Microsoft Technology Licensing, LLC

Apple Inc.

Yu Wu

Shujie Liu

Chengyi Wang

Jinyu Li

Nanshan Zeng

Furu Wei

Xie Chen

Kenichi Kumatani

Xuedong David Huang

Yao Qian

Jing Wang

Lijie Zhang

Zhiwei Wang

Qiang Miao

Lele Cui

Yun Gao

Aiman J Sang

Aspects are described for a user equipment (UE) comprising a transceiver configured to enable wireless communication with a base station; and a processor communicatively coupled to the transceiver. The processor is configured to receive a sounding reference signal (SRS) resource configuration from the base station, wherein the SRS resource configuration indicates at least a first SRS resource and a second SRS resource. The processor is further configured to transmit a first SRS via the first SRS resource to the base station using a first antenna coupled to the transceiver and transmit a second SRS via the second SRS resource to the base station using a second antenna coupled to the transceiver. The processor is further configured to receive an SRS resource indicator (SRI) from the base station based on at least one of the transmission of the first SRS or the second SRS; select the first antenna based on the SRI indicating the first SRS resource; and transmit uplink data to the base station using the first antenna.

Method of codebook sounding reference signal (SRS) antenna mapping to improve uplink performance

A system obtains a first training data set comprising labeled speech data or both labeled and unlabeled data corresponding to a high-resource data set as well as latent speech representations based on the first training data set. The system trains a machine learning model on the first training data set to learn phonetically aware speech representations corresponding to the first training data set. The system applies the latent speech representations to a transformer context network to generate contextual representations. The system aligns each of the contextual representations with a phoneme label to generate phonetically-aware contextual representations. The system causes a refinement engine to further refine the machine learning model.

Unified speech representation learning

The computing system trains an end-to-end (E2E) automatic speech recognition (ASR) model, using a transformer-transducer-based deep neural network that comprises a transformer encoder network and a transducer predictor network. The E2E ASR model is trained to have one or more adjustable hyperparameters that are configured to dynamically adjust an efficiency or a performance of the E2E ASR model when the E2E ASR model is deployed onto a device or executed by the device, by identifying one or more conditions of the device associated with computational power of the device and setting at least one of the one or more adjustable hyperparameters based on one or more conditions of the device.

Efficiency adjustable speech recognition system

Systems and methods are provided for training a machine learning model to learn speech representations. Labeled speech data or both labeled and unlabeled data sets is applied to a feature extractor of a machine learning model to generate latent speech representations. The latent speech representations are applied to a quantizer to generate quantized latent speech representations and to a transformer context network to generate contextual representations. Each contextual representation included in the contextual representations is aligned with a phoneme label to generate phonetically-aware contextual representations. Quantized latent representations are aligned with phoneme labels to generate phonetically aware latent speech representations. Systems and methods also include randomly replacing a sub-set of the contextual representations with quantized latent speech representations during their alignments to phoneme labels and aligning the phonetically aware latent speech representations to the contextual representations using supervised learning.

A computing system is configured to generate a transformer-transducer-based deep neural network. The transformer-transducer-based deep neural network comprises a transformer encoder network and a transducer predictor network. The transformer encoder network has a plurality of layers, each of which includes a multi-head attention network sublayer and a feed-forward network sublayer. The computing system trains an end-to-end (E2E) automatic speech recognition (ASR) model, using the transformer-transducer-based deep neural network. The E2E ASR model has one or more adjustable hyperparameters that are configured to dynamically adjust an efficiency or a performance of E2E ASR model when the E2E ASR model is deployed onto a device or executed by the device.

Growing community of inventors

Beijing, China

Yu Wu