Sparse Matrix Compression for Low-Loss Data Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data compression methods for large data transmissions in communication scenarios, such as point cloud and AI model data, result in high resource occupation and transmission delays with significant data loss.
Innovation Solution
Implementing data compression through dictionary learning and low-rank approximation on sparse matrices, combined with residual-based compression techniques to enhance the compression rate and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If existing data compression methods are used, then transmission resources are occupied and transmission delay increases, but data loss is significant
Solution Approach 1:
The patent segments the original data into multiple blocks and processes each block independently through dictionary learning to generate sparse matrices. This segmentation allows for more granular compression control while maintaining overall data integrity, reducing both transmission resource occupation and data loss simultaneously
Solution Approach 2:
The patent transforms the data representation parameters by converting original data into sparse matrices through dictionary learning. This parameter transformation enables more efficient compression with lower transmission resource requirements while preserving data information through the mathematical properties of sparse representations
2Productivity
If scalar quantization or vector quantization is used for data compression, then transmission resources are saved and transmission delay is reduced, but compression rate is limited and data loss is large
Solution Approach 1:
The patent replaces traditional quantization mechanisms with dictionary learning and sparse matrix transformation. Instead of using fixed quantization steps, the system learns adaptive dictionaries that capture data characteristics, achieving higher compression rates with minimal information loss through this mechanism substitution
Solution Approach 2:
The patent performs preliminary dictionary learning and sparse matrix transformation before actual data transmission. This preliminary action prepares the data in a compressed yet information-preserving format, enabling efficient transmission with both high compression rate and low data loss
3Productivity
If compression is performed on individual sparse matrices separately, then processing is simple, but overall compression efficiency is low
Solution Approach 1:
The patent merges multiple individual sparse matrix compressions into a unified joint compression process. By combining the compression operations and sharing computational resources across data blocks, the system achieves higher overall compression efficiency while managing processing complexity through coordinated operations
Data Source
AI summary
A data compression and transmission method, an apparatus, a device, and a storage medium. A first apparatus obtains M pieces of first data. One piece of subdata in the first data corresponds to one first sparse matrix, the first sparse matrix represents one piece of corresponding subdata in the first data based on a first dictionary matrix, and the first dictionary matrix includes features of M pieces of subdata respectively corresponding to the M pieces of first data. The first apparatus outputs compressed data of the M first sparse matrices. M is an integer greater than 1.