18420635. INSTANCE-ADAPTIVE IMAGE AND VIDEO COMPRESSION USING MACHINE LEARNING SYSTEMS simplified abstract (QUALCOMM Incorporated)
Contents
INSTANCE-ADAPTIVE IMAGE AND VIDEO COMPRESSION USING MACHINE LEARNING SYSTEMS
Organization Name
Inventor(s)
Ties Jehan Van Rozendaal of Amsterdam (NL)
Iris Anne Marie Huijben of Eindhoven (NL)
Taco Sebastiaan Cohen of Amsterdam (NL)
INSTANCE-ADAPTIVE IMAGE AND VIDEO COMPRESSION USING MACHINE LEARNING SYSTEMS - A simplified explanation of the abstract
This abstract first appeared for US patent application 18420635 titled 'INSTANCE-ADAPTIVE IMAGE AND VIDEO COMPRESSION USING MACHINE LEARNING SYSTEMS
Simplified Explanation: The patent application describes techniques for compressing data using machine learning systems and tuning these systems for compression. An example process involves a neural network compression system receiving input data, determining updates for the system, generating compressed versions of the input data and model parameters, and outputting these compressed versions for transmission.
Key Features and Innovation:
- Compression of data using machine learning systems
- Tuning machine learning systems for data compression
- Neural network compression system trained on a training dataset
- Generation of compressed bitstreams using latent and model priors
- Outputting compressed data for transmission
Potential Applications: The technology can be applied in various fields such as:
- Data compression in communication systems
- Image and video compression
- Efficient storage of large datasets
- Optimization of machine learning models
Problems Solved: The technology addresses the following problems:
- Inefficient data compression methods
- Limited storage capacity for large datasets
- Slow transmission speeds due to large data sizes
Benefits: The technology offers the following benefits:
- Improved data compression ratios
- Faster transmission speeds
- Enhanced storage efficiency
- Optimized machine learning model performance
Commercial Applications: Potential commercial applications include:
- Telecommunications industry for efficient data transmission
- Cloud storage providers for optimized data storage
- Video streaming services for improved compression algorithms
- Machine learning companies for enhanced model optimization
Prior Art: Readers can explore prior art related to data compression, machine learning systems, and neural network optimization techniques.
Frequently Updated Research: Stay updated on the latest advancements in data compression, machine learning, and neural network optimization for improved performance and efficiency.
Questions about Data Compression: 1. How does machine learning contribute to data compression techniques? 2. What are the key advantages of using neural networks for data compression?
Original Abstract Submitted
Techniques are described for compressing data using machine learning systems and tuning machine learning systems for compressing the data. An example process can include receiving, by a neural network compression system (e.g., trained on a training dataset), input data for compression by the neural network compression system. The process can include determining a set of updates for the neural network compression system, the set of updates including updated model parameters tuned using the input data. The process can include generating, by the neural network compression system using a latent prior, a first bitstream including a compressed version of the input data. The process can further include generating, by the neural network compression system using the latent prior and a model prior, a second bitstream including a compressed version of the updated model parameters. The process can include outputting the first bitstream and the second bitstream for transmission to a receiver.
