IP Library › Granted Patent US 12,748,960
Granted Patent B2
US 12,748,960 · App. 17/189,109 · Granted Sep 29, 2026

Analog hardware realization of neural networks

Inventors: Aleksandrs Timofejevs (Riga, LV); Boris Maslov (Newport Beach, CA); Nikolai Kovshov (Moscow, RU); Dmitri Godovskiy (Moscow, RU)
Assignee: PolyN Technology Limited
G06N3/065G06F1/3206G06F1/3287G06F30/39G06N3/044G06N3/049G06N3/0499G06N3/063G06N3/082G06N5/04
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 12,748,960
App. No.
17/189,109
Filed
Mar 1, 2021
Granted
Sep 29, 2026
Kind
B2
Examiner
TRAN, TAN H
Art Unit
2141
USPC
706/38
Abstract

Systems and methods are provided for analog hardware realization of neural networks. The method incudes obtaining a neural network topology and weights of a trained neural network. The method also includes transforming the neural network topology to an equivalent analog network of analog components. The method also includes computing a weight matrix for the equivalent analog network based on the weights of the trained neural network. Each element of the weight matrix represents a respective connection between analog components of the equivalent analog network. The method also includes generating a schematic model for implementing the equivalent analog network based on the weight matrix, including selecting component values for the analog components.

Claims (106)

1 . A method for hardware realization of neural networks, comprising:

obtaining a neural network topology and weights of a trained neural network;

transforming the neural network topology to an equivalent analog network of analog components corresponding to the neural network, including:

selecting a plurality of individual electronic analog components, wherein the plurality of individual analog components include components corresponding to analog neurons and components corresponding to connections between pairs of analog neurons;

computing a first weight matrix for the equivalent analog network based on the weights of the trained neural network, wherein each element of the weight matrix represents a respective connection between a respective pair of analog neurons;

generating a first schematic model for implementing the equivalent analog network as electronic circuitry based on the weight matrix, wherein:

generating the first schematic model includes generating a first resistance matrix from the weight matrix, each element of the first resistance matrix (i) representing a respective fixed resistance valuem, (ii) corresponding to a respective weight of the weight matrix, and (iii) defining a respective single predetermined signal pathway between a respective pair of analog neurons;

the first schematic model includes a first portion corresponding to topology of analog neurons and a second portion corresponding to connections between the analog neurons; and

generating the first schematic model includes:

selecting component values for each analog neuron; and

assigning fixed values to the connections based on the weight matrix;

obtaining new weights for an updated neural network that corresponds to the trained neural network after it has been re-trained;

computing a new weight matrix for an equivalent analog network corresponding to the updated neural network based on the new weights:

generating a new resistance matrix for the new weight matrix; and

generating a new schematic model for implementing the equivalent analog network of analog components corresponding to the updated neural network based on the new resistance matrix, so that:

a first set of one or more lithographic masks generated for fabricating the neural network includes one or more first lithographic masks corresponding to the first portion of the first schematic model and a second lithographic mask corresponding to the second portion of the first schematic model; and

a second set of one or more lithographic masks generated for fabricating the updated neural network includes the one or more first lithographic masks corresponding to the first portion of the first schematic model and a third lithographic mask corresponding to the second portion of the new schematic model, wherein:

the topology of analog neurons for the updated neural network is the same as the topology of analog neurons for the neural network prior to updating:

the new schematic model includes a third portion corresponding to topology of analog neurons and a fourth portion corresponding to connections between the analog neurons for the updated neural network;

the third portion of the new schematic model is the same as the first portion of the first schematic model; and

the fourth portion of the new schematic model is different from the second portion of the first schematic model.

2 . The method of claim 1 , further comprising:

obtaining new weights, for an updated neural network, wherein:

the new weights are different from the weights of the trained neural network; and

the updated neural network corresponds to the trained neural network after it has been re-trained;

computing a new weight matrix, for the equivalent analog network based on the new weights, wherein the new weight matrix is different from the weight matrix;

generating a new resistance matrix, for the new weight matrix, wherein the new resistance matrix is different from the resistance matrix; and

updating the first schematic model to implement the updated neural network with new weights, including:

assigning new fixed values to the new connections based on the new weight matrix; and

replacing the connections between the analog neurons in the second portion of the first schematic model with the new connections between the analog neurons.

3 . The method of claim 1 , wherein the neural network topology includes one or more layers of neurons, each layer of neurons computing respective outputs based on a respective mathematical function, and

transforming the neural network topology to the equivalent analog network of analog components comprises:

for each layer of the one or more layers of neurons:

identifying one or more function blocks, based on the respective mathematical function for the respective layer, wherein each function block has a respective schematic implementation with block outputs that conform to outputs of a respective mathematical function; and

generating a respective multilayer network of analog neurons based on arranging the one or more function blocks, wherein each analog neuron implements a respective function of the one or more function blocks, and each analog neuron of a first layer of the multilayer network is connected to one or more analog neurons of a second layer of the multilayer network.

4 . The method of claim 1 , wherein the neural network topology includes one or more layers of neurons, each layer of neurons computing respective outputs based on a respective mathematical function, and transforming the neural network topology to the equivalent analog network of analog components comprises:

decomposing a first layer of the neural network topology to a plurality of sub-layers, including decomposing a mathematical function corresponding to the first layer to obtain one or more intermediate mathematical functions, wherein each sub-layer implements an intermediate mathematical function; and

for each sub-layer of the first layer of the neural network topology:

selecting one or more sub-function blocks, based on a respective intermediate mathematical function for the respective sub-layer; and

generating a respective multilayer analog sub-network of analog neurons based on arranging the one or more sub-function blocks, wherein each analog neuron implements a respective function of the one or more sub-function blocks, and each analog neuron of a first layer of the multilayer analog sub-network is connected to one or more analog neurons of a second layer of the multilayer analog sub-network.

5 . The method of claim 4 , wherein the mathematical function corresponding to the first layer includes one or more weights, and decomposing the mathematical function includes adjusting the one or more weights such that combining the one or more intermediate functions results in the mathematical function.

6 . The method of claim 1 , further comprising:

generating an equivalent digital network of digital components for one or more output layers of the neural network topology; and

connecting output of one or more layers of the equivalent analog network to the equivalent digital network of digital components.

7 . The method of claim 1 , wherein the analog components include a plurality of operational amplifiers and a plurality of resistors, each operational amplifier represents an analog neuron of the equivalent analog network, and each resistor represents a connection between two analog neurons.

8 . The method of claim 7 , wherein selecting component values of the analog components includes performing a gradient descent method to identify possible resistance values for the plurality of resistors.

9 . The method of claim 1 , wherein the neural network topology includes one or more gated recurrent unit (GRU) or long short-term memory (LSTM) neurons, and transforming the neural network topology includes generating one or more signal delay blocks for each recurrent connection of the one or more GRU or LSTM neurons.

10 . The method of claim 9 , wherein the one or more signal delay blocks are activated at a frequency that matches a predetermined input signal frequency for the neural network topology.

11 . The method of claim 1 , wherein the neural network topology includes one or more layers of neurons that perform unlimited activation functions, and transforming the neural network topology includes applying one or more transformations selected from the group consisting of:

replacing the unlimited activation functions with limited activation; and

adjusting connections or weights of the equivalent analog network such that, for predetermined one or more inputs, a difference in output between the trained neural network and the equivalent analog network is minimized.

12 . The method of claim 1 , further comprising:

generating lithographic masks for fabricating a circuit implementing the equivalent analog network of analog components corresponding to the neural network based on the first resistance matrix; and

generating lithographic masks for fabricating a circuit implementing the equivalent analog network of analog components corresponding to the updated neural network based on the new resistance matrix.

13 . The method of claim 1 , wherein the trained neural network is trained using software simulations to generate the weights.

14 . The method of claim 1 , further comprising fabricating an electronic circuit based on the first set of one or more lithographic masks and the second lithographic masks using a lithographic process.

15 . The method of claim 1 , further comprising fabricating an electronic circuit at least partially based on (i) reusing the first set of one or more lithographic masks for a plurality of analog neurons of the equivalent analog network, and (ii) using the new lithographic mask for one or more connections between the plurality of analog neurons, wherein the new weights (i) correspond to the one or more connections and (ii) have new corresponding fixed resistance values.

16 . The method of claim 1 , further comprising transforming the first schematic model into a physical manufacturable circuit layout representing an integrated electronic circuit's arrangement on a semiconductor substrate, wherein the circuit layout includes the selected individual electronic analog components and fixed component values in accordance with the schematic model.

17 . The method of claim 1 , wherein the second portion can be updated, independently of the first portion, to have new fixed values.

18 . The method of claim 1 , wherein the second lithographic mask can be replaced, independently of the first set of one or more lithographic masks, by a new lithographic mask that includes new fixed values.

19 . A system for hardware realization of neural networks, comprising:

one or more processors;

memory;

wherein the memory stores one or more programs configured for execution by the one or more processors, and the one or more programs comprising instructions for:

obtaining a neural network topology and weights of a trained neural network;

transforming the neural network topology to an equivalent analog network of analog components corresponding to the neural network, including:

selecting a plurality of individual electronic analog components, wherein the plurality of individual analog components include components corresponding to analog neurons and components corresponding to connections between pairs of analog neurons;

computing a first weight matrix for the equivalent analog network based on the weights of the trained neural network, wherein each element of the weight matrix represents a respective connection between a respective pair of analog neurons;

generating a first schematic model for implementing the equivalent analog network as electronic circuitry based on the weight matrix, wherein:

generating the first schematic model includes generating a first resistance matrix from the weight matrix, each element of the first resistance matrix (i) representing a respective fixed resistance value, (ii) corresponding to a respective weight of the weight matrix, and (iii) defining a respective single predetermined signal pathway between a respective pair of analog neurons:

the first schematic model includes a first portion corresponding to topology of analog neurons and a second portion corresponding to connections between the analog neurons; and

generating the first schematic model includes:

 selecting component values for each analog neuron; and

 assigning fixed values to the connections based on the weight matrix;

obtaining new weights for an updated neural network that corresponds to the trained neural network after it has been re-trained;

computing a new weight matrix for an equivalent analog network corresponding to the updated neural network based on the new weights;

generating a new resistance matrix for the new weight matrix; and

generating a new schematic model for implementing the equivalent analog network of analog components corresponding to the updated neural network based on the new resistance matrix, so that:

a first set of one or more lithographic masks generated for fabricating the neural network includes one or more first lithographic masks corresponding to the first portion of the first schematic model and a second lithographic mask corresponding to the second portion of the first schematic model; and

a second set of one or more lithographic masks generated for fabricating the updated neural network includes the one or more first lithographic masks corresponding to the first portion of the first schematic model and a third lithographic mask corresponding to the second portion of the new schematic model, wherein:

the topology of analog neurons for the updated neural network is the same as the topology of analog neurons for the neural network prior to updating:

the new schematic model includes a third portion corresponding to topology of analog neurons and a fourth portion corresponding to connections between the analog neurons for the updated neural network;

the third portion of the new schematic model is the same as the first portion of the first schematic model; and

the fourth portion of the new schematic model is different from the second portion of the first schematic model.

20 . The system of claim 19 , wherein generating the schematic model includes generating a resistance matrix from the weight matrix, each element of the resistance matrix (i) representing a respective fixed resistance value (ii) corresponding to a respective weight of the weight matrix, and (iii) defining a respective single predetermined signal pathway between a respective pair of analog neurons.

21 . A non-transitory computer readable storage medium storing one or more programs configured for execution by a computer system having one or more processors, the one or more programs comprising instructions for:

obtaining a neural network topology and weights of a trained neural network;

transforming the neural network topology to an equivalent analog network of analog components corresponding to the neural network, including:

selecting a plurality of individual electronic analog components, wherein the plurality of individual analog components include components corresponding to analog neurons and components corresponding to connections between pairs of analog neurons;

computing a first weight matrix for the equivalent analog network based on the weights of the trained neural network, wherein each element of the weight matrix represents a respective connection between a respective pair of analog neurons;

generating a first schematic model for implementing the equivalent analog network as electronic circuitry based on the weight matrix, wherein:

generating the first schematic model includes generating a first resistance matrix from the weight matrix, each element of the first resistance matrix (i) representing a respective fixed resistance value, (ii) corresponding to a respective weight of the weight matrix, and (iii) defining a respective single predetermined signal pathway between a respective pair of analog neurons;

the first schematic model includes a first portion corresponding to topology of analog neurons and a second portion corresponding to connections between the analog neurons; and

generating the first schematic model includes:

selecting component values for each analog neuron; and

assigning fixed values to the connections based on the weight matrix;

obtaining new weights for an updated neural network that corresponds to the trained neural network after it has been re-trained;

computing a new weight matrix for an equivalent analog network corresponding to the updated neural network based on the new weights:

generating a new resistance matrix for the new weight matrix; and

generating a new schematic model for implementing the equivalent analog network of analog components corresponding to the updated neural network based on the new resistance matrix, so that:

a first set of one or more lithographic masks generated for fabricating the neural network includes one or more first lithographic masks corresponding to the first portion of the first schematic model and a second lithographic mask corresponding to the second portion of the first schematic model; and

a second set of one or more lithographic masks generated for fabricating the updated neural network includes the one or more first lithographic masks corresponding to the first portion of the first schematic model and a third lithographic mask corresponding to the second portion of the new schematic model, wherein:

the topology of analog neurons for the updated neural network is the same as the topology of analog neurons for the neural network prior to updating;

the new schematic model includes a third portion corresponding to topology of analog neurons and a fourth portion corresponding to connections between the analog neurons for the updated neural network;

the third portion of the new schematic model is the same as the first portion of the first schematic model; and

the fourth portion of the new schematic model is different from the second portion of the first schematic model.

Assignments (3)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 7, 2021
From: TIMOFEJEVS, ALEKSANDRS; MASLOV, BORIS; KOVSHOV, NIKOLAI; GODOVSKIY, DMITRI
To: POLYN TECHNOLOGY LIMITED
Reel/Frame 056782/0977 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Mar 3, 2021
From: TIMOFEJEVS, ALEKSANDRS; MASLOV, BORIS
To: POLYN TECHNOLOGY LIMITED
Reel/Frame 055484/0627 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Mar 3, 2021
From: TIMOFEJEVS, ALEKSANDRS; MASLOV, BORIS; KOVSHOV, NIKOLAI; GODOVSKIY, DMITRI
To: POLYN TECHNOLOGY LIMITED
Reel/Frame 055484/0649 →
Continuity (3)
Continuation In Part PCTEP2020067800 · Jun 25, 2020
Continuation PCTRU2020000306 · Jun 25, 2020
Related Publication 20210406661A1 · Dec 30, 2021
References Cited (179)
US 1080471A · Olsen · 1913 [cited by applicant]
US 3628053A · Weiss · 1971 [cited by applicant]
US 5047655A · Chambost et al. · 1991 [cited by applicant]
US 5315163A · Karinthi · 1994 [cited by applicant]
US 5361327A · Takahashi · 1994 [cited by applicant]
US 5422982A · Pernisz · 1995 [cited by applicant]
US 6507641B1 · Kondo · 2003 [cited by applicant]
US 7966992B2 · Glugla et al. · 2011 [cited by applicant]
US 9275328B1 · Minkovich et al. · 2016 [cited by applicant]
US 9853025B1 · Yang · 2017 [cited by applicant]
US 10090005B2 · Rumberg et al. · 2018 [cited by applicant]
US 10157629B2 · Van Der Made et al. · 2018 [cited by applicant]
US 10217512B1 · Leobandung · 2019 [cited by examiner]
US 10825536B1 · Yang · 2020 [cited by examiner]
US 10970441B1 · Zhang · 2021 [cited by examiner]
US 11092130B2 · Senda et al. · 2021 [cited by applicant]
US 11126897B2 · Cai et al. · 2021 [cited by applicant]
US 11429900B1 · Alvarenga Marinelli · 2022 [cited by applicant]
US 12301243B2 · Urban · 2025 [cited by applicant]
US 20010000427A1 · Miller · 2001 [cited by examiner]
US 20050258513A1 · Eshun · 2005 [cited by applicant]
US 20060166107A1 · Chen · 2006 [cited by examiner]
US 20100106044A1 · Linderman · 2010 [cited by applicant]
US 20130329524A1 · Lee et al. · 2013 [cited by applicant]
US 20140038674A1 · Srinivasan et al. · 2014 [cited by applicant]
US 20140164611A1 · Molettiere et al. · 2014 [cited by applicant]
US 20150120629A1 · Matsuoka et al. · 2015 [cited by applicant]
US 20160283842A1 · Pescianschi · 2016 [cited by applicant]
US 20160328642A1 · Himebaugh et al. · 2016 [cited by applicant]
US 20170017879A1 · Kataeva et al. · 2017 [cited by applicant]
US 20170056722A1 · Singh et al. · 2017 [cited by applicant]
US 20170060224A1 · Cao et al. · 2017 [cited by applicant]
US 20170083352A1 · Chiou et al. · 2017 [cited by applicant]
US 20170140262A1 · Wilson et al. · 2017 [cited by applicant]
US 20170169327A1 · Nestler et al. · 2017 [cited by applicant]
US 20170176287A1 · Ito et al. · 2017 [cited by applicant]
US 20170193395A1 · Limonad et al. · 2017 [cited by applicant]
US 20170249445A1 · Devries et al. · 2017 [cited by applicant]
US 20170300806A1 · Lee · 2017 [cited by examiner]
US 20170340292A1 · Rossi et al. · 2017 [cited by applicant]
US 20180018553A1 · Bach · 2018 [cited by applicant]
US 20180031605A1 · Li et al. · 2018 [cited by applicant]
US 20180091240A1 · Noda · 2018 [cited by applicant]
US 20180161623A1 · Nejezchleb · 2018 [cited by applicant]
US 20180197485A1 · Jose · 2018 [cited by applicant]
US 20180293736A1 · Rahimi et al. · 2018 [cited by applicant]
US 20180338223A1 · Park · 2018 [cited by applicant]
US 20180356771A1 · Basu · 2018 [cited by applicant]
US 20180357532A1 · Inoue · 2018 [cited by applicant]
US 20180357533A1 · Inoue · 2018 [cited by applicant]
US 20180364785A1 · Hu et al. · 2018 [cited by applicant]
US 20190026625A1 · Vorenkamp · 2019 [cited by examiner]
US 20190034790A1 · Busch et al. · 2019 [cited by applicant]
US 20190034791A1 · Busch et al. · 2019 [cited by applicant]
US 20190069795A1 · Kiranya · 2019 [cited by applicant]
US 20190104951A1 · Valys · 2019 [cited by applicant]
US 20190251279A1 · Emberson et al. · 2019 [cited by applicant]
US 20190251426A1 · Busch et al. · 2019 [cited by applicant]
US 20200026992A1 · Zhang · 2020 [cited by applicant]
US 20200043477A1 · Busch et al. · 2020 [cited by applicant]
US 20200046240A1 · Angle et al. · 2020 [cited by applicant]
US 20200053299A1 · Zhang et al. · 2020 [cited by applicant]
US 20200073483A1 · Berenzweig et al. · 2020 [cited by applicant]
US 20200089969A1 · Lakshmi Narayanan et al. · 2020 [cited by applicant]
US 20200105287A1 · Chang et al. · 2020 [cited by applicant]
US 20200110991A1 · Kataeva et al. · 2020 [cited by applicant]
US 20200166922A1 · Cella et al. · 2020 [cited by applicant]
US 20200167448A1 · Modarresi · 2020 [cited by applicant]
US 20200170549A1 · Baykaner et al. · 2020 [cited by applicant]
US 20200202206A1 · Rummens et al. · 2020 [cited by applicant]
US 20200211566A1 · Kang et al. · 2020 [cited by applicant]
US 20200222010A1 · Howard · 2020 [cited by applicant]
US 20200310541A1 · Reisman et al. · 2020 [cited by applicant]
US 20200311535A1 · Shrivastava · 2020 [cited by applicant]
US 20200364548A1 · Chang · 2020 [cited by applicant]
US 20200380192A1 · MacRae · 2020 [cited by applicant]
US 20210125049A1 · Khwa · 2021 [cited by applicant]
US 20210150328A1 · Dasalukunte et al. · 2021 [cited by applicant]
US 20210256988A1 · Mauri et al. · 2021 [cited by applicant]
US 20210270244A1 · He et al. · 2021 [cited by applicant]
US 20210278309A1 · Lefebvre et al. · 2021 [cited by applicant]
US 20210326393A1 · Aggarwal et al. · 2021 [cited by applicant]
US 20210335379A1 · Sugiyama et al. · 2021 [cited by applicant]
US 20210365778A1 · Dey et al. · 2021 [cited by applicant]
US 20210374546A1 · Tsai · 2021 [cited by examiner]
US 20210389817A1 · Spinelli et al. · 2021 [cited by applicant]
US 20220012564A1 · Zjajo et al. · 2022 [cited by applicant]
US 20220027718A1 · Alam et al. · 2022 [cited by applicant]
US 20220028051A1 · Asano · 2022 [cited by applicant]
US 20220083865A1 · Riazi · 2022 [cited by applicant]
US 20220172050A1 · Dalli et al. · 2022 [cited by applicant]
US 20220188636A1 · Pham et al. · 2022 [cited by applicant]
US 20220222513A1 · Paramasivam · 2022 [cited by applicant]
US 20220249906A1 · Phillips et al. · 2022 [cited by applicant]
US 20220253675A1 · Ren · 2022 [cited by applicant]
US 20220327370A1 · Huynh · 2022 [cited by applicant]
US 20230148326A1 · Papel et al. · 2023 [cited by applicant]
US 20230206036A1 · Labreuche et al. · 2023 [cited by applicant]
US 20230335118A1 · Huang et al. · 2023 [cited by applicant]
US 20230363703A1 · Canavan et al. · 2023 [cited by applicant]
US 20230379757A1 · Mozaffari et al. · 2023 [cited by applicant]
US 20240193760A1 · Fukuda · 2024 [cited by applicant]
US 20240404053A1 · Schwab et al. · 2024 [cited by applicant]
CN 102456157A · 2012 [cited by applicant]
CN 108431895A · 2018 [cited by applicant]
CN 111052154A · 2020 [cited by applicant]
EP 0432462A1 · 1991 [cited by applicant]
EP 3663988A1 · 2020 [cited by applicant]
JP 2003021032A · 2003 [cited by applicant]
JP 2019003464A · 2019 [cited by applicant]
JP 2019003546A · 2019 [cited by applicant]
JP 2019016159A · 2019 [cited by applicant]
JP 7371235B2 · 2022 [cited by applicant]
KR 20190052587A · 2019 [cited by applicant]
KR 102120756B1 · 2020 [cited by applicant]
KR 102191736B1 · 2020 [cited by applicant]
WO WO2006104144A1 · 2006 [cited by applicant]
WO WO2020082080A1 · 2020 [cited by applicant]
WO WO2021170735A1 · 2021 [cited by applicant]
WO WO2021262023A1 · 2021 [cited by applicant]
WO WO2023167607A1 · 2023 [cited by applicant]
Chen et al. (Accelerator-friendly Neural-network Training: Learning Variations and Defects in RRAM Crossbar; Design, Automation & Test in Europe Conference & Exhibition; published 2017, pp. 19-24) (Year: 2017). [cited by examiner]
Liew et al. (Bounded activation functions for enhanced training stability of deep neural networks on visual pattern recognition problems; Neurocomputing; published 2016, pp. 718-734). (Year: 2016). [cited by examiner]
Himavathi et al. (Feedforward Neural Network Implementation in FPGA Using Layer Multiplexing for Effective Resource Utilization, IEEE Transactions on Neural Networks, vol. 18, No. 3, May 2007, pp. 880-888) (Year: 2007). [cited by examiner]
Liu et al. (Training an Artificial Neural Network with Op-amp Integrators Based Analog Circuits; IEEE CSAA Guidance, Navigation and Control Conference (CGNCC); published 2018, pp. 1-6). (Year: 2018). [cited by examiner]
Wang et al. (Prandtl-Ishlinskii Modeling for Giant Magnetostrictive Actuator Based on Internal Time-Delay Recurrent Neural Network; Published in: IEEE Transactions on Magnetics ( vol. 54, Issue: 11, Nov. 2018); pp. 1-4)… [cited by examiner]
Cadence Design Systems: “From TensorFlow to RTL in three months,” Youtube, Nov. 14, 2018, XP054981600, retrieved from the Internet: URL: XP054981600, https://www.youtube.com > watch, 2 pgs. [cited by applicant]
Cadence: “Engineering Change Orders,” Dec. 8, 2019, XP055789511, Retrieved from the Internet: URL:https://web.archive.org/web/20191208162944if_https://www.cadence.com/content/dam/cadence-www/global/en_US/documents/tools… [cited by applicant]
Guyue Huang et al., “Machine Learning for Electronic Design Automation: A Survey,” arxiv.org, Cornell University Library, Mar. 8, 2021, arXiv:2102-03357v2, 44 pgs. [cited by applicant]
Monsen, Analog neural network-based helicopter gearbox health monitoring system, Acoustical Society of America, vol. 96, No. Dec. 6, 1995, 15 pgs. [cited by applicant]
Polyn Technology Limited, International Search Report and Written Opinion, PCT/EP2020/067800, Apr. 12, 2021, 10 pgs. [cited by applicant]
Polyn Technology Limited, International Preliminary Report on Patentability, PCT/EP2020/067800, Dec. 13, 2022, 8 pgs. [cited by applicant]
Polyn Technology Limited, International Search Report and Written Opinion, PCT/RU2020/000306, Mar. 4, 2021, 12 pgs. [cited by applicant]
Polyn Technology Limited, International Preliminary Report on Patentability, PCT/RU2020/000306, Dec. 13, 2022, 9 pgs. [cited by applicant]
Polyn Technology Limited, International Search Report and Written Opinion, PCT/US2021/058266, Feb. 18, 2022, 16 pgs. [cited by applicant]
Ronao, Charissa Ann, Human activity recognition with smartphone sensors using deep learning neural networks, Expert Systems with Applications 59, (2016) 235-244, 10 pgs. [cited by applicant]
Timofejevs, Office Action, U.S. Appl. No. 17/733,932, Apr. 21, 2023, 20 pgs. [cited by applicant]
Amelia Dalton, “TensorFlow to RTL with High-Level Synthesis—Cadence Design Systems”, EE Journal, Apr. 17, 2020, 1 page. [cited by applicant]
Andrew Muscat et al., “Electromagnetic Vibrational Energy Harvesters: A Review”, Sensors 2022, vol. 22, No. 15, Jul. 25, 2022, 17 pgs. [cited by applicant]
Anonymous, “Thoughts on Interfacing Piezo Vibration Sensor”, Aug. 22, 2013, 7 pgs., Retrieved from the Internet: https://scienceprog.com/thoughts-on-interfacing-piezo-vibration-sensor/. [cited by applicant]
Fredrik Sandin et al., “Synaptic Delays for Insect-Inspired Temporal Feature Detection in Dynamic Neuromorphic Processors”, Embedded Intelligent Systems Lab (EISLAB), Frontiers in Neuroscience, vol. 14, Article 150, Feb… [cited by applicant]
Hossam Abdelbaki et al., “Analog Hardware Implementation of the Random Neural Network Model”, Neural Networks, IJCNN 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Jul. 24, 2000, 5 pgs. [cited by applicant]
Hyeong-Ju Kang, “Accelerator-Aware Pruning for Convolutional Neural Networks”, arXiv, Sep. 5, 2020, 11 pgs. [cited by applicant]
Kaveri Mahapatra et al., “Power System Disturbance Classification with Online Event-Driven Neuromorphic Computing”, arXiv, Dec. 15, 2020, 11 pgs. [cited by applicant]
Khoa Van Pham et al., “Partial-Gated Memristor Crossbar for Fast and Power-Efficient Defect-Tolerant Training”, Micromachines, Apr. 13, 2019, 18 pgs., Retrieved from the Internet: www.mdpi.com/journal/micromachines. [cited by applicant]
M.V. Valueva et al., “Application of the Residue Number System to Reduce Hardware Costs of the Convolutional Neural Network Implementation”, Mathematics and Computers in Simulation, vol. 177 Nov. 2020, 8 pgs. [cited by applicant]
P. Sibi et al., “Analysis of Different Activation Functions Using Back Propagation Neural Networks”, Journal of Theoretical and Applied Information Technology, Jan. 31, 2013, vol. 47, No. 3, 5 pgs. [cited by applicant]
Paul M. Solomon, “Analog Neuromorphic Computing Using Programmable Resistor Arrays”, Solid State Electronics, vol. 155, May 2019, 10 pgs. [cited by applicant]
Peng Yao et al., “Fully Hardware-implemented Memristor Convolutional Neural Network”, Nature, vol. 577, Jan. 30, 2020, 21 pgs. [cited by applicant]
PolyN Technology Limited, Supplementary International Search Report, PCT/RU2020/000306, Jul. 26, 2022, 15 pgs. [cited by applicant]
PolyN Technology Limited, International Preliminary Report on Patentability, PCT/US2021/058266, Sep. 12, 2023, 13 pgs. [cited by applicant]
PolyN Technology Limited, International Search Report and Written Opinion, PCT/RU2022/000064, Dec. 1, 2022, 9 pgs. [cited by applicant]
PolyN Technology Limited, International Preliminary Report on Patentability, PCT/RU2022/000064, Sep. 10, 2024, 8 pgs. [cited by applicant]
PolyN Technology Limited, International Search Report and Written Opinion, PCT/US2023/022139, Oct. 2, 2023, 13 pgs. [cited by applicant]
PolyN Technology Limited, International Preliminary Report on Patentability, PCT/US2023/022139, Nov. 7, 2024, 9 pgs. [cited by applicant]
PolyN Technology Limited, International Search Report and Written Opinion, PCT/US2024/028993, Sep. 30, 2024, 11 pgs. [cited by applicant]
PolyN Technology Limited, International Search Report and Written Opinion, PCT/US2023/031692, Dec. 6, 2023, 15 pgs. [cited by applicant]
PolyN Technology Limited, International Preliminary Report on Patentability, PCT/US2023/031692, Mar. 1, 2025, 12 pgs. [cited by applicant]
Qiang Wang et al., “Compressive Sensing Reconstruction for Vibration Signals Based on the Improved Fast Iterative Shrinkage-Thresholding Algorithm”, Measurement, vol. 142, Aug. 31, 2019, 9 pgs. [cited by applicant]
Renée St. Amant et al., “General-Purpose Code Acceleration with Limited-Precision Analog Computation”, Proceedings of the 41st International Symposium on Computer Architecture, 2014, 12 pgs. [cited by applicant]
Renesas, “Introduction to Electronic Circuits: Op-Amp Comparator”, Tutorial, published 2015, 22 pgs. [cited by applicant]
Sharon Shea: “What is LPWAN (Low-Power Wide Area Network) ?: Definition from Tech Target”, Sep. 30, 2017, 4 pgs., Retrieved from the Internet: https://www.techtarget.com/iotagenda/definition/LPWAN-low-power-wide-area-ne… [cited by applicant]
Siddharth Sharma et al., “Activation Functions in Neural Networks”, International Journal of Engineering Applied Sciences and Technology, 2020, vol. 4, Issue 12, ISSN No. 2455-2143, 7 pgs. [cited by applicant]
Talha Furkan Canan et al., “4-Input NAND and NOR Gates Based on Two Ambipolar Schottky Barrier FinFETs”, 2020 27th IEEE International Conference on Electronics, Circuits and Systems (ICECS), 4 pgs. [cited by applicant]
Tiago Oliveira Weber et al., “Amplifier-based MOS Analog Neural Network Implementation and Weights Optimization”, 2019 32nd Symposium on Integrated Circuits and Systems Design (SBCCI), ACM, Aug. 26, 2019, 6 pgs. [cited by applicant]
Yun Long et al., “ReRAM-Based Processing-in-Memory Architecture for Recurrent Neural Network Acceleration”, Abstract, IEEE Journals & Magazine, Jul. 3, 2018, 4 pgs. [cited by applicant]
Chakma et al. (A Mixed-Signal Approach to Memristive Neuromorphic System Design; Published in: 2017 IEEE 60th International Midwest Symposium on Circuits and Systems (MWSCAS), pp. 1-4). (Year: 2017). [cited by applicant]
Chapeton et al. (Effects of homeostatic constraints on associative memory storage and synaptic connectivity of cortical circuits, Frontiers in Computational Neuroscience, vol. 9, Article 74, published Jun. 18, 2015, pp.… [cited by applicant]
Cheng et al. (MS-LSTM: a Multi-Scale LSTM Model for BGP Anomaly Classification, 2016 IEEE 24th International Conference on Network Protocols (ICNP), pp. 1-6) (Year: 2016). [cited by applicant]
Lin et al. (Learning the sparsity for ReRAM: mapping and pruning sparse neural network for ReRAM based accelerator; ASPDAC ' 19: Proceedings of the 24th Asia and South Pacific Design Automation Conference; published 201… [cited by applicant]
Polyn Technology Limited, International Search Report and Written Opinion, PCT/RU2021/000630, Mar. 17, 2022, 11 pgs. [cited by applicant]
Sang et al. (Micro Hand Gesture Recognition System Using Ultrasonic Active Sensing, Published in: IEEE Sensors Journal vol. 19, Issue: 18, Sep. 15, 2019), pp. 1-9) (Year: 2019). [cited by applicant]
Timofejevs, Office Action, U.S. Appl. No. 17/198,198, Apr. 23, 2024, 20 pgs. [cited by applicant]
Timofejevs, Office Action, U.S. Appl. No. 17/199,373, May 9, 2024, 32 pgs. [cited by applicant]
Timofejevs, Office Action, U.S. Appl. No. 17/199,407, Jun. 6, 2024, 25 pgs. [cited by applicant]
Timofejevs, Office Action, U.S. Appl. No. 17/199,422, Jul. 10, 2024, 19 pgs. [cited by applicant]
Timofejevs, Office Action, U.S. Appl. No. 17/200,707, Aug. 28, 2024, 38 pgs. [cited by applicant]
Eric K. Garcia et al., “Completely Lazy Learning,” IEEE Transactions on Knowledge and Data Engineering, IEEE Service Center, vol. 22, No. 9, Sep. 1, 2010, 12 pgs. [cited by applicant]
PolyN Technology Limited, Extended European Search Report, EP 22930051.2, Oct. 22, 2025, 10 pgs. [cited by applicant]