one memory including computer program code, at least one memory and the computer program code being configured to, with at least one processing core, cause the apparatus at least to obtain, from a first sequential input (102), a first output from a first recurrent neural network, the first sequential input being of a first modality, obtain, from a second sequential input (103), a second output from a second recurrent neural network, the second sequential input being of a second modality, and process the first output and the second output to obtain a correlation of the first and second sequential inputs.

This present invention relates to a system for processing distributed intelligent information, and discloses a method that when there are no actions beyond the threshold value among the analyzed actions, a virtual global workspace (GW) is formed, the global workspace including one or more thinking devices, and the thinking device operates as a first type device or a second type device, in which an action alternative is requested for the global workspace in the case of operating as the first type device, and an action solution is proposed for the global workspace in the case of operating as the second type device.

Computer-implemented techniques for learning composite machine learned models are disclosed. Benefits to implementors of the disclosed techniques include allowing non-machine learning experts to use the techniques for learning a composite machine learned model based on a learning dataset, reducing or eliminating the explorative trial and error process of manually tuning architectural parameters and hyperparameters, and reducing the computing resource requirements and model learning time for learning composite machine learned models. The techniques improve the operation of distributed learning computing systems by reducing or eliminating straggler effects and by reducing or minimizing synchronization latency when executing a composite model search algorithm for learning a composite machine learned model.

The present disclosure relates to an interconnect system for interfacing an electronic subsystem to a qubit package, wherein the qubit package has a plurality of independent qubits. The system makes use of an optical fiber cable having a plurality of optical fibers, which is interfaced to the electronic subsystem. A 3D optical structure is used which has a plurality of internal waveguides, and which is configured to interface the optical fiber cable to the qubit package. The 3D optical structure further has at least one subsystem for using the plurality of waveguides to receive signals of a first type from at least one of the qubits package or the optical fiber cable, to convert the signals from the first type to a second type, and to transmit the signals in the second type to the other one of the fiber optic cable or the qubit package.

The present disclosure relates to an interconnect system for interfacing an electronic subsystem to a qubit package, wherein the qubit package has a plurality of independent qubits. The system makes use of an optical fiber cable having a plurality of optical fibers, which is interfaced to the electronic subsystem. A 3D optical structure is used which has a plurality of internal waveguides, and which is configured to interface the optical fiber cable to the qubit package. The 3D optical structure further has at least one subsystem for using the plurality of waveguides to receive signals of a first type from at least one of the qubits package or the optical fiber cable, to convert the signals from the first type to a second type, and to transmit the signals in the second type to the other one of the fiber optic cable or the qubit package.

Methods and apparatus for training a matrix-based differentiable program using a photonics-based processor. The matrix-based differentiable program includes at least one matrix-valued variable associated with a matrix of values in a Euclidean vector space. The method comprises configuring components of the photonics-based processor to represent the matrix of values as an angular representation, processing, using the components of the photonics-based processor, training data to compute an error vector, determining in parallel, at least some gradients of parameters of the angular representation, wherein the determining is based on the error vector and a current input training vector, and updating the matrix of values by updating the angular representation based on the determined gradients.