Method and device for decoding data

Patent metadata

Field	Value
Publication number	US-11769079-B2
Application number	US-202217729628-A
Country	US
Kind code	B2
Filing date	Apr 26, 2022
Priority date	Apr 30, 2021
Publication date	Sep 26, 2023
Grant date	Sep 26, 2023

Abstract

A method for decoding data by an electronic device is provided. The method includes receiving, by the electronic device, encoded data, determining, by the electronic device, a sparsity of a plurality of Machine Learning (ML) models of a turbo decoder of the electronic device for decoding the encoded data based on Quality-of-Service (QoS) parameters, and decoding, by the electronic device, the encoded data using the turbo decoder based on the determined sparsity.

First claim

We claim:

1. A method for decoding data by an electronic device, the method comprises:
receiving, by the electronic device, encoded data;
determining, by the electronic device, a sparsity of a plurality of Machine Learning (ML) models for decoding the encoded data based on quality of service (QoS) parameters; and
decoding, by the electronic device, the encoded data using a turbo decoder based on the determined sparsity.

2. The method of claim 1 , wherein the QoS parameters comprise a QoS Class Identifier (QCI), a QoS, a Bandwidth Part (BWP), a Subcarrier Spacing (SCS), a Bandwidth (BW), a coherence BW, a coherence time, an interference, a noise, an operating frequency, a User Equipment (UE) capability, a Multiple-Input Multiple-Output (MIMO) capability, and a transmission mode.

3. The method of claim 1 , wherein the method further comprises:
determining, by the electronic device, whether an accuracy of the decoded data from the turbo decoder with respect to real decoded data of the encoded data is greater than a threshold value; and
increasing, by the electronic device, the sparsity of the plurality of ML models, in response to determining that the accuracy of the decoded data from the turbo decoder with respect to the real decoded data of the encoded data is greater than the threshold value.

4. The method of claim 3 wherein, increasing the sparsity of trained Machine Learning (ML) models of a plurality of turbo decoders, comprises:
receiving, by a central electronic device, weights of each layer of the trained ML models of the plurality of turbo decoders;
determining, by the central electronic device, an average of the weights of each layer of the trained ML models of the plurality of turbo decoders; and
updating, by the central electronic device, the weights of each layer of the trained ML models of the plurality of turbo decoders with the average of the weights for increasing the sparsity of the ML models of the plurality of decoders.

5. The method of claim 1 , wherein determining the sparsity of the plurality of ML models based on the QoS parameters, comprises:
determining, by the electronic device, the QoS parameters comprising a code word size, a code rate, a Signal to Noise Ratio, long/short filters, interference, a load of neighboring base station, a sub carrier spacing, and an operating frequency; and
predicting, by the electronic device, the sparsity of the plurality of ML models by providing the QoS parameters to a ML model.

6. The method of claim 1 , wherein determining the sparsity of the plurality of ML models based on the QoS parameters, comprises:
assigning, by the electronic device, default weights for each layer of the plurality of ML models and a default sparsity to the plurality of ML models;
training, by the electronic device, the plurality of ML models in a fully connected mode without removing any connections based on the QoS parameters;
updating, by the electronic device, weights of each layer of the plurality of ML models by performing to Stochastic Gradient Descent (SGD) and Stochastic Weight Averaging (SWA) on the default weights;
hierarchically training, by the electronic device, the plurality of ML models while incrementally increasing the sparsity of the plurality of ML models; and
determining, by the electronic device, the sparsity of the plurality of ML models.

7. The method of claim 6 , wherein hierarchically training, by the electronic device, the plurality of ML models while incrementally increasing the sparsity of the plurality of ML models, comprises:
sorting, by the electronic device, the updated weights of each layer in an increasing order;
determining, by the electronic device, whether the updated weights of each layer meet a threshold condition;
identifying, by the electronic device, weak connections between layers of the plurality of ML models in response to determining the layers with the updated weights meet the threshold condition; and
deleting, by the electronic device, the weak connections between layers of the plurality of ML models for incrementally increasing the sparsity of the plurality of ML models.

8. An electronic device for decoding data, the electronic device comprising:
a memory;
a processor;
a turbo decoder comprising a plurality of Machine Learning (ML) models; and
a sparsity controller, coupled to the memory and the processor, the sparsity controller being configured to:
receive encoded data,
determine a sparsity of the plurality of ML models of the turbo decoder for decoding the encoded data based on quality of service (QoS) parameters, and
decode the encoded data using the turbo decoder based on the determined sparsity.

9. The electronic device of claim 8 , wherein the QoS parameters comprising a QoS Class Identifier (QCI), a QoS, a Bandwidth Part (BWP), a Subcarrier Spacing (SCS), a Bandwidth (BW), a coherence BW, a coherence time, an interference, a noise, an operating frequency, a User Equipment (UE) capability, a Multiple-Input Multiple-Output (MIMO) capability, and a transmission mode.

10. The electronic device of claim 8 , wherein the sparsity controller is further configured to:
determine whether an accuracy of the decoded data from the turbo decoder with respect to real decoded data of the encoded data is greater than a threshold value; and
increase the sparsity of the plurality of ML models, in response to determining that the accuracy of the decoded data from the turbo decoder with respect to the real decoded data of the encoded data is greater than the threshold value.

11. The electronic device of claim 8 , wherein determining the sparsity of the plurality of ML models based on the QoS parameters, comprises:
determining the QoS parameters comprising a code word size, a code rate, a Signal to Noise Ratio (SNR), long/short filters, interference, a load of neighboring base station, a sub carrier spacing, and an operating frequency; and
predicting the sparsity of the plurality of ML models by providing the QoS parameters to a ML model.

12. The electronic device of claim 8 , wherein determining the sparsity of the plurality of ML models based on the QoS parameters, comprises:
assigning default weights for each layer of the plurality of ML models and a default sparsity to the plurality of ML models;
training the plurality of ML models in a fully connected mode without removing any connections based on the QoS parameters;
updating weights of each layer of the plurality of ML models by performing to Stochastic Gradient Descent (SGD) and Stochastic Weight Averaging (SWA) on the default weights;
hierarchically training the plurality of ML models while incrementally increasing the sparsity of the plurality of ML models; and
determining the sparsity of the plurality of ML models.

13. The electronic device of claim 12 , wherein hierarchically training the plurality of ML models while incrementally increasing the sparsity of the plurality of ML models, comprises:
sorting the updated weights of each layer in an increasing order;
determining whether the updated weights of each layer meet a threshold condition;
identifying weak connections between layers of the plurality of ML models in response to determining the layers with the updated weights meet the threshold condition; and
deleting the weak connections between layers of the plurality of ML models for incrementally increasing the sparsity of the plurality of ML models.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Kumar Vankayala Satya
• Singh Sukhmeet
• Kumar Swaraj
• Yoon Seungil
• Kommineni Issaac

Classifications

• G06N20/00 Primary

  Machine learning · CPC title

• H03M13/2963

  Turbo-block codes, i.e. turbo codes based on block codes, e.g. turbo decoding of product codes · CPC title

• H04L1/0009

  by adapting the channel coding (H04L1/1812 takes precedence) · CPC title

• H04W28/0268 Primary

  using specific QoS parameters for wireless networks, e.g. QoS class identifier [QCI] or guaranteed bit rate [GBR] (negotiating SLA or negotiating QoS H04W28/24) · CPC title

• H03M13/2703

  the interleaver involving at least two directions · CPC title