Neural Upsampling of Genomic Data After Lossy Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data compression techniques, particularly lossy compression methods like HEVC, result in data loss that is challenging to recover, especially in complex and diverse data sets such as genomic data, where information loss can compromise biological significance and accuracy.
Innovation Solution
A system and method utilizing a neural network for upsampling decompressed genomic data after lossy compression, incorporating a novel AI deblocking network with recurrent layers for feature extraction and a channel-wise transformer with attention to capture inter-channel dependencies, effectively mitigating compression artifacts and improving data quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If lossy compression is applied to genomic data, then compression ratio is improved, but information loss occurs
Solution Approach 1:
A neural network is introduced as an intermediary component between the lossy compression process and the final decompressed output. The neural network takes the decompressed data as input and applies learned transformations to recover information that was lost during compression, thereby mediating between the compressed representation and the original data quality
Solution Approach 2:
The system changes the state of the decompressed data by applying neural network transformations that modify the data parameters. The neural network learns optimal parameter adjustments during training to reconstruct lost information, effectively changing the data from a compressed state back toward its original state
2Loss of information
If lossless compression is applied to genomic data, then information retention is improved, but compression ratio deteriorates
Solution Approach 1:
Instead of applying full lossless compression, the system applies partial compression followed by neural network enhancement. This partial action approach achieves sufficient information retention for many applications while avoiding the excessive compression overhead of complete lossless methods
Solution Approach 2:
The patent replaces traditional mechanical compression algorithms with a neural network-based system. The neural network learns compression and reconstruction patterns from data, substituting rule-based mechanical compression with learned, adaptive transformations that can achieve better performance
3Quantity of substance
If compression algorithms are applied to diverse genomic data, then data volume reduction is improved, but algorithm scalability deteriorates
Solution Approach 1:
The neural network is designed with universal applicability to handle diverse genomic data types. By training on multiple data types and using architecture components that can process different input formats, the system achieves multi-functionality that scales across various genomic datasets without requiring algorithm redesign
Data Source
AI summary
A system and methods for upsampling of decompressed genomic data after lossy compression using a neural network integrates AI-based techniques to enhance compression quality. It incorporates a novel deep-learning neural network that upsamples decompressed data to restore information lost during lossy compression, taking advantage of cross-correlations between genomic data sets.