# US 12,015,424 B2 · Network-based hyperdimensional system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,015,424 B2 - Document ID: US12015424B2 - Status: Granted - Kind: B2 - Application number: 17/895,177 - Filed: 2022-08-25 - Published: 2023-03-16 - Granted: 2024-06-18 - Earliest claimed: 2021-08-27 - Assignees: The Regents of the University of California - Applicants: The Regents of the University of California - Inventors: Mohsen Imani - Web page: https://patentgrep.com/patent/US12015424B2 ## CPC - H03M 13/1108: {Hard decision decoding, e.g. bit flipping, modified or weighted bit flipping} - H04L 1/0041: {Arrangements at the transmitter end} - H03M 13/1128: {Judging correct decoding and iterative stopping criteria other than syndrome check and upper limit for decoding iterations} - H03M 13/6312: {Error control coding in combination with data compression} - H04L 1/0045: {Arrangements at the receiver end} - H04L 1/0056: {Systems characterized by the type of code used (H04L1/08 takes precedence)} - H03M: CODING; DECODING; CODE CONVERSION IN GENERAL (using fluidic means F15C4/00; optical analogue/digital converters G02F7/00; coding, decoding or code conversion, specially adapted for particular applications, see the relevant subclasses, e.g. G01D, G01R, G06F, G06T, G09G, G10L, G11B, G11C, H04B, H04L, H04M, H04N; ciphering or deciphering for cryptography or other purposes involving the need for secrecy G09C) - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract Disclosed is a network-based hyperdimensional system having an encoder configured to receive input data and encode the input data using hyperdimensional computing to generate a hypervector having encoded data bits that represent the input data. The network-based hyperdimensional system further includes a decoder configured to receive the encoded data bits, decode the encoded data bits, and reconstruct the input data from the decoded data bits. In some embodiments, the encoder is configured for direct hyperdimensional learning on transmitted data with no need for data decoding by the decoder. ## Claims ### Claim 1 A network-based hyperdimensional system comprising: an encoder implemented in hardware and configured to: receive input data in a bitstream; divide the bit stream into chunks; generate chunk hypervectors from the chunks; generate a position hypervector for each chunk hypervector; bind each position hypervector with each associated chunk hypervector to generate bound hypervectors; and add the bound hypervectors together to generate encoded data bits; and a decoder implemented in hardware and configured to: receive the encoded data bits over a network; and decode the encoded data bits by comparing predetermined bit sequences with the encoded data bits until a convergence to a maximum match between the predetermined bit sequences and the encoded data bits occurs within a finite number of decoding iterations and thereby reconstruct the input data from the decoded data bits. ### Claim 2 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the finite number of decoding iterations is between 1 and 10. ### Claim 3 (depends on 1) The network-based hyperdimensional system of claim 1 wherein no error correction is performed by the decoder. ### Claim 4 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the decoder further comprises a hardware accelerator with content addressable memory (CAM) that stores the predetermined bit sequences in a look-up table. ### Claim 5 (depends on 4) The network-based hyperdimensional system of claim 4 wherein the CAM is configured for row-parallel search operations. ### Claim 6 (depends on 4) The network-based hyperdimensional system of claim 4 wherein the CAM is configured for searching hypervectors that have real and imaginary components. ### Claim 7 (depends on 6) The network-based hyperdimensional system of claim 6 wherein both the real and imaginary components are nonzero. ### Claim 8 (depends on 6) The network-based hyperdimensional system of claim 6 wherein the real and imaginary components are stored in adjacent memory cells of the CAM. ### Claim 9 (depends on 8) The network-based hyperdimensional system of claim 8 wherein the CAM is further configured to determine Hamming distance of both the real and imaginary components. ### Claim 10 (depends on 9) The network-based hyperdimensional system of claim 9 wherein the CAM is configured to determine the Hamming distances by sensing discharging current on match-lines that are CAM rows. ### Claim 11 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the encoded data represent a totality of the bit stream. ### Claim 12 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the encoder is configured to operate based on parameters that include but are not limited to chunk size C representing a total number of bits to be transmitted per layer, dimension D representing a number of channels to be transmitted, and a number of layers V to be encoded. ### Claim 13 (depends on 12) The network-based hyperdimensional system of claim 12 wherein the chunk size C multiplied by the number of layers V to be encoded is equal to the total number of bits to be transmitted. ### Claim 14 (depends on 13) The network-based hyperdimensional system of claim 13 wherein a coding rate of the encoder is equal to the total number of bits to be transmitted C×V divided by the dimension D. ### Claim 15 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the encoder and decoder hardware implementations include but are not limited to field programmable gate arrays, graphics processing units, central processing units and CAM. ### Claim 16 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the HDC is configured to perform single-pass training. ### Claim 17 (depends on 1) The network-based hyperdimensional system of claim 1 further comprising a learning module that is configured to perform classification and clustering of the input data. ### Claim 18 (depends on 1) The network-based hyperdimensional system of claim 1 wherein the encoder is configured to provide dynamic data compression during encoding of the input data.