Abstract: Various operations may be performed based on a distance-related function associated with two or more devices. For example, an association procedure for two or more devices may be based on one or more determined distances. Similarly, presence management may be based on one or more determined distances. A distance-related function may take various form including, for example, a distance between devices, two or more distances between devices, a rate of change in a relative distance between devices, relative acceleration between devices, or some combination of two or more of the these distance-related functions.

Abstract: A method of quantizing a floating point machine learning network to obtain a fixed point machine learning network using a quantizer may include selecting at least one moment of an input distribution of the floating point machine learning network. The method may also include determining quantizer parameters for quantizing values of the floating point machine learning network based at least in part on the at least one selected moment of the input distribution of the floating point machine learning network to obtain corresponding values of the fixed point machine learning network.

Abstract: A method for selecting bit widths for a fixed point machine learning model includes evaluating a sensitivity of model accuracy to bit widths at each computational stage of the model. The method also includes selecting a bit width for parameters, and/or intermediate calculations in the computational stages of the mode. The bit width for the parameters and the bit width for the intermediate calculations may be different. The selected bit width may be determined based on the sensitivity evaluation.

Abstract: A method of adaptively selecting a configuration for a machine learning process includes determining current system resources and performance specifications of a current system. A new configuration for the machine learning process is determined based at least in part on the current system resources and the performance specifications. The method also includes dynamically selecting between a current configuration and the new configuration based at least in part on the current system resources and the performance specifications.

Abstract: A method of online training of a classifier includes determining a distance from one or more feature vectors of an object to a first predetermined decision boundary established during off-line training for the classifier. The method also includes updating a decision rule as a function of the distance. The method further includes classifying a future example based on the updated decision rule.

Abstract: An apparatus for data processing, according to one or more aspects of the disclosure, includes a processing system configured to communicate with at least one of a plurality of reference nodes worn on body parts to obtain body positioning data relating to relative position between the body parts, and provide body tracking based on the body positioning data. The body positioning data relates to ranging and/or angular position between each of the reference nodes and a reference plane defined by one or more of the reference nodes.

Abstract: Hyper-parameters are selected for training a deep convolutional network by selecting a number of network architectures as part of a database. Each of the network architectures includes one or more local logistic regression layer and is trained to generate a corresponding validation error that is stored in the database. A threshold error for identifying a good set of network architectures and a bad set of network architectures may be estimated based on validation errors in the database. The method also includes choosing a next potential hyper-parameter, corresponding to a next network architecture, based on a metric that is a function of the good set of network architectures. The method further includes selecting a network architecture, from among next network architectures, with a lowest validation error.

Abstract: Compressing a machine learning network, such as a neural network, includes replacing one layer in the neural network with compressed layers to produce the compressed network. The compressed network may be fine-tuned by updating weight values in the compressed layer(s).

Abstract: Apparatuses and methods that provide for implementing a body area network are disclosed. An example of a method described herein includes determining a plurality of paths from a source node to a destination node in a set of nodes; determining, for each path in the plurality of paths, a Path Figure of Merit (PFM); associating two or more paths in the plurality of paths with a connection based on a total PFM threshold for the connection; and communicating from the source node to the destination node using at least two of the paths associated with the connection. Apparatuses for implementing the methods are also disclosed herein.

Abstract: A method for classifying an object includes applying multiple confidence values to multiple objects. The method also includes determining a metric based on the multiple confidence values. The method further includes determining a classification of a first object from the multiple objects based on a knowledge-graph when the metric is above a threshold.

Abstract: A system for detecting anatomical gestures based on the interpretation of radio signal transmissions. A user may place wireless devices on his/her head and wrist(s) that communicate through short-range radio signals. The wireless devices may collect information regarding signal exchanges which may be analyzed by a computing device to determine positional information about the user's hands. The computing device may compile the positional information into sequences and evaluate the sequences against predefined patterns of movement data. The computing device may interpret recognized movements as computer input commands. In an embodiment, multiple wireless devices may be placed on the user's wrists to enable multi-touch input commands. In an embodiment, data from motion sensors, such as accelerometers and gyroscopes, may be combined with movement data based on radio signal information.

Abstract: Various operations may be performed based on a distance-related function associated with two or more devices. For example, an association procedure for two or more devices may be based on one or more determined distances. Similarly, presence management may be based on one or more determined distances. A distance-related function may take various form including, for example, a distance between devices, two or more distances between devices, a rate of change in a relative distance between devices, relative acceleration between devices, or some combination of two or more of the these distance-related functions.

Abstract: Systems and methods are described that facilitate evaluating conditions of nodes (e.g., access points, access terminals, etc.) in a wireless communication environment to determine a level of disadvantage for a given node relative to other nodes. A first node may receive a resource utilization message (RxRUM) and may determine a level of disadvantage for a node that sent the RxRUM. The first node may then compare its own level of disadvantage to the sending node in order to permit a determination of an appropriate course of action in response to the RxRUM.

Abstract: A method of updating a set of classifiers includes applying a first set of classifiers to a first set of data. The method further includes requesting, from a remote device, a classifier update based on an output of the first set of classifiers or a performance measure of the application of the first set of classifiers.

Abstract: A method of generating a classifier model includes distributing a common feature model to two or more users. Multiple classifiers are trained on top of the common feature model. The method further includes distributing a first classifier of the multiple classifiers to a first user and a second classifier of the multiple classifiers to a second user.

Abstract: A method of learning a model includes receiving model updates from one or more users. The method also includes computing an updated model based on a previous model and the model updates. The method further includes transmitting data related to a subset of the updated model to the a user(s) based on the updated model.

Abstract: Methods and apparatuses are disclosed for providing flexible channel information feedback. In some cases this may be according to reporting types assigned by one or more sectors, with different reporting types for each sector. In other cases, a reporting mode is determined to determine the reporting types to be utilized.

Abstract: Certain aspects of the present disclosure support methods and apparatus for temporal spike encoding for temporal learning in an artificial nervous system. The temporal spike encoding for temporal learning can comprise obtaining sensor data being input into the artificial nervous system, processing the sensor data to generate feature vectors, converting element values of the feature vectors into delays, and causing at least one artificial neuron of the artificial nervous system to spike at times based on the delays.

Abstract: The disclosure relates to systems and methods for controlling a delay probability distribution associated with receiving a response to a page. The method entails performing a series of page operations, wherein each page operation entails transmitting a page and scanning for a page response. The method further entails adjusting at least one timing parameter associated with performing the series of page operations based on a characteristic of one or more scans for the page performed by the at least one remote device. The characteristic may be the period of periodic page scans performed by the at least one remote device.

Abstract: Techniques for efficiently choosing a desired decoder in the presence of varying Doppler and/or mobility are disclosed. In one aspect, a method for selecting a filter for decoding forward-link (FL) information in a wireless communication network comprises the acts of decoding reverse-link (RL) information by a plurality of decoders, each being optimized based on a different parameter, and comparing a plurality of outputs from the decoders, based on a metric, to determine a desired decoder or parameter for reporting to an access terminal.