Abstract:
Methods, systems, and devices for wireless communication are described. The method, systems, and devices may include receiving a plurality of sets of input bits associated with respective transmission symbol periods at an encoder of a transmitting device, the plurality of sets of input bits associated with a single input vector to be encoded into a single codeword. The encoder may process the plurality of sets of input bits to generate a plurality of sets of output bits associated with respective transmission symbol periods, and output a first of the plurality of sets of output bits associated with a first of the plurality of sets of input bits prior receiving a second of the plurality of sets of input bits, the second of the plurality of sets of input bits being received at the encoder subsequent to the first of the plurality of sets of input bits.
Abstract:
Techniques are described for wireless communication. One method includes identifying a transmission time interval (TTI)-level control region, where the cascaded control region includes a TTI-level common control region and a TTI-level UE-specific control region, and where the TTI-level common control region has a pointer to a location of the TTI-level UE-specific control region; and demodulating at least one of the TTI-level common control region and the TTI-level UE-specific control region. Some techniques enable control information in a first region to point to a second region that includes additional control information. Some techniques enable two-stages of control. Semi-persistent scheduling/prescheduling control may be configured to cause an entity to refrain from demodulating one of the stages of control.
Abstract:
A method for object classification by an electronic device is described. The method includes obtaining an image frame that includes an object. The method also includes determining samples from the image frame. Each of the samples represents a multidimensional feature vector. The method further includes adding the samples to a training set for the image frame. The method additionally includes pruning one or more samples from the training set to produce a pruned training set. One or more non-support vector negative samples are pruned first. One or more non-support vector positive samples are pruned second if necessary to avoid exceeding a sample number threshold. One or more support vector samples are pruned third if necessary to avoid exceeding the sample number threshold. The method also includes updating classifier model weights based on the pruned training set.
Abstract:
A method for object classification by an electronic device is described. The method includes obtaining an image frame that includes an object. The method also includes determining samples from the image frame. Each of the samples represents a multidimensional feature vector. The method further includes adding the samples to a training set for the image frame. The method additionally includes pruning one or more samples from the training set to produce a pruned training set. One or more non-support vector negative samples are pruned first. One or more non-support vector positive samples are pruned second if necessary to avoid exceeding a sample number threshold. One or more support vector samples are pruned third if necessary to avoid exceeding the sample number threshold. The method also includes updating classifier model weights based on the pruned training set.
Abstract:
A processor-implemented method for image compression using an artificial neural network (ANN) includes receiving, at an encoder of the ANN, an image and a spatial segmentation map corresponding to the image. The spatial segmentation map indicates one or more regions of interest. The encoder compresses the image according to a controllable spatial bit allocation. The controllable spatial bit allocation is based on a learned quantization bin size.
Abstract:
Systems and techniques are described herein for processing media data using a neural network system. For instance, a process can include obtaining a latent representation of a frame of encoded image data and generating, by a plurality of decoder transformer layers of a decoder sub-network using the latent representation of the frame of encoded image data as input, a frame of decoded image data. At least one decoder transformer layer of the plurality of decoder transformer layers includes: one or more transformer blocks for generating one or more patches of features and determine self-attention locally within one or more window partitions and shifted window partitions applied over the one or more patches; and a patch un-merging engine for decreasing a respective size of each patch of the one or more patches.
Abstract:
Methods, systems, and devices for encoding and decoding are described. To encode a vector, an encoder allocates information bits of the vector to channel instances of a channel that are separated into groups. The groups may vary in size and allocation of the information bits is based on a base sequence of a given length. During decoding, a decoder assigns different bit types to channels instances by dividing a codeword into a plurality of groups and assigning bit types to channel instances of the plurality of groups using the base sequence.
Abstract:
Techniques are described for wireless communication. One method for wireless communication includes transmitting, to a user equipment (UE), an indication of a control channel subcarrier spacing to be used by the UE; and transmitting a control message having a subcarrier spacing in accordance with the indication. Another method for wireless communication includes transmitting, to a UE, an indicator channel identifying at least a subcarrier spacing to be used in one or more subsequent control channels. One or more subsequent control channels are then transmitted in accordance with the subcarrier spacing indicated by the indicator channel.
Abstract:
Aspects of the present disclosure provide various flexible search space designs that can handle UEs with various capabilities and limitations. In some aspects of the disclosure, the downlink bandwidth may be divided into several self-contained sub-bands. Each sub-band may include one or more OFDM subcarriers or tones, and each UE may be configured to monitor one or more of the sub-bands for its downlink control channel. The sub-band is self-contained in such a way that each sub-band includes CCEs that are mapped to resource elements contained in the same sub-band. In some aspects of the disclosure, different sub-bands can be configured with the different transmission modes and pilot densities. The transmission mode, pilot density, and layer-ID may be determined as a function of each UE's search space.
Abstract:
A method and apparatus for determining rate matching behavior for a plurality of control resource sets (CORESETs) is provided. A base station (BS) bundles a plurality of CORESETs into a global resource set. The BS configures the global resource set for a User Equipment (UE). The BS determines resources assigned to a downlink data channel overlap at least a portion of the global resource set and determines whether data on the downlink data channel is to be rate matched around the global resource set or is to use resources in the global resource set. The BS transmits the data on the downlink data channel based on the determination of the rate matching.