IP Library Granted Patent US 12,657,442
Granted Patent B2
US 12,657,442 · App. 17/892,876 · Granted Jun 16, 2026

Neuromorphic synthesizer

Inventors: Pieter Vorenkamp (Laguna Beach, CA); Kurt F. Busch (Laguna Hills, CA); Stephen W. Bailey (Irvine, CA); Jeremiah H. Holleman, III (Davidson, NC)
Assignee: SYNTIANT
G06N3/063G06F30/327G06F30/36G06N3/0442G06N3/065G06N3/105H10D84/01
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 12,657,442
App. No.
17/892,876
Granted
Jun 16, 2026
Kind
B2
Abstract

Disclosed herein is a method for automatically generating an integrated circuit. The method includes receiving a behavioral description of at least the first layer of the neural network, converting the behavioral description of the first layer of the neural network into the computational graph, converting a computational graph to a circuit netlist based on a correlation of: (i) operations described in the computational graph, and (ii) an analog cell library including a plurality of predetermined circuit blocks that describe known neural network operations, generating a circuit layout that corresponds to at least a first layer of a neural network, and performing additional actions configured to cause generation of the integrated circuit based on the circuit layout. In some situations, the behavioral description defines an architecture of machine learning logic that represents at least a portion of the neural network. Additionally, in some situations, each cell of the integrated circuit includes a metal-oxide-semiconductor field-effect transistor (“MOSFET”).

Claims (38)

1 . A non-transitory computer readable storage medium having stored thereon instructions, the instructions being executable by one or more processors to perform actions comprising:

receiving a behavioral description of a first layer of a neural network comprising one or more corresponding layer types;

parsing the behavioral description using a logic engine and converting the behavioral description of a machine learning logic module to a computational graph representing a plurality of levels of operation that describe the first layer of the neural network to a circuit netlist based on a correlation of:

(i) operations described in the computational graph, and

(ii) an analog cell library including a plurality of predetermined circuit blocks that describe known neural network operations; and

generating a circuit layout that corresponds to at least the first layer of the neural network;

wherein the logic engine is configured to route circuit blocks of the one or more corresponding layer types.

2 . The non-transitory computer readable storage medium of claim 1 , wherein the instructions being executable by the one or more processors of to perform further actions comprising:

creating a corresponding connection that connects the circuit blocks implementing corresponding computational elements.

3 . The non-transitory computer readable storage medium of claim 1 , wherein the levels of operation comprise primitive and complex operations.

4 . The non-transitory computer readable storage medium of claim 1 , wherein generating the circuit layout includes converting syntax of the circuit netlist to a representation of a circuit layout.

5 . The non-transitory computer readable storage medium of claim 1 , wherein at least one circuit block is instantiated with size parameters.

6 . The non-transitory computer readable storage medium of claim 1 , wherein if the operations are not defined in the analog cell library, the conversion includes using a set of predetermined rules to generate the netlist.

7 . The non-transitory computer readable storage medium of claim 1 , wherein the instructions being executable by the one or more processors of to perform further actions comprising:

performing additional actions configured to cause generation of an integrated circuit based on the circuit layout.

8 . The non-transitory computer readable storage medium of claim 1 , wherein the behavioral description includes machine code being a set of one or more instructions that are executable by a computer processing unit (CPU).

9 . The non-transitory computer readable storage medium of claim 1 , wherein the behavioral description defines an architecture of machine learning logic that represents at least a portion of the neural network.

10 . The non-transitory computer readable storage medium of claim 7 , wherein the integrated circuit includes a plurality of cells, each cell including a metal-oxide-semiconductor field-effect transistor (“MOSFET”).

11 . The non-transitory computer readable storage medium of claim 1 , wherein converting the computational graph to the circuit netlist further includes:

for each component element implementing an operation included in the computational graph, selecting one of the plurality of predetermined circuit blocks of the analog cell library that implements the operation.

12 . A method for automatically generating an integrated circuit, the method comprising:

receiving a behavioral description of a first layer of a neural network comprising one or more corresponding layer types;

parsing the behavioral description using a logic engine and converting the behavioral description of a machine learning logic module to a computational graph representing a plurality of levels of operation that describe the first layer of the neural network to a circuit netlist based on a correlation of:

(i) operations described in the computational graph, and

(ii) an analog cell library including a plurality of predetermined circuit blocks that describe known neural network operations; and

generating a circuit layout that corresponds to at least the first layer of the neural network;

wherein the logic engine is configured to route circuit blocks of the one or more corresponding layer types.

13 . The method of claim 12 , wherein the instructions being executable by the one or more processors of to perform further actions comprising:

creating a corresponding connection that connects the circuit blocks implementing corresponding computational elements.

14 . The method of claim 12 , wherein the levels of operation comprise primitive and complex operations.

15 . The method of claim 12 , wherein generating the circuit layout includes converting syntax of the circuit netlist to a representation of a circuit layout.

16 . The method of claim 12 , wherein at least one circuit block is instantiated with size parameters.

17 . The method of claim 12 , wherein if the operations are not defined in the analog cell library, the conversion includes using a set of predetermined rules to generate the netlist.

18 . The method of claim 12 , wherein the instructions being executable by the one or more processors of to perform further actions comprising:

performing additional actions configured to cause generation of an integrated circuit based on the circuit layout.

19 . The method of claim 12 , wherein the behavioral description includes machine code being a set of one or more instructions that are executable by a computer processing unit (CPU).

20 . The method of claim 12 , wherein converting the computational graph to the circuit netlist further includes:

for each component element implementing an operation included in the computational graph, selecting one of the plurality of predetermined circuit blocks of the analog cell library that implements the operation.

Assignments (2)
SECURITY INTEREST Recorded Dec 27, 2024
From: SYNTIANT CORP.; PILOT AI LABS, INC.; SYNTIANT TAIWAN LLC; SYNTIANT HOLDINGS LLC
To: OCEAN II PLO LLC
Reel/Frame 069687/0757 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 22, 2022
From: VORENKAMP, PIETER; BUSCH, KURT F.; BAILEY, STEPHEN W.; HOLLEMAN, JEREMIAH H., III
To: SYNTIANT
Reel/Frame 060861/0365 →
Continuity (3)
Continuation 16039142 · Jul 18, 2018
Provisional Application 62534188 · Jul 18, 2017
Related Publication 20220414439A1 · Dec 29, 2022
References Cited (8)
US 6424959B1 · Bennett et al. · 2002 [cited by applicant]
US 20110119214A1 · Breitwisch et al. · 2011 [cited by applicant]
US 20140282180A1 · Orofino · 2014 [cited by applicant]
US 20150286766A1 · Singh et al. · 2015 [cited by applicant]
US 20230252215A1 · Zhang · 2023 [cited by examiner]
PCT Search Report, PCT Application No. PCT/US2018/042749, dated Nov. 28, 2018. [cited by applicant]
Joshi, M., “Implementation of an Integrated Artificial Neural Network trained with Back Propagation Algorithm, Master of Technology in VLSI Design”, 2012, accessed from http:I/117.203.246.91:8080/jspul/bitstream/10266/1… [cited by applicant]
Rampasek, L. et al., “TensorFlow: Biology's Gateway to Deep Learning?”, Cell Systems, vol. 2, Issue 1, 2016, pp. 12-14, ISSN 2405-4712, https://doi.org/10.1016/j.cels.2016.01.009 (Year: 2016). [cited by applicant]