Abstract: In some embodiments, a system includes an artificial intelligence (AI) chip and a processor coupled to the AI chip and configured to receive an input image, crop the input image into a plurality of cropped images, and execute the AI chip to produce a plurality of feature maps based on at least a subset of the plurality of cropped images. The system may further merge at least a subset of the plurality of feature maps to form a merged feature map, and produce an output image based on the merged feature map. The cropping and merging operations may be performed according to a same pattern. The system may also include a training network configured to train weights of the CNN model in the AI chip in a gradient descent network. Cropping and merging may be performed over the training sample images in the training work in a similar manner.

Abstract: A system for structuring a directed energy beam includes one or more coherent light sources that emit one or more initial light beams, one or more spatial light modulators that modulate the one or more initial light beams, and a beam combiner that coherently adds orbital angular momentum beams to create a reconfigurable spatial region of localized power that forms the directed energy beam. Each spatial light modulator is loaded with a pattern that receives an incident light beam and outputs an orbital angular momentum beam. The pattern encodes one or more orthogonal orbital angular momentum functions. Characteristically, each orbital angular momentum having an associated complex weight with which each orbital angular momentum beam is weighted in forming the coherent addition.

Abstract: An integrated circuit may include multiple cellular neural networks (CNN) processing engines coupled in a loop circuit and configured to perform an AI task. Each CNN processing engine includes multiple convolution layers, a first memory buffer to store imagery data and a second memory buffer to store filter coefficients. The CNN processing engines are configured to perform convolution operations over an input image simultaneously in one or more iterations. In each iteration, various sub-images of the input image are loaded to the first memory buffer circularly. A portion of the filter coefficients corresponding to the sub-image are loaded to the second memory buffer in a cyclic order. Data may be arranged in the second memory buffer to facilitate loading of duplicate filter coefficients among at least two convolution layers without requiring duplicate memory space. Methods of training a CNN model having duplicate weights are also provided.

Abstract: An integrated circuit may include multiple cellular neural networks (CNN) processing engines coupled to at least one input/output data bus and a clock-skew circuit in a loop circuit. Each CNN processing engine includes multiple convolution layers, a first memory buffer to store imagery data and a second memory buffer to store filter coefficients. Each of the CNN processing engines is configured to perform convolution operations over an input image simultaneously in a first clock cycle to generate output to be fed to an immediate neighbor CNN processing engine for performing convolution operations in a next clock cycle. The second memory buffer may store a first subset of filter coefficients for a first convolution layer of the CNN processing engine and store a reference location to the first subset of filter coefficients for a second convolution layer, where the filter coefficients for the first and second convolution layers are duplicate.

Abstract: A system for training an artificial intelligence (AI) model for an AI chip may include a forward network and a backward propagation network. The AI model may be a convolution neural network (CNN). The forward network may infer the output of the AI chip based on the training data. The backward network may use the output of the AI chip and the ground truth data to train the weights of the AI model. In some examples, the system may train the AI model using a gradient descent method. The system may quantize the weights and update the weights during the training. In some examples, the system may perform a uniform quantization over the weights. The system may also determine the distribution of the weights. If the weight distribution is not symmetric, the system may group the weights and quantize the weights based on the grouping.

Abstract: A system for training an artificial intelligence (AI) model for an AI chip may include an AI training unit to train weights of an AI model in floating point, and one or more quantization units for updating the weights of the AI model while accounting for the hardware constraints in the AI chip. The system may also include customization unit for performing one or more linear transformations on the updated weights. The system may also perform output equalization for one or more convolution layers of the AI model to equalize the inputs and/or outputs of each layer of the AI model to within the range allowed in the physical AI chip. The system may further update the weights by performing shift-based quantization that mimics the characteristics of a hardware chip. The updated weights may be stored in fixed point and uploadable to an AI chip implementing an AI task.

Abstract: An artificial intelligence (AI) semiconductor having an embedded convolution neural network (CNN) may include a first convolution layer and a second convolution layer, in which the weights of the first layer and the weights of the second layer are quantized in different bit-widths, thus at different compression ratios. In a VGG neural network, the weights of a first group of convolution layers may have a different compression ratio than the weights of a second group of convolution layers. The weights of the CNN may be obtained in a training system including convolution quantization and/or activation quantization. Depending on the compression ratio, the weights of a convolution layer may be trained with or without re-training. An AI task, such as image retrieval, may be implemented in the AI semiconductor having the CNN described above.

Abstract: A system for training an artificial intelligence (AI) model for an AI chip to implement an AI task may include an AI training unit to train weights of an AI model in floating point, a convolution quantization unit for quantizing the trained weights to a number of quantization levels, and an activation quantization unit for updating the weights of the AI model so that output of the AI model based at least on the updated weights are within a range of activation layers of the AI chip. The updated weights may be stored in fixed point and uploadable to the AI chip. The various units may be configured to account for the hardware constraints in the AI chip to minimize performance degradation when the trained weights are uploaded to the AI chip and expedite training convergence. Forward propagation and backward propagation may be combined in training the AI model.

Abstract: Methods, systems, and devices for data encoding and channel hopping. The system includes a signal source for providing a signal. The system includes an optical switch having an input port and multiple output paths. The optical switch is configured to receive, at the input port, the signal. The optical switch is configured to route the signal to an output path of the multiple output paths. The system includes a mode converter that is connected to the optical switch and configured to select an orbital angular momentum (OAM) mode. The mode converter is configured to encode or channel hop the signal using the OAM mode and combine the signal from each output path. The system includes a transmitter configured to propagate the signal.

Abstract: In some examples, given an AI model in floating point, a system may use one or more artificial intelligence (AI) chips to train a global gain vector for use to convert the AI model in floating point to an AI model in fixed point for uploading to a physical AI chip. The system may determine initial gain vectors, and in each of multiple iterations, obtain the performance values of the AI chips based on the gain vectors and update the gam vectors for the next iteration. The gain vectors are updated based on a velocity of gain. The performance value may be based on feature maps of an AI model before and after the converting. The performance value may also be based on interference over a test dataset. Upon completion of the iterations, the system determines the global gain vector that resulted in the best performance value during the iterations.

Abstract: A system may include multiple client devices and a processing device communicatively coupled to the client devices. Each client device includes an artificial intelligence (AI) chip and is configured to generate an AI model. The processing device may be configured to (i) receive a respective AI model and an associated performance value of the respective AI model from each of the plurality of client devices; (ii) determine an optimal AI model based on the performance values associated with the respective AI models from the plurality of client devices; and (iii) determine a global AI model based on the optimal AI model. The system may load the global AI model into an AI chip of a client device to cause the client device to perform an AI task based on the global AI model in the AI chip. The AI model may include a convolutional neural network.

Abstract: A device for obtaining a local optimal AI model may include an artificial intelligence (AI) chip and a processing device configured to receive a first initial AI model from the host device. The device may load the initial AI model into the AI chip to determine a performance value of the AI model based on a dataset, and determine a probability that a current AI model should be replaced by the initial AI model. The device may determine, based on the probability, whether to replace the current AI model with the initial AI model. If it is determined that the current AI model be replaced, the device may replace the current AI model with the initial AI model. The device may repeat the above processes and obtain a final current AI model. The device may transmit the final current AI model to the host device.

Abstract: A system may be configured to obtain a global artificial intelligence (AI) model for uploading into an AI chip to perform AI tasks. The system may implement a training process including receiving updated AI models from one or more client devices, determining a global AI model based on the received AI models from the client devices, and updating initial AI models for the client devices. Each client device may receive an initial AI model and train an updated AI model by training the entire parameters of the AI model together, by training a subset of the parameters of the AI model in a layer by layer fashion, or by training a subset of the parameters by parameter types. Each client device may include one or more AI chips configured to run an AI task to measure performance of an AI model. The AI model may include a convolutional neural network.

Abstract: A system includes a transmitter with multiple transmit devices each having an OAM multiplexer that converts multiple input signals into an OAM beam. Each transmit device outputs a coaxial group of orthogonal OAM beams. The system also includes a receiver that has multiple receive devices each having an OAM demultiplexer that receives the group of OAM beams from a corresponding transmit device. The OAM demultiplexer also converts the coaxial group of mutually orthogonal OAM beams into a plurality of received signals corresponding to input signals represented by the OAM beams. The receiver also includes a MIMO processor that has an equalizer that determines a transfer function corresponding to crosstalk of each of the plurality of received signals. The MIMO processor also reduces the crosstalk of each of the plurality of received signals based on the transfer function and updates the transfer function.

Abstract: A system includes a transmitter having a first transmit device having a first transmit antenna and a first OAM multiplexer designed to receive two input signals and to convert the input signals to orthogonal OAM beams. The first transmit antenna is designed to transmit a first output signal that includes the OAM beams. The transmitter also includes a second transmit device that functions in a similar manner as the first transmit device. A receiver includes a first receive device having a first receive antenna designed to receive the first output signal and a first OAM demultiplexer designed to convert the first output signal to received signals corresponding to the input signals. The receiver also includes a second receive device having similar features as the first receive device. The receiver also includes a MIMO processor designed to reduce interference between the received signals.

Abstract: A system for structuring a directed energy beam includes one or more coherent light sources that emit one or more initial light beams, one or more spatial light modulators that modulate the one or more initial light beams, and a beam combiner that coherently adds orbital angular momentum beams to create a reconfigurable spatial region of localized power that forms the directed energy beam. Each spatial light modulator is loaded with a pattern that receives an incident light beam and outputs an orbital angular momentum beam. The pattern encodes one or more orthogonal orbital angular momentum functions. Characteristically, each orbital angular momentum having an associated complex weight with which each orbital angular momentum beam is weighted in forming the coherent addition.

Abstract: Methods, systems, and devices for data encoding and channel hopping. The system includes a signal source for providing a signal. The system includes an optical switch having an input port and multiple output paths. The optical switch is configured to receive, at the input port, the signal. The optical switch is configured to route the signal to an output path of the multiple output paths. The system includes a mode converter that is connected to the optical switch and configured to select an orbital angular momentum (OAM) mode. The mode converter is configured to encode or channel hop the signal using the OAM mode and combine the signal from each output path. The system includes a transmitter configured to propagate the signal.

Abstract: A system includes a transmitter with multiple transmit devices each having an OAM multiplexer that converts multiple input signals into an OAM beam. Each transmit device outputs a coaxial group of orthogonal OAM beams. The system also includes a receiver that has multiple receive devices each having an OAM demultiplexer that receives the group of OAM beams from a corresponding transmit device. The OAM demultiplexer also converts the coaxial group of mutually orthogonal OAM beams into a plurality of received signals corresponding to input signals represented by the OAM beams. The receiver also includes a MIMO processor that has an equalizer that determines a transfer function corresponding to crosstalk of each of the plurality of received signals. The MIMO processor also reduces the crosstalk of each of the plurality of received signals based on the transfer function and updates the transfer function.

Abstract: An adaptive optics compensation approach for an OAM multiplexed FSO communication system is described, in which a Gaussian beam is used to probe the turbulence-induced wavefront distortions and derive the correction pattern for compensating the OAM beams. Using this approach, we demonstrate simultaneous compensation of multiple OAM beams each carrying a 100-Gbit/s data channel through emulated atmospheric turbulence. The results indicate that the turbulence-induced crosstalk and power penalty could be efficiently mitigated by ˜12.5 dB and ˜11 dB respectively.

Abstract: In at least one aspect, a device for Orbital Angular Momentum (OAM) based optical communication includes a first spatial light modulator configured to down-convert a first plurality of higher-order OAM modes from a communication signal to a second plurality of higher-order OAM modes and a first Gaussian mode, a second spatial light modulator configured to drop the first Gaussian mode and add a second Gaussian mode to the second plurality of higher-order OAM modes, and a third spatial light modulator configured to up-convert the second plurality of higher-order OAM modes and the second Gaussian mode to a third plurality of higher-order OAM modes for further communications.