Samsung patents method to compress AI model data for faster loading
Samsung describes a system that shrinks the data needed to run specific AI tasks before moving it into the device's fast memory. The system expands the data back to its original size only when the AI task is ready to run.
In the filing’s own words
Disclosed is an electronic device. The electronic device may store weights related to an artificial intelligence (AI) model in a non-volatile memory. The electronic device may compress, by means of a compressor, some weights related to a designated operation among the weights stored in the non-volatile memory on the basis of a load request for the designated operation, and load the compressed weights to a first volatile memory. The electronic device may decompress the some weights compressed and loaded to the first volatile memory by means of a decompressor and load the decompressed weights to a second volatile memory. The electronic device may perform the designated operation on the basis of the some weights decompressed and loaded to the second volatile memory.
What exactly they claim (claim 1)
1. An electronic device comprising: a compressor compressing data based on a compression algorithm; a decompressor decompressing data based on a decompression algorithm corresponding to the compression algorithm; at least one processor; non-volatile memory storing instructions; a first volatile memory; and a second volatile memory, wherein the instructions, are loaded into the first volatile memory, and, when executed by the at least one processor, individually or collectively, cause the electronic device to: store, in the non-volatile memory, weights related to an artificial intelligence, artificial intelligence (AI), model, based on a load request for a specified operation, compress, through the compressor, partial weights related to the specified operation, wherein the partial weights correspond to some of the weights among the weights stored in the non-volatile memory, and load the compressed partial weights into the first volatile memory, decompress, through the decompressor, the partial weights with a compressed state, loaded into the first volatile memory, and load the decompressed partial weights into the second volatile memory, and perform the specified operation based on the partial weights with a decompressed state, loaded into the second volatile memory.
Facts
- Company
- Samsung (Samsung Electronics Co., Ltd.)
- Official title
- ELECTRONIC DEVICE, METHOD, AND NON-TRANSITORY COMPUTER-READABLE RECORDING MEDIUM FOR PROTECTING ARTIFICIAL INTELLIGENCE MODEL
- Publication number
- US20260291516A1
- Application number
- 19676287
- Filed
- May 13, 2026
- Published
- Sep 24, 2026
- Inventors
- 7
- Claims
- 20
Classified under
H03M 7/3077 Electricity & electronicsH03M 7/6005 Electricity & electronicsH03M 7/6011 Electricity & electronicsMore from Samsung in AI & software
- Samsung patents tool to erase objects from photos using AI Sep 24, 2026
- Samsung patents drawing objects into photos with AI Oct 15, 2026
- Samsung patents AI that creates personalized exercise videos Sep 10, 2026
- Samsung patents AI summaries for paused TV shows based on your viewing habits Sep 24, 2026
- Samsung patents turning chat conversations into image prompts Oct 1, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.