Abstract: A method and apparatus are provided. The method includes configuring a plurality of long short term memory (LSTM) networks, wherein each of the plurality of LSTM networks is at a different network layer, configuring a plurality of memory cells in a spatial domain of the plurality of LSTM networks, configuring the plurality of memory cells in a temporal domain of the plurality of LSTM networks, controlling an output of each of the plurality of LSTM networks based on highway connections to outputs from at least one previous layer and at least one previous time of the plurality of LSTM networks, and controlling the plurality of memory cells based on highway connections to memory cells from the at least one previous time.

Abstract: Apparatuses and methods of manufacturing same, systems, and methods for training deep learning machines are described. In one aspect, candidate units, such as detection bounding boxes in images or phones of an input audio feature, are classified using soft labelling, where at least label has a range of possible values between 0 and 1 based, in the case of images, on the overlap of a detection bounding box and one or more ground-truth bounding boxes for one or more classes.

Abstract: Apparatuses and methods of manufacturing same, systems, and methods for object detection using a region-based deep learning model are described. In one aspect, a method is provided, in which a region proposal network (RPN) is used to identify regions of interest (RoI) in an image by assigning a confidence levels, the assigned confidence levels of the RoIs are used to boost the background score assigned by the downstream classifier to each RoI, and the background scores are used in a softmax function to calculate the final class probabilities for each object class.

Abstract: An apparatus and a method. The apparatus includes a plurality of polarization processors, including n inputs and n outputs, where n is an integer; and at least one permutation processor, including n inputs and n outputs, wherein each of the at least one permutation processor is connected between two of the plurality of polarization processors, and connects the n outputs of a first of the two of the plurality of polarizations processors to the n inputs of a second of the two of the plurality of polarization processors between which each of the at least one permutation processor is connected in a permutation pattern that maximally polarizes the n outputs of the second of the two of the plurality of polarization processors.

Abstract: A system, method and device for wireless communication is provided. The method includes receiving, by a receiver, data from a wireless transmitter, receiving, by the receiver, retransmission of the data from the wireless transmitter, storing, by the receiver, the data, processing, by the receiver, the data without knowledge of the data length, and determining, by the receiver, a data length of the data.

Abstract: A communication system includes: a validation module configured to transmit a repeat request corresponding to a preceding data including a communication content; an inter-block processing module, coupled to the validation module, configured to determine a previous communication value based on the preceding data; a detection module, coupled to the inter-block processing module, configured to identify a repeat data corresponding to the repeat request from a receiver signal; an accumulator module, coupled to the detection module, configured to generate an accumulation output based on the preceding data and the repeat data; and a decoding module, coupled to the accumulator module, configured to determine the communication content using the previous communication value and the accumulation output across instances of transmission blocks for communicating with a device.

Abstract: Apparatuses and methods of manufacturing same, systems, and methods for performing network parameter quantization in deep neural networks are described. In one aspect, diagonals of a second-order partial derivative matrix (a Hessian matrix) of a loss function of network parameters of a neural network are determined and then used to weight (Hessian-weighting) the network parameters as part of quantizing the network parameters. In another aspect, the neural network is trained using first and second moment estimates of gradients of the network parameters and then the second moment estimates are used to weight the network parameters as part of quantizing the network parameters. In yet another aspect, network parameter quantization is performed by using an entropy-constrained scalar quantization (ECSQ) iterative algorithm. In yet another aspect, network parameter quantization is performed by quantizing the network parameters of all layers of a deep neural network together at once.

Abstract: Apparatuses and methods of manufacturing same, systems, and methods for performing network parameter quantization in deep neural networks are described. In one aspect, multi-dimensional vectors representing network parameters are constructed from a trained neural network model. The multi-dimensional vectors are quantized to obtain shared quantized vectors as cluster centers, which are fine-tuned. The fine-tuned and shared quantized vectors/cluster centers are then encoded. Decoding reverses the process.

Abstract: A method and apparatus are provided. The method includes receiving an image, detecting an object in the image, determining, by a primary object detector, a primary confidence detection score of the object, determining, by a classification network, a confidence scaling factor of the object, and adjusting the primary confidence detection score based on multiplying the primary confidence detection score by the confidence scaling factor.

Abstract: A system, method and device for object identification is provided. The method of identifying objects includes, but is not limited to, calculating feature vectors of the object, calculating feature vectors of the object's context and surroundings, combining feature vectors of the object, calculating likelihood metrics of combined feature vectors, calculating verification likelihood metrics against contact list entries, calculating a joint verification likelihood metric using the verification likelihood metrics, and identifying the object based on the joint verification likelihood metric.

Abstract: An apparatus and a method. The apparatus includes a receiver to receive a polar codeword of length mj; a processor configured to determine a decoding node tree structure with mj leaf nodes for the received codeword, and receive i indicating a level at which parallelism of order m is applied to the decoding node tree structure, wherein i indicates levels of the decoding node tree structure, and wherein the mj leaf nodes are at level j; and m successive cancellation list decoders (SCLDs) applied to each child node of each node in the decoding node tree structure at level i?1, wherein each of the m SCLDs executes in parallel to determine log likelihood ratios (LLRs) for a codeword of length mj?i, and wherein each of the m SCLDs uses LLRs of an associated parent node without using a hard decision or a soft reliability estimate of any other node of the other m SCLDs.

Abstract: An apparatus and a method. The apparatus includes a first recurrent network in a first layer; a second recurrent network in a second layer connected to the first recurrent network; a distant input gate connected to the second recurrent network; a first highway gate connected to the distant input gate and the second recurrent network; a first elementwise product projection gate connected to the distant input gate, the highway gate, and the second recurrent network; a second highway gate connected to the first recurrent network and the second recurrent network; and a second elementwise product projection gate connected to the first recurrent network, the second highway gate, and the second recurrent network.

Abstract: An apparatus and a method. The apparatus includes a plurality of long short term memory (LSTM) networks, wherein each of the plurality of LSTM networks is at a different network layer, wherein each of the plurality of LSTM networks is configured to determine a residual function, wherein each of the plurality of LSTM networks includes an output gate to control what is provided to a subsequent LSTM network, and wherein each of the plurality of LSTM networks includes at least one highway connection to compensate for the residual function of a previous LSTM network.

Abstract: A system to recognize objects in an image includes an object detection network outputs a first hierarchical-calculated feature for a detected object. A face alignment regression network determines a regression loss for alignment parameters based on the first hierarchical-calculated feature. A detection box regression network determines a regression loss for detected boxes based on the first hierarchical-calculated feature. The object detection network further includes a weighted loss generator to generate a weighted loss for the first hierarchical-calculated feature, the regression loss for the alignment parameters and the regression loss of the detected boxes. A backpropagator backpropagates the generated weighted loss.

Abstract: A method, apparatus, and chipset are provided for constructing hybrid automatic repeat request (HARQ) rate-compatible polar codes for communication channels. The method includes constructing, in a terminal, a base polar code of length 2n; and determining a sequence of m<2n bits to puncture in the base polar code by testing a predetermined criterion at most (22n+2n)/2?1 times.

Abstract: A system, method and device for wireless communication is provided. The method includes receiving, by a receiver, data from a wireless transmitter, receiving, by the receiver, retransmission of the data from the wireless transmitter, storing, by the receiver, the data, processing, by the receiver, the data without knowledge of the data length, and determining, by the receiver, a data length of the data.

Abstract: A system, method and device for object identification is provided. The method of identifying objects includes, but is not limited to, calculating feature vectors of the object, calculating feature vectors of the object's context and surroundings, combining feature vectors of the object, calculating likelihood metrics of combined feature vectors, calculating verification likelihood metrics against contact list entries, calculating a joint verification likelihood metric using the verification likelihood metrics, and identifying the object based on the joint verification likelihood metric.

Abstract: A computing system includes: an inter-device interface configured to communicate content; and a communication unit, coupled to the inter-device interface, configured to process the content based on a polar communication mechanism utilizing multiple processing dimensions for communicating the content, including: generating a node result with a first orthogonal mechanism, and processing the node result from the first orthogonal mechanism with a second orthogonal mechanism.

Abstract: A computing system includes: a communication unit configured to: determine a relaxed coding profile including a polar-processing range for processing content data over a bit channel; process the content data based on a total polarization level being within the polar-processing range, the polar-processing range for controlling a polar processing mechanism or a portion therein corresponding to the bit channel for the content data; and an inter-device interface, coupled to the communication unit, configured to communicate the content data.

Abstract: An apparatus and method of constructing a universal polar code is provided. The apparatus includes a first function block configured to polarize and degrade a class of channels Wj to determine a probability of error Pe,j of each bit-channel of Wj, wherein j?{1, 2, . . . , s}, in accordance with a bit-channel index i; a second function block configured to determine a probability of error Pe(i) for the universal polar code for each bit-channel index i; a third function block configured to sort the Pe(i); and a fourth function block configured to determine a largest number k of bit-channels such that a sum of corresponding k bit-channel error probabilities Pe(i) is less than or equal to a target frame error rate Pt for the universal polar code, wherein the indices corresponding to the k smallest Pe(i) are good bit-channels for the universal polar code.