Granted Patent B2

US 12,216,623 · App. 18/822,208 · Granted Feb 4, 2025

Inventors: Joshua Cooper (Columbia, SC); Charles Yeomans (Orinda, CA); Brian Galvin (Silverdale, WA)

Assignee: ATOMBEAM TECHNOLOGIES INC

G06F16/1752G06F3/0608G06F3/0641G06F3/067

Abstract

Claims (55)

1. A system for random-access manipulation of a compacted data file, comprising:

a computing device comprising a memory, a processor, and a non-volatile data storage device;

a random access engine comprising a first plurality of programming instructions that, when operating on the processor, cause the computing device to:

receive a data search query directed to the compacted data file, wherein the compacted file is compacted using a reference codebook;

organize the reference codebook into a hierarchical representation;

traverse the hierarchical representation to identify a start codeword corresponding to the beginning of the data search query; and

send the start codeword and a plurality of immediately following codewords from the compacted data file to a decoder; and

an estimator module comprising a second plurality of programming instructions that, when operating on the processor, cause the computing device to:

estimate a first starting bit location in the compacted data file;

refine the first starting bit location by:

determining a plurality of codeword boundaries by performing distinct value estimation on the compacted data file, wherein the distinct value estimates correspond to a codeword boundary; and

determining whether a bit sequence starting at the first starting bit location corresponds to a codeword boundary of the plurality of codeword boundaries and, if not, traversing the hierarchical representation until a codeword boundary is located at a new starting bit.

2. The system of claim 1 , further comprising a dyadic distribution compression and encryption subsystem comprising a third plurality of programming instructions that, when operating on the processor, cause the computing device to:

analyze input data to determine its properties;

create a transformation matrix based on the properties of the input data;

transform the input data into a dyadic distribution;

generate a main data stream of transformed data and a secondary data stream of transformation information; and

compress the main data stream.

3. The system of claim 2 , further comprising a large codeword model with a latent transformer core comprising a fourth plurality of programming instructions that, when operating on the processor, cause the computing device to:

receive the compressed data as input vectors;

generate latent space vectors by processing the input vectors through a variational autoencoder's encoder;

process the latent space vectors through a transformer to learn relationships between the vectors, wherein the transformer does not include an embedding layer and a positional encoding layer; and

decode output latent space vectors through a variational autoencoder's decoder to produce final output data.

4. The system of claim 1 , wherein the random access engine uses the refined starting bit location as an initial point to begin traversing the hierarchical representation.

5. The system of claim 4 , further comprising an end to end training subsystem configured to jointly train the random access engine, the dyadic distribution compression and encryption module, the large codeword model, and the estimator module.

6. The system of claim 5 , wherein the end to end training subsystem is configured to optimize performance across all components from initial compression and encryption to final random access capabilities.

7. The system of claim 2 , wherein the dyadic distribution compression and encryption subsystem uses trainable parameters for transforming the input data and compressing the main data stream.

8. The system of claim 3 , wherein the large codeword model is configured to adapt its processing based on the output of the dyadic distribution compression and encryption module.

9. The system of claim 4 , wherein the estimator module uses learnable parameters to estimate the starting bit location and refine the estimate.

10. A method for random-access manipulation of a compacted data file, comprising the steps of:

receiving a data search query directed to the compacted data file, wherein the compacted file is compacted using a reference codebook;

organizing the reference codebook into a hierarchical representation;

traversing the hierarchical representation to identify a start codeword corresponding to the beginning of the data search query;

sending the start codeword and a plurality of immediately following codewords from the compacted data file to a decoder;

estimating, using an estimator module, a first starting bit location in the compact data file; and

refining the first starting bit location by:

determining a plurality of codeword boundaries by performing distinct value estimation on the compacted data file, wherein the distinct value estimates correspond to a codeword boundary; and

determining whether a bit sequence starting at the first starting bit location corresponds to a codeword boundary of the plurality of codeword boundaries and, if not, traversing the hierarchical representation until a codeword boundary is located at a new starting bit.

11. The method of claim 10 , further comprising:

analyzing input data to determine its properties;

creating a transformation matrix based on the properties of the input data;

transforming the input data into a dyadic distribution;

generating a main data stream of transformed data and a secondary data stream of transformation information; and

compressing the main data stream.

12. The method of claim 10 , further comprising:

receiving the compressed data as input vectors;

generating latent space vectors by processing the input vectors through a variational autoencoder's encoder;

processing the latent space vectors through a transformer to learn relationships between the vectors, wherein the transformer does not include an embedding layer and a positional encoding layer; and

decoding output latent space vectors through a variational autoencoder's decoder to produce final output data.

13. The method of claim 10 , wherein traversing the hierarchical representation uses the refined starting bit location as an initial point.

14. The method of claim 11 , further comprising jointly training, in an end-to-end manner, the steps of organizing, traversing, and sending along with the steps of analyzing, creating, transforming, generating, compressing, and estimating.

15. The method of claim 14 , wherein the joint training optimizes performance across all steps from initial compression and encryption to final random access capabilities.

16. The method of claim 12 , wherein transforming the input data and compressing the main data stream use trainable parameters.

17. The method of claim 12 , wherein processing the latent space vectors adapts based on the output of the compression step.

18. The method of claim 13 , wherein estimating the first starting bit location and refining the estimate use learnable parameters.

Assignments (1)

Continuity (14)

References Cited (9)