Abstract: This disclosure is directed to methods and systems for protecting a deep learning model from piracy and unauthorized uses. The protection may be implemented by embedding an ownership detection mechanism such that unauthorized use of the model may be detected using a detection input data and corresponding model signature. In addition, the deep learning model may be used in conjunction with a secret or license protected data encoder such that the deep learning model may generate meaningful output only when processing encoded input data. An unauthorized user who does not have access to the secret data encoder may not be able to use a pirated copy of the deep learning model to generate meaningful output. Under such a scheme, a deep learning model itself may be widely distributed without restriction and without license-protection.

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: Computing a non-linear function ƒ(x) in hardware or embedded systems can be complex and resource intensive. In one or more aspects of the disclosure, a method, a computer-readable medium, and an apparatus are provided for computing a non-linear function ƒ(x) accurately and efficiently in hardware using look-up tables (LUTs) and interpolation or extrapolation. The apparatus may be a processor. The processor computes a non-linear function ƒ(x) for an input variable x, where ƒ(x)=g(y(x),z(x)). The processor determines an integer n by determining a position of a most significant bit (MSB) of an input variable x. In addition, the processor determines a value for y(x) based on a first look-up table and the determined integer n. Also, the processor determines a value for z(x) based on n and the input variable x, and based on a second look-up table. Further, the processor computes ƒ(x) based on the determined values for y(x) and z(x).

Abstract: Computing a non-linear function ƒ(x) in hardware or embedded systems can be complex and resource intensive. In one or more aspects of the disclosure, a method, a computer-readable medium, and an apparatus are provided for computing a non-linear function ƒ(x) accurately and efficiently in hardware using look-up tables (LUTs) and interpolation or extrapolation. The apparatus may be a processor. The processor computes a non-linear function ƒ(x) for an input variable x, where ƒ(x)=g(y(x),z(x)). The processor determines an integer n by determining a position of a most significant bit (MSB) of an input variable x. In addition, the processor determines a value for y(x) based on a first look-up table and the determined integer n. Also, the processor determines a value for z(x) based on n and the input variable x, and based on a second look-up table. Further, the processor computes ƒ(x) based on the determined values for y(x) and z(x).

Abstract: Techniques are described for wireless communication. One method includes implementing, at a first node, a first access protocol to contend for access to a wireless communication medium shared by a plurality of nodes; determining whether a triggering event has occurred; and implementing, at the first node, a second access protocol to contend for access to the wireless communication medium based at least in part on a determination that the triggering event has occurred.

Abstract: A method of detecting unknown classes is presented and includes generating a first classifier for multiple first classes. In one configuration, an output of the first classifier has a dimension of at least two. The method also includes designing a second classifier to receive the output of the first classifier to decide whether input data belongs to the multiple first classes or at least one second class.

Abstract: Techniques are described for wireless communication. A first method may include transmitting a reference signal from an access point over a shared spectrum to a user equipment (UE), and receiving a measurement of the reference signal from the UE. A second method may include receiving at a UE over a shared spectrum a reference signal from an access point, and transmitting a measurement of the reference signal to the access point. In each of the first and second methods, the measurement may indicate interference with the shared spectrum from outside of a clear channel assessment (CCA) range of the access point.

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 reducing computational complexity for a fixed point neural network operating in a system having a limited bit width in a multiplier-accumulator (MAC) includes reducing a number of bit shift operations when computing activations in the fixed point neural network. The method also includes balancing an amount of quantization error and an overflow error when computing activations in the fixed point neural network.

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: The present disclosure presents a method and an apparatus for transmitting discovery signaling from a base station. For example, the method may include encoding a wireless fidelity (Wi-Fi) beacon at the base station for transmission and transmitting the encoded Wi-Fi beacon from the base station to one or more neighboring wireless nodes. The Wi-Fi beacon is generated by a Wi-Fi access point (AP) co-located at the base station which is a long term evolution (LTE) or LTE advanced in unlicensed spectrum base station. As such, other wireless nodes can discover the LTE or LTE advanced in unlicensed spectrum base station.

Abstract: A selection process is disclosed for a user equipment (UE) to select between a WAN and a WLAN. Instead of defaulting to select one or the other networks, as long as the connection is available, a better load balancing and maximization of the system capacity may be achieved when the UE selects the network according to a priority order between the WAN/WLAN. The priority order takes into account various network metrics, such as loading, channel conditions, bandwidth, and service requirements.

Abstract: Techniques for sending signaling messages with beacon signals in a wireless communication network are described. In one design, a transmitter station may map a signaling message (e.g., a reduce interference request) to multiple code symbols. The transmitter station may select multiple resource elements from among a plurality of resource elements based on the multiple code symbols. In one design, each code symbol may be sent across frequency by selecting one of multiple subcarriers in one symbol period. In another design, each code symbol may be sent across time by selecting one of multiple symbol periods on one subcarrier. The transmitter station may generate a beacon signal having transmit power on the selected resource elements and no transmit power on remaining resource elements. The transmitter station may send the beacon signal to at least one receiver station.

Abstract: Systems and methodologies are described herein that facilitate techniques for design of relay backhaul to support mobility of relay nodes in a wireless communication system. According to various aspects herein, techniques are provided to enable and support the use of mobile relays and to facilitate handover of mobile relays between respective donor cells. More particularly, techniques are provided herein for relay backhaul control channel assignment associated with hand in or hand out of mobile relays, access/backhaul resource partitioning for mobile relays, and management of quality of service (QoS) requirements associated with a relay handover.

Abstract: Methods and apparatuses are provided that include selecting resources for assigning to a device to mitigate relay self-interference when also communicating with a base station. The resources can be selected based on one or more factors, such as based on resources that are negotiated with the base station, or based on resources indicated as not desired for allocation from the base station, etc. In other examples, reference signals and control data can be communicated such as to mitigate relay self-interference as well.

Abstract: Techniques are described for wireless communication. One method includes implementing, at a first node, a first access protocol to contend for access to a wireless communication medium shared by a plurality of nodes; determining whether a triggering event has occurred; and implementing, at the first node, a second access protocol to contend for access to the wireless communication medium based at least in part on a determination that the triggering event has occurred.

Abstract: Methods and apparatuses are provided that facilitate associating with relays in a wireless network. A device can select whether to utilize relay assistance where present based at least in part on measuring one or more determined or projected parameters related to the relay. Where utilizing a relay results in user-plane data channel conditions above a threshold level and control channel conditions below a threshold level, a serving base station can determine whether to employ another base station to serve the device, jointly serve a relay with an additional base station, and/or the like.

Abstract: Techniques are described for wireless communication. A first method may include transmitting a reference signal from an access point over a shared spectrum to a user equipment (UE), and receiving a measurement of the reference signal from the UE. A second method may include receiving at a UE over a shared spectrum a reference signal from an access point, and transmitting a measurement of the reference signal to the access point. In each of the first and second methods, the measurement may indicate interference with the shared spectrum from outside of a clear channel assessment (CCA) range of the access point.