Prosecution Insights
Last updated: November 20, 2026
Application No. 17/893,462

SYSTEMS AND METHODS FOR A STORAGE BIT IN AN ARTIFICIAL NEURAL NETWORK

Non-Final OA §103
Filed
Aug 23, 2022
Priority
Mar 07, 2022 — provisional 63/268,953
Examiner
SAX, STEVEN PAUL
Art Unit
2100
Tech Center
2100 — Computer Architecture & Software
Assignee
Everspin Technologies Inc.
OA Round
3 (Non-Final)
69%
Grant Probability
Favorable
3-4
OA Rounds
0m
Est. Remaining
99%
With Interview

Examiner Intelligence

Grants 69% — above average
69%
Career Allowance Rate
325 granted / 470 resolved
+14.1% vs TC avg
Strong +46% interview lift
Without
With
+45.8%
Interview Lift
resolved cases with interview
Typical timeline
4y 1m
Avg Prosecution
15 currently pending
Career history
493
Total Applications
across all art units

Statute-Specific Performance

§101
11.3%
-28.7% vs TC avg
§103
58.1%
+18.1% vs TC avg
§102
12.1%
-27.9% vs TC avg
§112
10.2%
-29.8% vs TC avg
Black line = Tech Center average estimate • Based on career data from 470 resolved cases

Office Action

§103
DETAILED ACTION Notice of Pre-AIA or AIA Status 1. The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . 2. The RCE (Request for Continued Examination) filed 9/25/26 has been entered. 3. Accordingly, the amendment filed 9/21/26 has been entered. Claims 1-19 and 21 are pending. Claim Rejections - 35 USC § 103 4. In the event the determination of the status of the application as subject to AIA 35 U.S.C. §102 and §103 (or as subject to pre-AIA 35 U.S.C. §102 and §103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status. The following is a quotation of 35 U.S.C. §103 which forms the basis for all obviousness rejections set forth in this Office action: A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made. This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. §102(b)(2)(C) for any potential 35 U.S.C. §102(a)(2) prior art against the later invention. Claims 1-4, 9-10, 12-14, 18, and 21 are rejected under 35 U.S.C. §103 as being unpatentable over Jung et al. (US 2022/0114427), hereinafter Jung, and Ito et al. (US 2023/0176818), hereinafter Ito, and Rohm et al (CN 112154460 A), hereinafter Rohm. (Please see the attached copy of Rohm that numbers paragraphs in the same manner as that used in this Action). Claim 1 Regarding: “A device, comprising: input circuitry; weight operation circuitry electrically connected to the input circuitry” Jung [0044], [0052], [0059], [0070], [0074], [0078] teach a neural network operation comprising memory cells and resistances corresponding to the weights, when voltage corresponding to an input activation is applied to each of the plurality of word lines and transistors can be turned on or turned off to control the supply of current to the variable resistance element, as well as other word and column line input circuitry; Jung [0042], [0043], [0052], [0060] teach at least weight filters which are physical circuits and weighted channel circuitry such as through resistance circuits, which are connected to channel input lines. “bias operation circuitry electrically connected to the weight operation circuitry” Jung [0078] teaches a neural network operation comprising memory cells and resistances corresponding to weights, currents are output from memory cells arranged along one column circuit, summation currents and output along the plurality of bit lines; Jung [0052], [0053] teach bias connection valves and channel bias which connect with the weighted channel lines. “storage circuitry electrically connected to the weight operation circuitry and the bias operation circuitry” Jung [0089] teaches when the neural network operation is performed, the column circuit may be connected to one the bit line and output the summation current and a weight value stored in each memory cell; Jung Figure 8, [0060], [0072], [0091], [0111] further teach column circuits and memory circuits for storing weight values and which are connected to [weighted] channel lines. “activation function circuitry electrically connected to the bias operation circuitry” Jung [0053], [0054] teach activation function channel lines which connect to the channel bias. Jung does not explicitly teach “wherein at least the weight operation circuitry, the bias operation circuitry, and the storage circuitry are located on a same chip.” Ito [0077], [[0111], [0120], [0294], [0295] however do teach a System on Chip semiconductor chip integrated circuit. The integrated circuit includes storage, bias, and weight operation circuits. See also Ito [0059-0062] which show the storage circuit with switching circuits (including product-sum operation circuits) capable of bias and weight functions. Jung and Ito disclose analogous art. Jung teaches weight, bias, storage, and activation function circuitries and Ito shows “weight, bias, and storage operation circuitry on one chip” as explained above. Hence, it would have been obvious to one ordinary skilled in the art before the effective filing date of the claimed invention to have the weight operation circuitry, the bias operation circuitry, and the storage circuitry all located on the same chip, because it would provide and efficient way to interconnect and process the circuitry together. Doing so using the technique of Ito would also reduce power consumption needs (Ito [0294]). Jung and Ito do not explicitly teach “the storage circuitry comprising non-volatile magneto-resistive random access memory (MRAM) storing a weight value applied by the weight operation circuitry and a bias value applied by the bias operation circuitry.” Rohm [151-152, 154] however do show storing a weight value (applied by weight operation circuitry) and a bias value (applied by bias operation circuitry) in a non-volatile memory which is described as having storage circuitry components, and Rohm [111] shows this non-volatile memory may be a magneto-resistive random access memory. Jung (and Ito) and Rohm disclose analogous art. Jung teaches weight, bias, storage, and activation function circuitries, and Rohm shows the magneto-resistive random access memory storage circuitry which stores the weight value and bias value. Hence, it would have been obvious to one ordinary skilled in the art before the effective filing date of the claimed invention to have the storage circuitry comprise non-volatile magneto-resistive random access memory (MRAM) storing a weight value applied by the weight operation circuitry and a bias value applied by the bias operation circuitry, because it would provide an efficient way to store the weight value and bias value Rohm [111]. Claim 2 Regarding: “wherein the weight operation circuitry comprises first weight operation circuitry and second weight operation circuitry” Jung [0078][0113] teaches spreading the weight value of one memory cell to the remaining memory cells, when data “1” and “0” is stored in the first memory cell, current may flow in the direction of an arrow that data “1” can be stored in the first memory cell, the weight value stored in the second memory cell may be stored in the third memory cell and the fourth memory cell in the same manner, “wherein the storage circuitry comprises first storage circuitry electrically connected to the first weight operation circuitry, second storage circuitry electrically connected to the second weight operation circuitry, and third storage circuitry electrically connected to the bias operation circuitry” Jung [0078][0113] teaches the first weight and the second weight connected, and second and third storage connected to summation operation circuit. Claim 3 Regarding: “wherein the weight operation circuitry comprises first weight operation circuitry and second weight operation circuitry, and wherein the storage circuitry is electrically connected to the first weight operation circuitry, the second weight operation circuitry, and the bias operation circuitry” Jung [0078][0113] teaches the first weight and the second weight connected, and second and third storage connected to summation operation circuit. Claim 4 Regarding: “wherein the storage circuitry comprises one or more storage bits” Jung [0078] teaches a neural network operation comprising memory cells and resistances corresponding to the weights, currents are output from memory cells arranged along one column circuit, summation currents and output along the plurality of bit lines. Claim 9 Regarding: “wherein the device comprises a hardware neuron in an artificial neural network” Jung [0050] shows the neuron channel processing elements that are hardware implemented nodes of the neural network. Claim 10 Regarding: “output circuitry electrically connected to the activation function circuitry” Jung [0078] teaches a neural network operation comprising memory cells and resistances corresponding to the weights, when voltage corresponding to an input activation is applied to each of the plurality of word lines and resistances corresponding to weights, currents are output from memory cells arranged along one column circuit, summation currents and output along the plurality of bit lines. Claim 12 Claim 12 is rejected for the similar rationale given for claim 1. Additionally, Rohm [117, 137, 154, 180] show the artificial neural network comprises at least a first and second neuron devices, each of them comprising the input circuitry, weight operation circuitry, bias operation circuitry, and storage circuitry are of the first and second neuron devices. Furthermore, Rohm [112, 115, 180-181] show the storage circuitry used for the first neuron device is different from the storage circuitry used for the second neuron device. Claims 13-14 & 18 Claims 13-14, and 18 are rejected for the similar rationale given for claims 2-3, and 10 respectively. Additionally, note that Rohm [117, 137, 154, 180] shows the weight operation circuitry and storage circuitry (in claims 13-14) and output circuitry (in claim 18) are of the first and second neuron devices Claim 21 Claim 21 is rejected for the similar rationale given for claim 1. Additionally, note that Jung [0044], [0052], [0059], [0070], [0074], [0078] teach the multiple/plurality of input activation and column line input circuitry, and Jung [0042], [0043], [0052], [0060] teach a plurality of weight filters which are physical circuits and weighted channel circuitry such as through resistance circuits, which are connected to channel input lines. Claims 5-6 and 15-16 are rejected under 35 U.S.C. §103 as being unpatentable over Jung and Ito and Rohm, and Shainline (US 2020/0301874), hereinafter Shainline. Claim 5 Jung and Ito and Rohm do not explicitly teach “wherein the one or more storage bits each comprise one or more resistive elements and a voltage amplifier” Shainline [0031][0037] however does show the dendritic integration resistor and superconducting voltage amplifier in communication with dendritic integrated current. Jung, Ito, Rohm, and Shainline disclose analogous art. Jung teaches weight, bias, storage, and activation function circuitries. Ito teaches locating all circuitries on a same chip. Rohm shows the MRAM. Shainline shows the storage bits have the resistive elements and voltage amplifier. Hence, it would have been obvious to one ordinary skilled in the art at the time the present invention was made to incorporate said feature of Shainline into Jung as modified by Ito and Rohm, because it would provide an efficient way to enhance its storage functions with more resistive elements and a voltage amplifier. Claim 6 In addition to that mentioned for claim 5, regarding: “wherein the one or more resistive elements comprise at least four resistive elements, wherein at least two first resistive elements are electrically connected in series and at least two second resistive elements are electrically connected in series, wherein the at least two first resistive elements are electrically connected in parallel to the at least two second resistive elements,” Shainline Figure 10 illustrates resistive elements connected in series 201 and more than one synapse connected in parallel; “wherein an input of the voltage amplifier is electrically connected to a first electrode between the at least two first resistive elements and to a second electrode between the at least two second resistive elements” Shainline Figure 9 & [0037] teaches amplifier 480 connected to electronic dendrite 342 output signal. Motivation to combine with Shainline with Jung as modified by Ito is the same as that modified by claim 5. Claims 15-16 Claims 15 and 16 are rejected for the similar rationale given for claims 5 and 6 respectively. Additionally, note for the claim 15 that Rohm [117, 137, 154, 180] shows the storage circuitry is of the first and second neuron devices. For claim 16 note that Shainline [36, 103] shows the resistive elements would thus be of the first and second neuron devices. Claim 7 is rejected under 35 U.S.C. §103 as being unpatentable over Jung and Ito and Rohm and Shainline, and Woo et al. (US 2022/0359617), hereinafter Woo. Claim 7 In addition to that mentioned for claim 5, Jung and Ito and Rohm and Shainline do not explicitly teach “wherein each of the one or more resistive elements comprise a magnetic tunnel junction (MTJ)” Woo [0003] however teaches a magnetic tunnel junction(MTJ) structure of a resistive memory device. Jung, Ito, Rohm, Shainline, and Woo disclose analogous art. Jung teaches weight, bias, storage, and activation function circuitries. Ito teaches locating all circuitries on a same chip. Shainline teaches resistive elements and a voltage amplifier. Rohm teaches the MRAM. Woo teaches the “magnetic tunnel junction” as recited above. Hence, it would have been obvious to one ordinary skilled in the art at the time the present invention was made to incorporate said feature of Woo into Jung as modified by Ito and Rohm and Shainline, because it would provide and efficient way to further enhance its storage functions with a magnetic tunnel junction as one of the resistive elements. Claims 8, 11, 17, and 19 rejected under 35 U.S.C. §103 as being unpatentable over Jung and Ito and Rohm, and Woo et al. (US 2022/0359617), hereinafter Woo. Claim 8 In addition to that mentioned for claim 4, Jung and Ito and Rohm do not explicitly teach “wherein the one or more storage bits are included in a single array of bits” Woo [0030][0031] however does teach multiple control lines including source lines and bit lines, and various blocks connected in the 2D array of RPU cells, which are configured to perform vector-matrix multiply functions. Jung, Ito, Rohm, and Woo disclose analogous art. Jung teaches weight, bias, storage, and activation function circuitries. Ito teaches locating all circuitries on a same chip. Rohm teaches the MRAM. Woo teaches the “one or more storage bits included in a single array of bits” feature as recited above. Hence, it would have been obvious to one ordinary skilled in the art at the time the present invention was made to incorporate said feature of Woo into Jung as modified by Ito and Rohm, because it would provide an efficient way to further enhance its storage functions with array of bit lines control operations. Claim 11 In addition to that mentioned for claim 4, regarding “wherein each of the one or more storage bits comprises: a first set of resistive elements and a second set of resistive elements, first read circuitry electrically connected to the first set of resistive elements and second read circuitry electrically connected to the second set of resistive elements, cross-coupled inverter circuitry electrically connected to the first read circuitry and the second read circuitry, and third read circuitry electrically connected to the cross-coupled inverter circuitry” Jung [0063], [0078], [0091] show the sets of resistive elements connected with the read line circuitry. Jung and Ito and Rohm do not teach the cross-coupled inverter circuitry. Woo [0020], [0026] however teaches a circuit structure with inverters with cross-points, and read/sensing operations, e.g., read a weight value of given RPU cell. It would have been obvious to one ordinary skilled in the art at the time the present invention was made to incorporate the inverter circuitry of Woo into Jung as modified by Ito and Rohm, because it would provide and efficient way to implement high speed read operation using resistive elements. Claims 17 & 19 Claims 17 and 19 are rejected for the similar rationale given for claims 8 and 11 respectively. Response to Arguments 5. Applicant's arguments filed 9/21/26 have been fully considered but they are not persuasive. Applicant argues that the prior art used in the previous rejection does not show the newly amended features. However, Rohm is brought in to show these newly amended features, as explained in the Action. Conclusion 6. The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: a) Sudo (US 2019/0238092 A1) shows an integrated circuit with input, bias, storage, and activation function circuitry. b) Timofejevs (WO 2021259482 A1) shows integrated circuit hardware for neural networks containing input, bias, storage, and activation function circuitry, and which also uses inverter circuitry. c) Higashi (CN 113711205 A) shows a neural network with a plurality of neurons and that uses a magneto-resistive random access memory to store weighting and bias values. 7. Any inquiry concerning this communication or earlier communications from the examiner should be directed to STEVEN PAUL SAX whose telephone number is (571)272-4072. The examiner can normally be reached Monday - Friday, 9:30 - 6:00 Est. Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Usmaan Saeed, can be reached at 571-272-4046. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000. /STEVEN P SAX/ Primary Examiner, Art Unit 2146
Read full office action

Prosecution Timeline

Show 2 earlier events
Dec 16, 2025
Non-Final Rejection mailed — §103
Mar 09, 2026
Response Filed
Jul 28, 2026
Examiner Interview (Telephonic)
Jul 30, 2026
Final Rejection mailed — §103
Sep 21, 2026
Response after Non-Final Action
Sep 25, 2026
Request for Continued Examination
Sep 29, 2026
Response after Non-Final Action
Oct 05, 2026
Non-Final Rejection mailed — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12801892
METHOD FOR FUSING OPERATORS OF NEURAL NETWORK, AND RELATED PRODUCT
2y 7m to grant Granted Nov 17, 2026
Patent 12801927
SYSTEMS AND METHODS FOR SHAPE OPTIMIZATION OF STRUCTURES USING PHYSICS INFORMED NEURAL NETWORKS
2y 9m to grant Granted Nov 17, 2026
Patent 12783856
METHOD TO DETECT AND FIX HALLUCINATIONS IN GENERATIVE LARGE LANGUAGE MODELS
3y 3m to grant Granted Nov 03, 2026
Patent 12775861
Quantum Number Preserving Circuits For Preparing Quantum States Representing Fermions In Computational Units-Based Quantum Computers
3y 7m to grant Granted Oct 27, 2026
Patent 12755156
NEURAL NETWORK MODEL COMPRESSION METHOD AND APPARATUS, STORAGE MEDIUM, AND CHIP
4y 7m to grant Granted Oct 06, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

3-4
Expected OA Rounds
69%
Grant Probability
99%
With Interview (+45.8%)
4y 1m (~0m remaining)
Median Time to Grant
High
PTA Risk
Based on 470 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month