# US 12,216,621 B2 · Hyperparameter optimization in file compression using sequence alignment > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,216,621 B2 - Document ID: US12216621B2 - Status: Granted - Kind: B2 - Application number: 17/658,930 - Filed: 2022-04-12 - Published: 2023-10-12 - Granted: 2025-02-04 - Earliest claimed: 2022-04-12 - Assignees: Dell Products L.P. - Applicants: Dell Products L.P. - Inventors: Ofir Ezrielev; Ilan Buyum; Jehuda Shemer - Web page: https://patentgrep.com/patent/US12216621B2 ## CPC - G06F 16/1744: {using compression, e.g. sparse files} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Compressing files is disclosed. An input file to be compressed is first aligned. During or prior to aligning the input file, hyperparameters are set, determined, or configured. The hyperparameters may be set, determined, or configured to achieve a particular performance characteristic. Aligning the file includes splitting the file into sequences that can be aligned. The result is a compression matrix, where each row of the matrix corresponds to part of the file. A consensus sequence id determined from the compression matrix. Using the consensus sequence, pointer pairs are generated. Each pointer pair identifies a subsequence of the consensus matrix. The compressed file includes the pointer pairs and the consensus sequence. ## Claims ### Claim 1 A method, comprising: receiving an input file into a compression engine; inputting hyperparameters associated with the compression engine, wherein the hyperparameters include one or more of initial splits and a minimal consensus length reduction; aligning the input file, based on the hyperparameters, to create a compression matrix that includes sequences, wherein aligning the input file includes splitting the input file one or more times starting from the initial splits to generate sequences, wherein each of the sequences corresponds to a different portion of the input file, wherein the sequences are arranged as rows of the compression matrix, wherein each column of the sequences in the compression matrix includes identical data after the sequences are aligned in the compression matrix, wherein the data in a column of the compression matrix are identical when each cell of the column, after the sequences are aligned in the compression matrix, comprise the same characters or the same characters and at least one dash; determining a consensus sequence from the compression matrix, wherein the consensus sequence includes an entry from each of the columns and does not include any dashes; and generating a compressed file that includes the consensus sequence and pointer pairs, wherein each pointer pair identifies a subsequence of the consensus sequence and each of the rows of the compression matrix is replaced by at least one of the pointer pairs, the pointer pairs being smaller than the corresponding sequences. ### Claim 2 (depends on 1) The method of claim 1, further comprising inputting hyperparameters in order to achieve a particular goal, wherein the particular goal is one of a compression ratio, a computation time, resource usage, ease of retrieval, ease of sharding, or combination thereof. ### Claim 3 (depends on 1) The method of claim 1 wherein the at least one dash is not represented by pointers in the compressed file and the at least one dash represents a gap or a space. ### Claim 4 (depends on 2) The method of claim 2, wherein the hyperparameters further include one or more of a letter size, scoring function parameters, and hierarchical hyperparameters. ### Claim 5 (depends on 2) The method of claim 2, further comprising optimizing the hyperparameters. ### Claim 6 (depends on 5) The method of claim 5, further comprising optimizing the hyperparameters by making changes to one or more of the hyperparameters to achieve the particular goal. ### Claim 7 (depends on 6) The method of claim 6, further comprising optimizing the hyperparameters to improve a performance characteristic while ensuring that a reduction in other performance characteristics is kept within a threshold. ### Claim 8 (depends on 6) The method of claim 6, further comprising optimizing the hyperparameters individually or as a group. ### Claim 9 (depends on 1) The method of claim 1, further comprising adjusting the hyperparameters while aligning the input file or prior to aligning the input file. ### Claim 10 (depends on 1) The method of claim 1, further comprising generating multiple compressed files, wherein each of the compressed files is associated with different hyperparameters. ### Claim 11 A non-transitory storage medium having stored therein instructions that are executable by one or more hardware processors to perform operations comprising: receiving an input file into a compression engine; inputting hyperparameters associated with the compression engine, wherein the hyperparameters include one or more of initial splits and a minimal consensus length reduction; aligning the input file, based on the hyperparameters, to create a compression matrix that includes sequences, wherein aligning the input file includes splitting the input file one or more times starting from the initial splits to generate sequences, wherein each of the sequences corresponds to a different portion of the input file, wherein the sequences are arranged as rows of the compression matrix, wherein each column of the sequences in the compression matrix includes identical data after the sequences are aligned in the compression matrix, wherein the data in a column of the compression matrix are identical when each cell of the column, after the sequences are aligned in the compression matrix, comprise the same characters or the same characters and at least one dash; determining a consensus sequence from the compression matrix, wherein the consensus sequence includes an entry from each of the columns and does not include any dashes; and generating a compressed file that includes the consensus sequence and pointer pairs, wherein each pointer pair identifies a subsequence of the consensus sequence and each of the rows of the compression matrix is replaced by at least one of the pointer pairs, the pointer pairs being smaller than the corresponding sequences. ### Claim 12 (depends on 11) The non-transitory storage medium of claim 11, further comprising inputting hyperparameters in order to achieve a particular goal, wherein the particular goal is one of a compression ratio, a computation time, resource usage, ease of retrieval, ease of sharding, or combination thereof. ### Claim 13 (depends on 11) The non-transitory storage medium of claim 11, wherein the at least one dash is not represented by pointers in the compressed file and the at least one dash represents a gap or a space. ### Claim 14 (depends on 12) The non-transitory storage medium of claim 12, wherein the hyperparameters further include one or more of a letter size, scoring function parameters, and hierarchical hyperparameters. ### Claim 15 (depends on 12) The non-transitory storage medium of claim 12, further comprising optimizing the hyperparameters. ### Claim 16 (depends on 15) The non-transitory storage medium of claim 15, further comprising optimizing the hyperparameters by making changes to one or more of the hyperparameters to achieve the particular goal. ### Claim 17 (depends on 16) The non-transitory storage medium of claim 16, further comprising optimizing the hyperparameters to improve a performance characteristic while ensuring that a reduction in other performance characteristics is kept within a threshold. ### Claim 18 (depends on 16) The non-transitory storage medium of claim 16, further comprising optimizing the hyperparameters individually or as a group. ### Claim 19 (depends on 11) The non-transitory storage medium of claim 11, further comprising adjusting the hyperparameters while aligning the input file or prior to aligning the input file. ### Claim 20 (depends on 11) The non-transitory storage medium of claim 11, further comprising generating multiple compressed files, wherein each of the compressed files is associated with different hyperparameters.