Abstract:
Certain aspects provide techniques and apparatuses for efficiently processing inputs in a neural network using multiple receptive field sizes. An example method includes partitioning a first input into a first set of channels and a second set of channels. At a first layer of a neural network, the first set of channels and the second set of channels are convolved into a first output having a smaller dimensionality a dimensionality of the first input. The first set of channels and the first output are concatenated into a second input. The second input is convolved into a second output via a second layer of the neural network, wherein the second output merges a first receptive field generated by the first layer with a larger second receptive field generated by the second layer. One or more actions are taken based on at least one of the first output and the second output.
Abstract:
A processor-implemented method for adaptive quantization in an artificial neural network (ANN) includes receiving an ANN model. The ANN model has multiple channels of target activations. A quantization module is incorporated between a first linear layer of the ANN and a second linear layer of the ANN to generate an adapted ANN. The quantization module scales a first set of weights and biases of the first linear layer based on a learnable quantization module parameter and scales a second set of weights of the second linear layer based on an inverse of the learnable quantization module parameter.
Abstract:
The present disclosure presents a method and an apparatus for planning interference measurement resources (IMRs). For example, the example method may include assigning a transmission group identifier to a cell in a wireless network, mapping the transmission group identifier assigned to the cell to a corresponding transmission pattern of a combination of zero power (ZP) and non-ZP (NZP) channel state information-reference signals (CSI-RSs) transmitted from the cell and neighbors of the cell, and receiving, at the cell, a CSI report from a user equipment (UE) in communication with the cell, wherein the CSI report is received from the UE based at least on an interference measured by an IMR at the UE corresponding to the transmission pattern. As such, IMR planning may be achieved.
Abstract:
The present disclosure presents a method and an apparatus for coordinated scheduling at a cell. For example, the example method may include receiving a plurality of channel state information (CSI) reports from one or more user equipments (UEs) served by the cell, generating a plurality of cell reports based at least on the plurality of CSI reports received from the one or more UEs, transmitting the plurality of cell reports to a central scheduling entity, receiving a selected global transmission configuration, receiving one or more additional CSI reports from the one or more UEs subsequent to transmitting the plurality of cell reports to the central scheduling entity, and identifying a UE of the one or more UEs to serve based at least on the selected global transmission configuration and the one or more additional CSI reports. As such, coordinated scheduling at a cell may be achieved.
Abstract:
The disclosure provides techniques for reducing interference caused by a first device to a second device receiving a satellite-based positioning signal. A device such as a user equipment (UE) or base station (eNB) determines a threshold transmission power for a transmission frequency of the device. The device also determines a signal strength of the satellite-based positioning system signal at the device. The device then controls a transmission property of the device based on the signal strength of the satellite-based positioning system signal when a transmission power of the device at the transmission frequency satisfies the threshold. The device may also determine that reception of the satellite-based positioning system signal by the second device is likely to be affected by a transmission from the device at a transmission power that satisfies the threshold and control the transmission property when reception of the satellite-based positioning system signal is likely to be affected.
Abstract:
A method, an apparatus, and a computer program product for wireless communication are provided in connection with making UE handover decisions. In one example, a node is equipped to obtain one or more UE measurement values associated with a UE, determine a mobility state of the UE based at least in part on the obtained one or more UE measurement values, and adjust one or more handover parameters based at least in part on the determined mobility state of the UE. In another example, a UE is equipped to obtain one or more UE measurements, determine a handover state of a UE based on the obtained one or more UE measurements, and perform a handover-related action based on the determined handover state of the UE.
Abstract:
A method for wireless communication may comprise, for example, automatically determining first power settings for a first radio access technology (RAT) based on second power settings for a second RAT, and automatically applying the first power settings as power settings for the first RAT.
Abstract:
Resources are reserved for handover signaling. Handover reliability is thereby improved since the primary interference for handover signaling will be from the handover signaling of nearby cells (e.g., as opposed to data traffic). Given the relatively sparse nature of handover signaling, a significant reduction in interference may be achieved in a network employing this technique in comparison to conventional networks. In some implementations, a set of neighboring access points cooperate to determine which resources are reserved for handover signaling. In some implementations, resources for handover signaling are reserved on a network-wide basis or an access point cluster-wide basis.
Abstract:
Wireless communication in a heterogeneous network may be facilitated by establishing connections with femto nodes based on page notifications from macro nodes to user equipment (UE). A UE in idle mode may monitor paging resources provided by a macro node, discover presence of a femto node based on signals received from the femto node, and establish a communicative connection with the femto node based on a page notification received from the macro node over the paging resources.
Abstract:
A processor-implemented method for training-free architecture searching for a transformer model includes generating a set of transformer model candidates for a target device. Each transformer model candidate of the set of transformer model candidates is initialized with random weights. A set of data samples are randomly sampled to produce random data samples for inputting at each transformer model candidate. An attention confidence score is computed for each transformer model candidate based on the random data samples and the random weights. A transformer model candidate for the target device is selected based on the attention confidence score.