IP Library › Granted Patent US 12,260,913
Granted Patent B2
US 12,260,913 · App. 18/166,484 · Granted Mar 25, 2025

Hyperdimensional computing device

Inventors: Yu-Hsuan Lin (Taichung, TW); Po-Hao Tseng (Taichung, TW)
Assignee: MACRONIX International Co., Ltd.
G11C16/08G11C16/0483G11C16/26
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,260,913
App. No.
18/166,484
Granted
Mar 25, 2025
Kind
B2
Abstract

A hyperdimensional computing device includes a non-volatile memory cell array and a first operation circuit. The non-volatile memory cell array is coupled to a plurality of first word lines. The non-volatile memory cell array has a plurality of memory cell groups, a plurality of first memory cells of each of the memory cell groups are coupled to a same first word line of the first word lines, and the memory cell groups respectively store a plurality of data vectors. The first operation circuit receives at least one of the data vectors through bit lines and generates a bundled data vector according to the at least one of the data vectors.

Claims (22)

1. A hyperdimensional computing device, comprising:

a non-volatile memory cell array, coupled to a plurality of first word lines, the memory cell array having a plurality of first memory cell groups, a plurality of first memory cells of each of the first memory cell groups being coupled to a same first word line of the first word lines, the first memory cell groups respectively storing a plurality of data vectors; and

a first operation circuit, coupled to a plurality of first bit lines of the non-volatile memory cell array, receiving at least one of the data vectors through the bit lines, and generating a bundled data vector according to the at least one of the data vectors,

wherein each of the data vectors is an n-gram vector, and the data vectors comprise all permutations and combinations of a plurality of characters.

2. The hyperdimensional computing device according to claim 1 , further comprising an encoder generating the n-gram vectors and respectively storing the n-gram vectors into the first memory cell groups.

3. The hyperdimensional computing device according to claim 1 , further comprising:

a word line driver, turning on a plurality of selected word lines of the first word lines according to input data.

4. The hyperdimensional computing device according to claim 3 , wherein the word line driver turns on each of the selected word lines one by one.

5. The hyperdimensional computing device according to claim 4 , wherein the operation circuit is a counter.

6. The hyperdimensional computing device according to claim 3 , wherein the word line driver synchronously turns on the selected word lines.

7. The hyperdimensional computing device according to claim 6 , wherein the operation circuit is a sensing amplifier.

8. A hyperdimensional computing device, comprising:

a non-volatile memory cell array, coupled to a plurality of first word lines, the memory cell array having a plurality of first memory cell groups, a plurality of first memory cells of each of the first memory cell groups being coupled to a same first word line of the first word lines, the first memory cell groups respectively storing a plurality of data vectors; and

a first operation circuit, coupled to a plurality of first bit lines of the non-volatile memory cell array, receiving at least one of the data vectors through the bit lines, and generating a bundled data vector according to the at least one of the data vectors, wherein the data vectors respectively comprise a plurality of possible component information and a plurality of position information corresponding to a graphic image.

9. The hyperdimensional computing device according to claim 8 , wherein the operation circuit generates the bundled data vector according to a plurality of selected component information and the position information respectively corresponding to the selected component information.

10. The hyperdimensional computing device according to claim 8 , further comprising:

an associative memory array, coupled to the operation circuit and configured to store a plurality of classification vectors, the associative memory array having a plurality of second memory cell groups, a plurality of second memory cells of each of the second memory cell groups being respectively coupled to a plurality of second word lines, the second memory cell groups being respectively coupled to a plurality of second bit lines, wherein the associative memory array receives the bundled data vector and performs an in-memory-searching operation according to the bundled data vector; and

a second operation circuit, coupled to the associative memory to generate an optimal similar result according to the in-memory-searching operation.

11. The hyperdimensional computing device according to claim 10 , wherein the second operation circuit generates the optimal similar result according to a current value of each of the second word lines.

12. The hyperdimensional computing device according to claim 11 , wherein the second operation circuit generates the optimal similar result according to one of the second word lines having a maximum current value.

13. The hyperdimensional computing device according to claim 10 , wherein the associative memory array is a NAND flash memory or a NOR flash memory.

14. The hyperdimensional computing device according to claim 1 , wherein the non-volatile memory cell array is a NAND flash memory or a NOR flash memory.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 15, 2023
From: LIN, YU-HSUAN; TSENG, PO-HAO
To: MACRONIX INTERNATIONAL CO., LTD.
Reel/Frame 062713/0864 →
Continuity (1)
Related Publication 20240274199A1 · Aug 15, 2024
References Cited (11)
US 10460817B2 · Li et al. · 2019 [cited by applicant]
US 10762957B2 · Chen · 2020 [cited by examiner]
US 20120221775A1 · Kim · 2012 [cited by examiner]
US 20140143473A1 · Kim · 2014 [cited by examiner]
US 20190164600A1 · Castro · 2019 [cited by examiner]
CN 109213614 · 2022 [cited by applicant]
TW I718626 · 2021 [cited by applicant]
TW 202147311 · 2021 [cited by applicant]
TW 202211216 · 2022 [cited by applicant]
Lu Lu Ge et al., “Classification using Hyperdimensional Computing: A Review,” IEEE Circuits and Systems Magazine, Apr. 2020, pp. 1-16. [cited by applicant]
“Office Action of Taiwan Counterpart Application”, issued on Sep. 13, 2023, pp. 1-5. [cited by applicant]