DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, 365(c), or 386(c) is acknowledged.
The examiner acknowledges the priority benefit to U.S. Application No. 16/531,085, filed on 8/04/2019. The present application is a continuation-in-part of Application No. 16/531,085.
Applicant has not complied with one or more conditions for receiving the benefit of an earlier filing date under 35 U.S.C. 119(e) as follows:
As discussed below, the claimed features of the independent claims are not supported by U.S. Application No. 16/531,085 (hereinafter “the parent application”). Therefore, claims 1-20 do not enjoy the priority benefit of the parent application 16/531,085.
The disclosure of the prior-filed parent application fails to provide adequate support or enablement in the manner provided by 35 U.S.C. 112(a) or pre-AIA 35 U.S.C. 112, first paragraph for one or more claims of this application. Independent claim 1 recites “wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered hexagonal prism made of Mu-metal and ferrite materials, and substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons” (see, the last 4 lines of claim 1). The as-filed specification of the parent application fails to provide adequate support or enablement for at least these elements of independent claim 1.
The above recited features of claim 1 are not defined or described in the specification of the parent application in such a way as to reasonably convey to one skilled in the relevant art that the inventor, at the time the application was filed, had possession of the claimed invention.
For example, the original specification of the parent application is silent regarding any conversation “three-layered hexagonal prism” or any “three-layered hexagonal prism made of Mu-metal and ferrite materials”, let alone “wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered hexagonal prism made of Mu-metal and ferrite materials, and substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons” as recited in claim 1. For example, aside from a single mention in paragraph 5 of the original specification of the parent application that “FIG. 1 illustrates an example of embodiments that consists of three layers of hexagon stereoscopic magnetic effect artificial neurons”, the terms “hexagonal” or “hexagon”, do not appear anywhere in the specification of the parent application. Further, for example the term “Mu-metal” (or μ-metal) does not appear anywhere in the original specification of the parent application. That is, the specification of the parent application is silent regarding any “three-layered hexagonal prism made of Mu-metal”. As such, the claim term “three-layered hexagonal prism made of Mu-metal” is not defined or described in the parent application.
Also, dependent claim 3 recites “wherein the middle head and the middle tail are made of ferrite, and the middle body is made of Mu-metal.” Further, dependent claim 4 recites, “wherein the top head, the top body and the top tail are made of Mu-metal.” Additionally, dependent claim 5 recites “wherein the bottom head, the bottom body and the bottom tail are made of Mu-metal.” The as-filed specification of the parent application fails to provide adequate support or enablement for at least these elements of dependent claims 3-5.
The specification of the instant application as-filed on 3/21/2023 appears to provide adequate support and enablement for at least the above-noted elements of independent claims 1 and 11, and dependent claims 3-5. Thus, the instant application appears to support at least independent claim 1, and dependent claims 3-5, as currently written. Therefore, the effective filing date for the claims of the instant application is the filing date of the instant application, 3/21/2023. Examiner will consider if the parent application supports each dependent claim (aside from dependent claims 3-5) if a rejection would need to rely upon an intervening reference between the actual filing date of 3/21/2023 for the instant application and the 8/04/2019 filing of the parent application.
Each claim will receive benefit of the earliest filing date above for which a continuous chain of support can be established for the entirety of the claim. As discussed above, the effective filing date for the claims of the instant application is the filing date of the instant application, 3/21/2023.
Drawings
The drawings are objected to as failing to comply with 37 CFR 1.84(p)(4) because reference character “332” has been used to designate both a bottom body (BB) 332 and a bottom tail (BT) in FIG. 2.
The drawings are additionally objected to as failing to comply with 37 CFR 1.84(p)(5) because they do not include the following reference signs mentioned in the description:
333 (see, e.g., paragraph 37 describing FIG. 2 and reciting “a bottom tail (BT) 333.”).
Corrected drawing sheets in compliance with 37 CFR 1.121(d) are required in reply to the Office action to avoid abandonment of the application. Any amended replacement drawing sheet should include all of the figures appearing on the immediate prior version of the sheet, even if only one figure is being amended. Each drawing sheet submitted after the filing date of an application must be labeled in the top margin as either “Replacement Sheet” or “New Sheet” pursuant to 37 CFR 1.121(d). If the changes are not accepted by the examiner, the applicant will be notified and informed of any required corrective action in the next Office action. The objection to the drawings will not be held in abeyance.
Specification
The disclosure is objected to because of the following informalities:
In the “CROSS-REFERENCE TO RELATED APPLICATION” section in paragraph 1 of the specification, reference is made to “U.S. Application Ser. No. 16/531,085”. This application should be identified by its U.S. Patent Application Publication number(s) and/or U.S. Patent number(s), if it has been published and/or issued. In particular, for example, the above-noted U.S. Patent Application No. 16/531,085 has been published as U.S. Patent Application Pub. No. 2021/0034954. Appropriate correction is required.
Paragraph 15 recites “a signal differentiator configured to direct signal to either the first pulsating direct current module or the second pulsating direct current module.” This recitation is grammatically incorrect and appears to be missing one or more words between “direct” and “signal”. If supported by applicant’s original specification, examiner suggests two possible ways to address this objection would be to amend “a signal differentiator configured to direct signal to either the first pulsating direct current module or the second pulsating direct current module” to read “a signal differentiator configured to transmit a direct signal to either the first pulsating direct current module or the second pulsating direct current module” or “a signal differentiator configured to transmit direct signals to either the first pulsating direct current module or the second pulsating direct current module” (see, e.g., paragraphs 4 and 39 of the specification reciting “a synapse, which can transmit a signal to other neurons.” and “a schematic diagram of signal transmission of the magnetic effect artificial neuron”). Appropriate correction is required.
With reference to FIG. 2, paragraph 37 recites “a bottom body (BB) 332 and a bottom tail (BT) 333.” However, FIG. 2 depicts a bottom body (BB) 332 and a bottom tail (BT) 332. As such, references in the specification to elements 332 and 333 are inconsistent and/or include one or more typographical errors. Appropriate correction is required.
Claim Objections
Claim 8 is objected to because of the following informalities:
Claim 8 recites “a signal differentiator configured to direct signal to either the first pulsating direct current module or the second pulsating direct current module.” This recitation is grammatically incorrect and appears to be missing one or more words between “direct” and “signal”. If supported by applicant’s original specification, examiner suggests two possible ways to address this objection would be to amend “a signal differentiator configured to direct signal to either the first pulsating direct current module or the second pulsating direct current module” to read “a signal differentiator configured to transmit a direct signal to either the first pulsating direct current module or the second pulsating direct current module” or “a signal differentiator configured to transmit direct signals to either the first pulsating direct current module or the second pulsating direct current module” (see, e.g., paragraphs 4 and 39 of the specification reciting “a synapse, which can transmit a signal to other neurons.” and “a schematic diagram of signal transmission of the magnetic effect artificial neuron”). Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
Claims 1-16 are rejected under 35 U.S.C. 112(b) as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor regards as the invention.
The last limitation of claim 1 recites “wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered hexagonal prism made of Mu-metal and ferrite materials, and substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons.” The term “substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention. In particular, it is unclear what metrics are used for ascertaining the requisite degree of attachment for the term "substantially” in the phrase "substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons.” Aside from merely repeating the claim language in stating “Each of the plurality of magnetic effect artificial neurons is shaped as a three-layered hexagonal prism made of Mu-metal and ferrite materials, and substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons.” and providing general examples in stating “Each neuron substantially attaches or is very close to adjacent neighbors and forms interactions by means of boundary effect, induction effect and diffusional effect of magnetic field.”, “Each of the magnetic effect artificial neurons 3 is shaped as a three-layered hexagonal prism, and substantially attaches or is very close to adjacent ones of the magnetic effect artificial neurons 3.” and “each magnetic effect artificial neuron is shaped as a three-layered hexagonal prism and substantially attaches to adjacent neurons, and thus, interactions between the neurons can be formed.” (see, e.g., paragraphs 8, 31-32 and 60), the specification does not explicitly define what is meant by the recited “vibrant constitutional guidance” or provide a standard for ascertaining the requisite degree of the term "substantially” of the claimed “wherein each of the plurality of magnetic effect artificial neurons … substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons.”
For the purposes of determining patent eligibility and comparison with the prior art, the examiner is interpreting the term “wherein each of the plurality of magnetic effect artificial neurons … substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons” as each of the artificial neurons being attached to, in close proximity to, or very close to adjacent ones of the plurality of magnetic effect artificial neurons. Appropriate correction is required.
Also, claims 2-16, which each depend directly or indirectly from claim 1, are rejected under 35 U.S.C. 112(b) as being indefinite under the same rationale as claim 1.
Claim Rejections - 35 USC § 103
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-5, 7-8 and 10-16 are rejected over Schneider et al. (U.S. Patent Application Pub. No. 2019/0207076 A1, hereinafter “Schneider”) in view of Choi et al. (U.S. Patent Application Pub. No 2019/0245136 A1, hereinafter “Choi”) and further in view of Friedman et al. (U.S. Patent Application Pub. No 2021/0232903 A1, hereinafter “Friedman”).
With respect to claim 1, Schneider discloses the invention as claimed including a magnetic effect artificial intelligence system1 (see, e.g., paragraphs 62, “The artificial synapses are compatible with single flux quantum (SFQ) Josephson Junction (JJ) circuits that provide a platform for a large neuromorphic system … The neural member is a synaptic element based on a dynamically reconfigurable JJ synapse … neuromorphic system that can include SFQ neurons.” [i.e., an artificial intelligence system including artificial neurons and synapses] and 63, “By aligning the net spins of several clusters, we increase the overall magnetic order in the JJ synapse, which tunes the critical current of that synapse. ” [i.e., system includes synapses with magnetic effects]), comprising:
an input pre-processing unit2 (see, e.g., paragraphs 61, “neural member as an artificial synapse with a … Josephson junction [JJ] that includes magnetic nanoclusters … current of the synapse junction, analogous to the synaptic weight, can be tuned using input voltage spikes” [i.e., magnetic effect artificial neural members/synapses/neurons accept input voltage], 63, “The neural member artificial synaptic element weights the input … signals to … neuronal elements” and 72, “A standard single SFQ JJ is used to represent a pre-synaptic neuron … This JJ provides the input pulses for the circuit. The JJ synapse element acts to weight the pulses … This synapse will pass more of the input pulse when it remains in the superconducting state and less when it enters the voltage state.” [i.e., the system comprises an element/unit that processes/weights input signals – an input pre-processing unit]);
a plurality of magnetic effect artificial neurons connected with the input pre-processing unit (see, e.g., paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., a plurality of neurons are connected/interconnected via wiring], 62, “neural member as an artificial synapse with a dynamically reconfigurable superconducting Josephson junction that includes magnetic nanoclusters … can be tuned using input voltage spikes that change the spin alignment of Mn nanoclusters.” [i.e., magnetic effect artificial neural members/synapses/neurons], 63, “artificial synapses are compatible with single flux quantum (SFQ) Josephson Junction (JJ) circuits that provide a platform for a large neuromorphic system. … neuromorphic system that can include SFQ neurons.” [i.e., magnetic effect artificial neurons and circuits], 73, “FIG. 9a shows a schematic of a JJ synapse … The amount of magnetic order between the clusters … The neural member artificial synaptic element weights the input … signals to … neuronal elements” [i.e., the neurons/neuronal elements are connected with the input unit to receive the input signals]); and
an output unit connected with the plurality of magnetic effect artificial neurons (see, e.g., paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., system members/neurons and components/units are connected/interconnected], 63, “The neural member artificial synaptic element weights the … output signals … from neuronal elements” [i.e., output signals from neuronal/elements/neurons are weighted by synaptic element/unit] and 72, “the JJ synapse. This JJ acts as the output for the circuit element” [i.e., the system comprises an element/JJ synapse unit that outputs signals to other neuronal elements – an output unit]),
wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered … made of Mu-metal and ferrite materials3 (see, e.g., paragraphs 28, “neural member 100 includes ferromagnetic layer 132 interposed between dendritical superconducting electrode 110 and synaptic barrier 112. Neural member 100 can include spacing layer 130 interposed between ferromagnetic layer 132 and synaptic barrier 112.”, 37, “Neural member 100 can include spacing layer 130 that includes, e.g., include Cu, Nb, and the like. In an embodiment, spacing layer 130 includes Cu. Spacing layer 130 provides decoupling of the magnetic clusters in the synaptic barrier 112 from the spin polarizing layer 132.” and 38, “Neural member 100 can include a ferromagnetic layer 132 that includes, e.g., include Fe, NiFe, and the like. … Ferromagnetic layer 132 provides a material to spin polarize the electrons impinging on the synaptic barrier 112.” [i.e., the neural members/magnetic effect artificial neurons are three-layer structures made of metal material having magnetic permeability and saturation, and ferrite – including Fe/Iron and NiFe ferromagnetic materials]), and substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons4 (see, e.g., FIG. 17 - showing a neural network with artificial neurons in an input layer that are connected/attached to adjacent ones of neurons in an output layer and paragraphs 23, “FIG. 17 shows neural network connections”, 49, “neural members 100 can be interconnected by superconducting wiring” and 94, “inputs are connected to 27 synaptic MJJs which have their critical currents adjusted to form the weights of the middle layer. These are then connected to 3 threshold MJJs forming the output layer.” [i.e., artificial neurons/neural members 100 in layers are connected/attached to adjacent neurons]).
Although Schneider substantially discloses the claimed invention, Schneider is not relied on for explicitly disclosing wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered hexagonal.
In the same field, analogous art Choi teaches wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered hexagonal5 (see, e.g., paragraphs 3, “a magnetic tunnel junction for storing data includes a reference layer, a barrier layer, and a free layer … includes a nucleation region configured to form a magnetic domain wall.”, 4, “a plurality of artificial neurons and a synapse array of multi-state magnetic memory cells coupling the artificial neurons … includes a pad-shaped region for forming a magnetic domain wall.” [i.e., magnetic effect artificial neurons shaped as a three-layer structure] and 96, “the nucleation region 602 may be a pad-shaped region of the free layer 600 … if a length (or longest dimension) of the region is comparable to the width (or shortest dimension) of the region … a pad-shaped region may be convex … a pad-shaped region may be convex. For example, a pad-shaped region may be a … hexagon” [i.e., the layered structure shaped as a hexagonal/hexagon having dimensions, viewable as having three dimensions]).
Schneider and Choi are analogous art because they are both related to neuromorphic computing and implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, and Choi, paragraph 4).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider to incorporate the teachings of Choi to provide “systems and methods for magnetoresistive random access memory” where “A magnetic tunnel junction for storing data may include a reference layer, a barrier layer, and a free layer … for neuromorphic computing” (See, e.g., Choi, Abstract and paragraphs 3-4). Doing so would have allowed Schneider to use Choi’s “plurality of magnetic tunnel junctions (MTJs) for storing data. In certain embodiments, an MTJ includes a reference layer, a barrier layer, and a free layer” as components of “an artificial neural network [that] includes an artificial neuron 306, and a plurality of synapses”, as suggested by Choi (See, e.g., Choi, paragraphs 47 and 52).
Although Schneider in view of Choi substantially teaches the claimed invention, Schneider in view of Choi is not relied on to teach wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered … prism.
In the same field, analogous art Friedman teaches wherein each of the plurality of magnetic effect artificial neurons is shaped as a three-layered … prism (see, e.g., FIG. 15B – depicting 3-layered prism-shaped trapezoidal neurons, and paragraphs 9, “a domain wall magnetic tunnel junction device comprising a number of ferromagnetic tracks, wherein each ferromagnetic track has a first fixed magnetization region at a first end and a second fixed magnetization region at a second end, and … each ferromagnetic track is trapezoidal having a first width at the first end and a second width at the second opposite end, wherein the second width is longer than the first width. A magnetic tunnel junction is located between the first and second ends of each ferromagnetic track, wherein each magnetic tunnel junction comprises a tunnel barrier on the ferromagnetic track and a fixed ferromagnet on top of the tunnel barrier.”, 13, “plurality of domain wall magnetic tunnel junction artificial neurons” [i.e., each of the magnetic effect artificial neurons is shaped as a three-layered structure] and 116-117, “The shape-based DW drift provides a native representation of neuron … the shape-based DW drift enables an artificial 3T-MTJ neuron”, “With this trapezoidal prism, the DW velocity is also influenced by the width” and “FIG. 15 illustrates combined integration and leaking behavior of the trapezoidal neuron” [i.e., each three-terminal, magnetic tunnel junction/3T-MTJ neuron is shaped as a three-layered trapezoidal prism]).
Schneider, Choi and Friedman are analogous art because they are each related to neuromorphic computing and implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, Choi, paragraph 4, and Friedman, Abstract and paragraphs 12-13).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider in view of Choi to incorporate the teachings of Friedman to provide “a domain wall magnetic tunnel junction device comprising a number of ferromagnetic tracks, wherein each ferromagnetic track has a first fixed magnetization region” for a “neural network [that] comprises a first crossbar array of domain wall magnetic tunnel junction synapses and a first plurality of domain wall [DW] magnetic tunnel junction [MTJ] artificial neurons” where a “shape-based DW drift provides a native representation of neuron[s]” (See, e.g., Friedman, paragraphs 9, 13 and 116). Doing so would have allowed Schneider in view of Choi to use Friedman’s trapezoidal prism domain wall (DW) magnetic tunnel junction (MTJ) device because “The shape-based DW drift provides a native representation of neuron leaking that enables simplification of the device structure … the shape-based DW drift enables an artificial 3T-MTJ neuron with an intrinsic leaking capability.” and with the “trapezoidal prism, the DW velocity is also influenced by the width, as discussed previously in relation to the leaking; the DW moves faster where the width is smaller”, as suggested by Friedman (See, e.g., Friedman, paragraphs 116-117).
Regarding claim 2, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 1.
Schneider further discloses wherein each of the plurality of magnetic effect artificial neurons comprises a top layer, a middle layer and a bottom layer (see, e.g., paragraphs 28, “neural member 100 includes ferromagnetic layer 132 interposed between dendritical superconducting electrode 110 and synaptic barrier 112. Neural member 100 can include spacing layer 130 interposed between ferromagnetic layer 132 and synaptic barrier 112.”, 37, “Neural member 100 can include spacing layer 130 that includes, e.g., include Cu, Nb, and the like. In an embodiment, spacing layer 130 includes Cu. Spacing layer 130 provides decoupling of the magnetic clusters in the synaptic barrier 112 from the spin polarizing layer 132.”, and 38, “Neural member 100 can include a ferromagnetic layer 132 that includes, e.g., include Fe, NiFe, and the like. … Ferromagnetic layer 132 provides a material to spin polarize the electrons impinging on the synaptic barrier 112.” [i.e., the magnetic effect artificial neurons comprise three layers – ferromagnetic, spacing and barrier layers corresponding to top, middle and bottom layers, the three layers each include three parts – Cu, Nb, Fe, NiFe, and the like]), the top layer comprises a top head, a top body and a top tail, the middle layer comprises a middle head, a middle body and a middle tail, and the bottom layer comprises a bottom head, a bottom body and a bottom tail (see, e.g., paragraphs 40, “a superconducting wire that is fabricated on top of the device separated by an insulating layer.” and 44-46, “making neural member 100 includes sputter depositing the axonal electrode of Nb followed by the deposition of Si and Mn … layered followed by the deposition of the dendritical electrode.”, “Spacing layer 130 can be included in neural member 100 by sputter deposition subsequent to the barrier layer.”, “Ferromagnetic layer 132 can be included in neural member 100 by sputter deposition subsequent to the spacer layer.” [i.e., the three layers each include three deposited parts layered for the barrier, ferromagnetic and spacer layers]).
Regarding claim 3, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 2.
Schneider further discloses wherein the middle head and the middle tail are made of ferrite (see, e.g., FIG. 2 – depicting neural element/member 100 with middle parts/layers 130, 132 and 112, and paragraphs 28, “neural member 100 includes ferromagnetic layer 132 interposed between dendritical superconducting electrode 110 and synaptic barrier 112. Neural member 100 can include spacing layer 130 interposed between ferromagnetic layer 132 and synaptic barrier 112.” and 38, “Neural member 100 can include a ferromagnetic layer 132 that includes, e.g., include Fe, NiFe, and the like. … Ferromagnetic layer 132 provides a material to spin polarize the electrons impinging on the synaptic barrier 112.” [i.e., the magnetic effect artificial neurons comprise layers with parts made of ferromagnetic, ferrite materials – Fe, NiFe and the like]), and the middle body is made of Mu-metal6 (see, e.g., paragraphs 29, “electrode 110 receives electric staple pulse 120 and communicates electric state pulse 120 through … magnetic clusters … neural element 100 can include magnet 190 that can provide magnetic field”, 37, “Neural member 100 can include spacing layer 130 that includes, e.g., include Cu, Nb, and the like. In an embodiment, spacing layer 130 includes Cu. Spacing layer 130 provides decoupling of the magnetic clusters in the synaptic barrier 112 from the spin polarizing layer 132.”, 38, “Neural member 100 can include a ferromagnetic layer 132 that includes, e.g., include Fe, NiFe, and the like.”, 67, “given enough pulses, the trend … will eventually saturate” and 71, “a voltage pulse across the junction causes the order parameter to increase … up to a saturation point.” [i.e., the neural element/magnetic effect artificial neuron includes layers with parts made of metal material having magnetic permeability and saturation]).
Regarding claim 4, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 2.
Schneider further discloses wherein the top head, the top body and the top tail are made of Mu-metal7 (see, e.g., FIG. 2 – depicting neural element/member 100 with parts and upper/top layers, and paragraphs 29, “electrode 110 receives electric staple pulse 120 and communicates electric state pulse 120 through … magnetic clusters … neural element 100 can include magnet 190 that can provide magnetic field”, 37, “Neural member 100 can include spacing layer 130 that includes, e.g., include Cu, Nb, and the like. In an embodiment, spacing layer 130 includes Cu. Spacing layer 130 provides decoupling of the magnetic clusters in the synaptic barrier 112 from the spin polarizing layer 132.”, 38, “Neural member 100 can include a ferromagnetic layer 132 that includes, e.g., include Fe, NiFe, and the like. … Ferromagnetic layer 132 provides a material to spin polarize the electrons impinging on the synaptic barrier 112.”, 67, “given enough pulses, the trend … will eventually saturate” and 71, “a voltage pulse across the junction causes the order parameter to increase … up to a saturation point.” [i.e., the neural element/magnetic effect artificial neuron includes an upper/top layer with head, body and tail parts made of metal material having magnetic permeability and saturation, and a nickel-iron/NiFe alloy ferromagnetic material – Mu metal]).
Regarding claim 5, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 2.
Schneider further discloses wherein the bottom head, the bottom body and the bottom tail are made of Mu-metal8 (see, e.g., FIG. 2 – depicting neural element/member 100 with parts and lower/bottom layers, and paragraphs 29, “electrode 110 receives electric staple pulse 120 and communicates electric state pulse 120 through … magnetic clusters … neural element 100 can include magnet 190 that can provide magnetic field”, 37, “Neural member 100 can include spacing layer 130 that includes, e.g., include Cu, Nb, and the like. In an embodiment, spacing layer 130 includes Cu. Spacing layer 130 provides decoupling of the magnetic clusters in the synaptic barrier 112 from the spin polarizing layer 132.”, 38, “Neural member 100 can include a ferromagnetic layer 132 that includes, e.g., include Fe, NiFe, and the like. … Ferromagnetic layer 132 provides a material to spin polarize the electrons impinging on the synaptic barrier 112.”, 67, “given enough pulses, the trend … will eventually saturate” and 71, “a voltage pulse across the junction causes the order parameter to increase … up to a saturation point.” [i.e., the neural element/magnetic effect artificial neuron includes a lower/bottom layer with head, body and tail parts made of metal material having magnetic permeability and saturation, and a nickel-iron/NiFe alloy ferromagnetic material – Mu metal]).
Regarding claim 7, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 1.
Schneider further discloses wherein each of the plurality of magnetic effect artificial neurons comprises:
a signal differential module9 connected to the input pre-processing unit (see, e.g., paragraphs 70, “A modified version of the resistively and capacitively shunted junction (RCSJ) model was used to simulate the reconfigurable JJ synapses” [i.e., artificial JJ synapses are connected to the shunt junction/unit], 82, “control of the magnetic cluster size allows one to vary the amount of energy required to adjust the weight of the artificial synapse.” and 94, “synaptic MJJs which have their critical currents adjusted to form the weights of the middle layer. These are then connected to 3 threshold MJJs forming the output layer.” [i.e., synaptic MJJs include a module that accepts and processes/adjusts signals and feedback/weights]);
a first pulsating direct current module and a second pulsating direct current module10, both of which are connected to the signal differential module (see, e.g., FIG. 12 – depicting positive and negative pulsating voltages [i.e., generated by positive and negative pulsating DC modules] and paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and modules], 89, “a positive voltage pulse across the junction … In the circuits presented here, we set the minimum pulse energy equal to 3 aJ” [i.e., a module/circuit generates positive pulsating DC signal] and 91, “The circuit operates when the input DC bias is turned on. At this point the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy [i.e., pulsating DC current]);
a first magnetoresistance … unit and a second magnetoresistance … unit, both of which are connected to the first pulsating direct current module, the second pulsating direct current module, and the signal differential module11 (see, e.g., paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and units and modules], 90, “The JJ in the center is a ‘synaptic’ MJJ which will weight the transmitted pulse from the input ‘neuron.’ We adjust the critical current of these junctions between 1 μA and 100 μA to achieve the desired ‘synaptic’ weighting. This adjustment would be accomplished physically by adjusting the magnetic order in the MJJ … adjust the range of the critical currents that the MJJ has by changing the size of the junction” and 91, “The resistor and inductor above the middle MJJ act to stabilize the circuit” [i.e., MJJs include first and second junctions/units with adjustable magnetic orders and resistors - magnetoresistance units that are both connected to the first and second direct current modules and the signal differential module via the interconnections/wiring]); and
a trigger unit12 connected between the first magnetoresistance and amplification unit and the second magnetoresistance and amplification unit (see, e.g., paragraphs 91, “the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy” and 92, “As can be seen in both the voltage and phase traces, the input and output JJs are firing in this configuration” [i.e., certain JJs are trigger units that fire/are triggered and are middle JJs/MJJs connected between other units]).
Although Schneider substantially discloses the claimed invention, Schneider is not relied on for explicitly disclosing a first magnetoresistance and amplification unit and a second magnetoresistance and amplification unit, wherein the first magnetoresistance and amplification unit is disposed inside the middle head and the second magnetoresistance and amplification unit is disposed inside the middle tail.
In the same field, analogous art Choi teaches a first magnetoresistance and amplification unit and a second magnetoresistance and amplification unit13, wherein the first magnetoresistance and amplification unit is disposed inside the middle head and the second magnetoresistance and amplification unit is disposed inside the middle tail (see, e.g., FIG. 5 – depicting a magnetic tunnel junctions (MTJ) 500 with units in middle layers 502, 504, 506 between terminals 522, 524, and paragraphs 47, “MRAM die, including a plurality of magnetic tunnel junctions (MTJs)” [i.e., magnetic/electromagnetic components], 82, “the resulting current through the free layer 502, the barrier layer 504, and the reference layer 506 may be measured or sensed to detect the resistance of the MTJ [i.e., magnetic measurement for magnetic tunnel junction/MTJ] … MRAM die 150, neuromorphic computing die 450, or the like may include sense amplifiers, latches, and the like, to convert a low power signal to a logic level representing a data value” and 86, “MTJs 500, sensing components such as sense amplifiers” [i.e., middle layer parts/portions of MTJs include first and second amplifier components that sense/measure and amplify a signal]).
Schneider and Choi are analogous art because they are both related to neuromorphic computing and implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, and Choi, paragraph 4).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider to incorporate the teachings of Choi to provide “systems and methods for magnetoresistive random access memory” where “A magnetic tunnel junction for storing data may include a reference layer, a barrier layer, and a free layer … for neuromorphic computing” (See, e.g., Choi, Abstract and paragraphs 3-4). Doing so would have allowed Schneider to use Choi’s “plurality of magnetic tunnel junctions (MTJs) for storing data. In certain embodiments, an MTJ includes a reference layer, a barrier layer, and a free layer” as components of “an artificial neural network [that] includes an artificial neuron 306, and a plurality of synapses”, as suggested by Choi (See, e.g., Choi, paragraphs 47 and 52).
Regarding claim 8, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 7.
Schneider further discloses wherein the signal differential module is a signal differentiator configured to direct signal to either the first pulsating direct current module or the second pulsating direct current module14 (see, e.g., FIG. 12 – depicting positive and negative pulsating voltages [i.e., signals directed to positive and negative pulsating DC modules] and paragraphs 63, “neural member artificial synaptic element weights the input and output signals to and from neuronal elements”, 81, “the signal and training pulses are … applied locally with a field control line”, 89, “a positive voltage pulse across the junction … In the circuits presented here, we set the minimum pulse energy equal to 3 aJ” [i.e., a module/circuit generates and directs a positive pulsating DC signal], 91, “The circuit operates when the input DC bias is turned on. At this point the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy [i.e., directing a pulsating DC current signal] and 94, “synaptic MJJs which have their critical currents adjusted to form the weights of the middle layer. These are then connected to 3 threshold MJJs forming the output layer.” [i.e., synaptic MJJs include a signal module that processes/differentiates and directs signals to a direct current/DC module]).
Regarding claim 10, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 7.
Although Schneider substantially discloses the claimed invention, Schneider is not relied on for explicitly disclosing wherein the first magnetoresistance and amplification unit and the second magnetoresistance and amplification unit are magnetoresistance amplifiers configured to measure magnetic field strength and work with resistance to generate signal gain.
In the same field, analogous art Choi teaches wherein the first magnetoresistance and amplification unit and the second magnetoresistance and amplification unit15 are magnetoresistance amplifiers configured to measure magnetic field strength and work with resistance to generate signal gain (see, e.g., paragraphs 82, “the resulting current through the free layer 502, the barrier layer 504, and the reference layer 506 may be measured or sensed to detect the resistance of the MTJ [i.e., magnetic field measurement for magnetic tunnel junction/MTJ] … MRAM die 150, neuromorphic computing die 450, or the like may include sense amplifiers … to convert a low power signal to a logic level representing a data value”, 86, “MTJs 500, sensing components such as sense amplifiers”, 101, “for an MTJ 500. A "pinning strength" for a location as used herein, may refer to any measurement” and 108-109, “parallel-magnetized volume has increased relative to the high resistance state of FIG. 6A, the resistance of the MTJ 500 has decreased”, “the parallel-magnetized domain has expanded in response to an increased write current from the controller.” [i.e., parts/portions of MTJs include first and second amplifier components/controllers that sense/measure magnetic field strength and amplify/increase a signal/generate signal gain based on resistance]).
Schneider and Choi are analogous art because they are both related to neuromorphic computing and implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, and Choi, paragraph 4).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider to incorporate the teachings of Choi to provide “systems and methods for magnetoresistive random access memory” where “A magnetic tunnel junction for storing data may include a reference layer, a barrier layer, and a free layer … for neuromorphic computing” (See, e.g., Choi, Abstract and paragraphs 3-4). Doing so would have allowed Schneider to use Choi’s “plurality of magnetic tunnel junctions (MTJs) for storing data. In certain embodiments, an MTJ includes a reference layer, a barrier layer, and a free layer” as components of “an artificial neural network [that] includes an artificial neuron 306, and a plurality of synapses”, as suggested by Choi (See, e.g., Choi, paragraphs 47 and 52).
Regarding claim 11, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 7.
Schneider further discloses wherein the trigger unit16 is a threshold-exceeded starter configured to conduct electrical current when an accumulated input signal reaches or exceeds a threshold voltage (see, e.g., paragraphs 90-91, “The JJ on the right side of the circuit acts as the output threshold "neuron". This JJ is also an MJJ but with a higher range of critical currents”, “the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy … The output JJ has a small constant DC bias that is below the spiking threshold”, 92, “As can be seen in both the voltage and phase traces, the input and output JJs are firing in this configuration” and 94, “inputs are connected to 27 synaptic MJJs which have their critical currents adjusted to form the weights of the middle layer. These are then connected to 3 threshold MJJs forming the output layer.” [i.e., some JJs are trigger units that fire/are triggered and conduct electrical current when an accumulated input current signal meets a threshold]).
Regarding claim 12, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 7.
Schneider further discloses wherein the input pre-processing unit17 comprises a shunt unit connected to the signal differential module and the second magnetoresistance and amplification unit (see, e.g., paragraphs 49, “neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and units/modules], 70, “A modified version of the resistively and capacitively shunted junction (RCSJ) model was used to simulate the reconfigurable JJ synapses” [i.e., artificial JJ synapses are connected to the shunt junction/unit], 91, “The circuit operates when the input DC bias is turned on. … as the middle junction enters the voltage state, it becomes resistive and more of the input current spike is shunted to the ground above the synaptic JJ.” and 94, “synaptic MJJs which have their critical currents adjusted … These are then connected to 3 threshold MJJs” [i.e., the input pre-processing unit that processes the input DC bias and current includes a shunt junction/unit to shunt the input current spike connected to the signal differentiation and 2nd unit/module with adjustable magnetic orders and resistors – 2nd magnetoresistance unit - via the interconnections/wiring]).
Regarding claim 13, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 2.
Schneider further discloses wherein each of the plurality of magnetic effect artificial neurons comprises:
a signal differential module18 connected to the input pre-processing unit (see, e.g., paragraphs 70, “A modified version of the resistively and capacitively shunted junction (RCSJ) model was used to simulate the reconfigurable JJ synapses” [i.e., artificial JJ synapses are connected to the shunt junction/unit], 82, “control of the magnetic cluster size allows one to vary the amount of energy required to adjust the weight of the artificial synapse.” and 94, “synaptic MJJs which have their critical currents adjusted to form the weights of the middle layer. These are then connected to 3 threshold MJJs forming the output layer.” [i.e., synaptic MJJs include a module that accepts and processes/adjusts signals and feedback/weights]);
a first pulsating direct current module and a second pulsating direct current module19, both of which are connected to the signal differential module (see, e.g., FIG. 12 – depicting positive and negative pulsating voltages [i.e., generated by positive and negative pulsating DC modules] and paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and modules], 89, “a positive voltage pulse across the junction … In the circuits presented here, we set the minimum pulse energy equal to 3 aJ” [i.e., a module/circuit generates positive pulsating DC signal] and 91, “The circuit operates when the input DC bias is turned on. At this point the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy [i.e., pulsating DC current]);
a first magnetoresistance … unit and a second magnetoresistance … unit, both of which are connected to the first pulsating direct current module, the second pulsating direct current module, and the signal differential module20 (see, e.g., paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and units and modules], 90, “The JJ in the center is a ‘synaptic’ MJJ which will weight the transmitted pulse from the input ‘neuron.’ We adjust the critical current of these junctions between 1 μA and 100 μA to achieve the desired ‘synaptic’ weighting. This adjustment would be accomplished physically by adjusting the magnetic order in the MJJ … adjust the range of the critical currents that the MJJ has by changing the size of the junction” and 91, “The resistor and inductor above the middle MJJ act to stabilize the circuit” [i.e., MJJs include first and second junctions/units with adjustable magnetic orders and resistors - magnetoresistance units that are both connected to the first and second direct current modules and the signal differential module via the interconnections/wiring]); and
a trigger unit21 connected between the first magnetoresistance and amplification unit and the second magnetoresistance and amplification unit (see, e.g., paragraphs 91, “the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy” and 92, “As can be seen in both the voltage and phase traces, the input and output JJs are firing in this configuration” [i.e., certain JJs are trigger units that fire/are triggered and are middle JJs/MJJs connected between other units]).
Although Schneider substantially discloses the claimed invention, Schneider is not relied on for explicitly disclosing a first magnetoresistance and amplification unit and a second magnetoresistance and amplification unit, wherein the first magnetoresistance and amplification unit is disposed inside the middle head and the second magnetoresistance and amplification unit is disposed inside the middle tail.
In the same field, analogous art Choi teaches a first magnetoresistance and amplification unit and a second magnetoresistance and amplification unit22, wherein the first magnetoresistance and amplification unit is disposed inside the middle head and the second magnetoresistance and amplification unit is disposed inside the middle tail (see, e.g., FIG. 5 – depicting a magnetic tunnel junctions (MTJ) 500 with units in middle layers 502, 504, 506 between terminals 522, 524, and paragraphs 47, “MRAM die, including a plurality of magnetic tunnel junctions (MTJs)” [i.e., magnetic/electromagnetic components], 82, “the resulting current through the free layer 502, the barrier layer 504, and the reference layer 506 may be measured or sensed to detect the resistance of the MTJ [i.e., magnetic measurement for magnetic tunnel junction/MTJ] … MRAM die 150, neuromorphic computing die 450, or the like may include sense amplifiers, latches, and the like, to convert a low power signal to a logic level representing a data value” and 86, “MTJs 500, sensing components such as sense amplifiers” [i.e., middle layer parts/portions of MTJs include first and second amplifier components that sense/measure and amplify a signal]).
Schneider and Choi are analogous art because they are both related to neuromorphic computing and implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, and Choi, paragraph 4).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider to incorporate the teachings of Choi to provide “systems and methods for magnetoresistive random access memory” where “A magnetic tunnel junction for storing data may include a reference layer, a barrier layer, and a free layer … for neuromorphic computing” (See, e.g., Choi, Abstract and paragraphs 3-4). Doing so would have allowed Schneider to use Choi’s “plurality of magnetic tunnel junctions (MTJs) for storing data. In certain embodiments, an MTJ includes a reference layer, a barrier layer, and a free layer” as components of “an artificial neural network [that] includes an artificial neuron 306, and a plurality of synapses”, as suggested by Choi (See, e.g., Choi, paragraphs 47 and 52).
Regarding claim 14, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 13.
Schneider further discloses wherein the trigger unit23 is disposed inside the middle body (see, e.g., paragraphs 91, “the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy” and 92, “As can be seen in both the voltage and phase traces, the input and output JJs are firing in this configuration” [i.e., certain JJs are trigger units that fire/are triggered and are middle JJs/MJJs disposed in the middle body, between other units]).
Regarding claim 15, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 3.
Schneider further discloses wherein each of the plurality of magnetic effect artificial neurons comprises:
a signal differential module24 connected to the input pre-processing unit (see, e.g., paragraphs 70, “A modified version of the resistively and capacitively shunted junction (RCSJ) model was used to simulate the reconfigurable JJ synapses” [i.e., artificial JJ synapses are connected to the shunt junction/unit], 82, “control of the magnetic cluster size allows one to vary the amount of energy required to adjust the weight of the artificial synapse.” and 94, “synaptic MJJs which have their critical currents adjusted to form the weights of the middle layer. These are then connected to 3 threshold MJJs forming the output layer.” [i.e., synaptic MJJs include a module that accepts and processes/adjusts signals and feedback/weights]);
a first pulsating direct current module and a second pulsating direct current module25, both of which are connected to the signal differential module (see, e.g., FIG. 12 – depicting positive and negative pulsating voltages [i.e., generated by positive and negative pulsating DC modules] and paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and modules], 89, “a positive voltage pulse across the junction … In the circuits presented here, we set the minimum pulse energy equal to 3 aJ” [i.e., a module/circuit generates positive pulsating DC signal] and 91, “The circuit operates when the input DC bias is turned on. At this point the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy [i.e., pulsating DC current]);
a first magnetoresistance … unit and a second magnetoresistance … unit, both of which are connected to the first pulsating direct current module, the second pulsating direct current module, and the signal differential module26 (see, e.g., paragraphs 49, “A plurality of neural members 100 can be interconnected by superconducting wiring” [i.e., interconnections connect neural members and units and modules], 90, “The JJ in the center is a ‘synaptic’ MJJ which will weight the transmitted pulse from the input ‘neuron.’ We adjust the critical current of these junctions between 1 μA and 100 μA to achieve the desired ‘synaptic’ weighting. This adjustment would be accomplished physically by adjusting the magnetic order in the MJJ … adjust the range of the critical currents that the MJJ has by changing the size of the junction” and 91, “The resistor and inductor above the middle MJJ act to stabilize the circuit” [i.e., MJJs include first and second junctions/units with adjustable magnetic orders and resistors - magnetoresistance units that are both connected to the first and second direct current modules and the signal differential module via the interconnections/wiring]); and
a trigger unit27 connected between the first magnetoresistance and amplification unit and the second magnetoresistance and amplification unit (see, e.g., paragraphs 91, “the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy” and 92, “As can be seen in both the voltage and phase traces, the input and output JJs are firing in this configuration” [i.e., certain JJs are trigger units that fire/are triggered and are middle JJs/MJJs connected between other units]).
Although Schneider substantially discloses the claimed invention, Schneider is not relied on for explicitly disclosing a first magnetoresistance and amplification unit and a second magnetoresistance and amplification unit, wherein the first magnetoresistance and amplification unit is disposed inside the middle head and the second magnetoresistance and amplification unit is disposed inside the middle tail.
In the same field, analogous art Choi teaches a first magnetoresistance and amplification unit and a second magnetoresistance and amplification unit28, wherein the first magnetoresistance and amplification unit is disposed inside the middle head and the second magnetoresistance and amplification unit is disposed inside the middle tail (see, e.g., FIG. 5 – depicting a magnetic tunnel junctions (MTJ) 500 with units in middle layers 502, 504, 506 between terminals 522, 524, and paragraphs 47, “MRAM die, including a plurality of magnetic tunnel junctions (MTJs)” [i.e., magnetic/electromagnetic components], 82, “the resulting current through the free layer 502, the barrier layer 504, and the reference layer 506 may be measured or sensed to detect the resistance of the MTJ [i.e., magnetic measurement for magnetic tunnel junction/MTJ] … MRAM die 150, neuromorphic computing die 450, or the like may include sense amplifiers, latches, and the like, to convert a low power signal to a logic level representing a data value” and 86, “MTJs 500, sensing components such as sense amplifiers” [i.e., middle layer parts/portions of MTJs include first and second amplifier components that sense/measure and amplify a signal]).
Schneider and Choi are analogous art because they are both related to neuromorphic computing and implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, and Choi, paragraph 4).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider to incorporate the teachings of Choi to provide “systems and methods for magnetoresistive random access memory” where “A magnetic tunnel junction for storing data may include a reference layer, a barrier layer, and a free layer … for neuromorphic computing” (See, e.g., Choi, Abstract and paragraphs 3-4). Doing so would have allowed Schneider to use Choi’s “plurality of magnetic tunnel junctions (MTJs) for storing data. In certain embodiments, an MTJ includes a reference layer, a barrier layer, and a free layer” as components of “an artificial neural network [that] includes an artificial neuron 306, and a plurality of synapses”, as suggested by Choi (See, e.g., Choi, paragraphs 47 and 52).
Regarding claim 16, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 15.
Schneider further discloses wherein the trigger unit29 is disposed inside the middle body (see, e.g., paragraphs 91, “the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ. This MJJ will fire with a pulse energy” and 92, “As can be seen in both the voltage and phase traces, the input and output JJs are firing in this configuration” [i.e., certain JJs are trigger units that fire/are triggered and are middle JJs/MJJs disposed in the middle body, between other units]).
Claim 6 is rejected over Schneider in view and further in view of Friedman as applied to claim 1 above, and further in view of Grollier (U.S. Patent Application Pub. No 2021/0201123 A1, hereinafter “Grollier”).
Regarding claim 6, as discussed above, Schneider in view of Choi and Friedman teaches the system of claim 1.
Although Schneider in view of Choi and Friedman substantially teaches the claimed invention, and Schneider discloses “The circuit operates when the input DC bias is turned on. At this point the JJ on the left will start firing with a frequency of roughly 0.35 GHz. These pulses are then seen as current pulses by the middle synaptic MJJ” [i.e., circuit includes component converting an input DC signal to a pulsating, alternating current/AC signal] (see, e.g., paragraph 91), Schneider in view of Choi and Friedman is not relied on to teach wherein the input pre-processing unit comprises a rectifier capable of converting a direct current signal to an alternating current signal.
In the same field, analogous art Grollier teaches wherein the input pre-processing unit comprises a rectifier capable of converting a direct current signal to an alternating current signal (see, e.g., paragraph 18, “input applied to the layers of neurons is a direct voltage and the output of the layers of neurons is an alternating current. Thus, the neurons of a lower layer send alternating current to the synaptic chains of the interconnection. The rectification circuit allows the signals at the terminals of the synaptic chains to be rectified. The rectification circuit then creates a direct voltage that is applied to the upper neuron layer.” [i.e., input processing circuit/unit for processing applied input includes a rectifier/rectification circuit capable of converting a direct current signal to an alternating current signal]).
Schneider, Choi, Friedman and Grollier are analogous art because they are each related to implementing magnetic neural devices (See, e.g., Schneider, Abstract and paragraph 61, Choi, paragraph 4, Friedman, Abstract and paragraphs 12-13, and Grollier, paragraph 53).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Schneider in view of Choi and Friedman to incorporate the teachings of Grollier to provide “a neural network comprising synaptic chains, each synaptic chain comprising synapses, … a lower layer being connected to an upper layer by an interconnection comprising an assembly of synaptic chains connected to rectification circuits” where “The rectification circuit allows the signals at the terminals of the synaptic chains to be rectified.” (See, e.g., Grollier, paragraphs 15 and 18). Doing so would have allowed Schneider in view of Choi and Friedman to use Grollier’s neural network including rectification circuits because “Such a neural network enables memory and computing to come closer together, to create fast, low-power neural networks suitable for learning in real time”, as suggested by Grollier (See, e.g., Grollier, paragraph 17).
Allowable Subject Matter
Upon overcoming of all the objections and rejections as discussed above in items 6-11, claim 9 is objected to as being dependent upon a rejected base claim (i.e., claim 1), but would be allowable if amended to address the above-noted objections and rejections under 35 U.S.C. 112(b) and 103, and rewritten in independent form including all of the limitations of the base claims (i.e., claim 1) and any intervening claims (i.e., claim 7).
For example, with regard to dependent claim 9, the prior art of record does not anticipate, nor do they render obvious in any reasonable combination to one of ordinary skill in the art at the time of Applicants' invention, the combination of recited limitations of claim 9, its base claim, independent claim 1, and its intervening claim, dependent claim 7.
As discussed above, Schneider in view of Choi and Friedman teaches the systems of claims 1 and 7.
However, the prior art of record does not anticipate or render obvious the limitations “wherein the first pulsating direct current module is a semiconductor diode bridge configured to generate a pulsating direct current with positive polarity, and the second pulsating direct current module is a semiconductor diode bridge configured to generate a pulsating direct current with negative polarity” as recited dependent claim 9 in combination with limitations of base claim 1 and intervening claim 7.
Conclusion
The prior art made of record, listed on form PTO-892, and not relied upon, is considered pertinent to applicant's disclosure.
The references listed on form PTO-892 are all generally related to hardware implementations of neurons, neuromorphic computing and implementing magnetic neural devices for artificial intelligence systems.
For example, non-patent literature Li et al. (“Magnetic skyrmion-based artificial neuron device." Nanotechnology 28.31 (2017): 31LT01: 1-7, hereinafter “Li”) discloses techniques “for developing artificial synapses and neurons” that include “magnetic skyrmions in neuromorphic computing design” using “a skyrmion-based artificial neuron by exploiting the tunable current-driven skyrmion motion dynamics, mimicking the leaky-integrate-fire function of a biological neuron. With a simple single-device implementation, this proposed artificial neuron may enable us to build a dense and energy-efficient spiking neuromorphic computing system” where “the input spikes are stochastic and non-homogeneous, … Once the membrane potential reaches a definite threshold value, the neuron will fire an output spike” and “When the membrane potential Vmem(t) reaches a given threshold, the neuron will ‘fire’ an output spike” [i.e., an output unit/neuron fires an output spike after matching input signals/spikes to a given, definite threshold value/abstract information] (see, Abstract and pages 1-3).
Also, for example, non-patent literature Ross et al. et al. (“Multilayer spintronic neural networks with radio-frequency connections." arXiv preprint arXiv:2211.03659 (2022), hereinafter “Ross”) discloses a “Spin-diode response of a chain of two synaptic junctions connected in series: rectified DC voltage versus frequency of the input signal” and “A full network can be broken into subsets … consisting of inputs coming from a layer of N oscillators which are amplified and passed to M chains of N diodes which perform the synaptic operation. Each chain output is then passed to an oscillator in the next layer. We have two layers of amplifiers, the first is an RF amplifier between the N oscillators and M chains and then a second layer of DC amplifiers between the M chains and M oscillators of the next layer.” (see, e.g., pages 4 and 25).
The examiner requests, in response to this office action, support be shown for language added to any original claims on amendment and any new claims. That is, indicate support for newly added claim language by specifically pointing to page(s) and line no(s) in the specification and/or drawing figure(s). This will assist the examiner in prosecuting the application.
When responding to this office action, Applicant is advised to clearly point out the patentable novelty which he or she thinks the claims present, in view of the state of the art disclosed by the reference cited or the objections made. He or she must also show how the amendments avoid such references or objections See 37 CFR 1.111 (c).
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RANDY K BALDWIN whose telephone number is (571)270-5222. The examiner can normally be reached on Mon - Fri 9:00-6:00.
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, Kamran Afshar can be reached on 571-272-7796. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/RANDALL K. BALDWIN/Primary Examiner, Art Unit 2125
1 Regarding the “artificial intelligence system”, applicant’s specification merely repeats the claim language in paragraphs 7-8 and 60 in stating “a magnetic effect artificial intelligence system, which employs the property of magnetism to simulate human neurons to learn, to store and to retrieve information”, “The magnetic effect artificial intelligence system includes an input pre-processing unit, a plurality of magnetic effect artificial neurons connected with the input pre-processing unit, and an output unit connected with the plurality of magnetic effect artificial neurons.” and “the magnetic effect artificial intelligence system includes the magnetic effect artificial neurons”, and provides general examples in paragraphs 30-31 in stating “the magnetic effect artificial intelligence system includes magnetic effect artificial neurons made of ferromagnetic or ferrite materials.” and “The magnetic effect artificial intelligence system utilizes the magnetic fields and electromagnetic property to train the neurons with feedback signals”. Therefore, “a magnetic effect artificial intelligence system”, under the broadest reasonable interpretation (BRI), in light of the specification, is any system including artificial neurons or synapses having magnetic properties or effects.
2 Paragraphs 19 and 42 of applicant’s specification state “the input pre-processing unit includes a shunt unit connected to the signal differential module and the second magnetoresistance and amplification unit” and “The input pre-processing unit 2 is to pre-process the input signal. For example, the input pre-processing unit 2 performs processes of accepting input sources, connecting circuits from the SU 21 to the SDM 34 and the MRA2 39 of each neuron 3, and converting the input signal to be alternating current (AC) signal.” Therefore, an “input pre-processing unit”, under the BRI, in light of the specification, is any unit, element, module or component that accepts and processes, converts or transforms input signals or data)
3 Paragraphs 35-37 of applicant’s specification state that different “parts” are “made of a material with high magnetic permeability and low magnetic saturation” and are “made of Mu-metal (μ-metal), which is a nickel-iron soft ferromagnetic alloy.” The plain meaning of “ferrite” is “the pure iron constituent of ferrous metals” or “any of a group of ferromagnetic highly resistive ceramic compounds” See https://www.dictionary.com/browse/ferrite. Therefore, a “structure made of Mu-metal and ferrite materials”, under the BRI, in light of the specification, is any structure with parts or components made of metal material having magnetic permeability and magnetic saturation, a nickel-iron (NiFe) allow, and ferromagnetic compounds or materials.
4 As indicated above in the section 112(b) rejection of this claim, “wherein each of the plurality of magnetic effect artificial neurons … substantially attaches to adjacent ones of the plurality of magnetic effect artificial neurons” has been interpreted as each of the artificial neurons being attached to, in close proximity to, or very close to adjacent ones of the plurality of magnetic effect artificial neurons.
5 Paragraphs 8 and 32 of applicant’s specification merely repeat the claim language and paragraphs 33 and 60 provide general examples in stating “The hexagonal prism is a three-dimensional geometric structure with two hexagonal bases connected by six rectangular faces. The top and bottom hexagonal bases of the hexagonal prism are in the shape of a hexagon and are congruent with each other.” and “The magnetic effect artificial neuron is shaped as a three-layered hexagonal”. These are the only mentions in the specification of any “hexagon” shape or hexagonal structure. Therefore, “shaped as a three-layered hexagonal”, under the BRI, in light of the specification, is any structure having a hexagonal (i.e., 6-sided) shape having three dimensions, such as length, height and width or depth)
6 Paragraphs 35-37 of applicant’s specification state that different “parts” are “made of a material with high magnetic permeability and low magnetic saturation” and are “made of Mu-metal (μ-metal), which is a nickel-iron soft ferromagnetic alloy.” The plain meaning of “ferrite” is “the pure iron constituent of ferrous metals” or “any of a group of ferromagnetic highly resistive ceramic compounds” See https://www.dictionary.com/browse/ferrite. Therefore, a “middle head and the middle tail are made of ferrite , and the middle body is made of Mu-metal”, under the BRI, in light of the specification, is any structure with the recited parts or components made of metal material having magnetic permeability and magnetic saturation, and ferromagnetic compounds or materials.
7 Paragraph 35 of applicant’s specification states that the “top head (TH) 311, a top body (TB) 312 and a top tail (TT) 313. Each of the TH 311, the TB 312 and the TT 313 is made of a material with high
magnetic permeability and low magnetic saturation. In an embodiment, each of the TH 311, the TB 312 and the TT 313 is made of Mu-metal (μ-metal), which is a nickel-iron soft ferromagnetic alloy.” Therefore, top layer parts “made of Mu-metal”, under the BRI, in light of the specification, are any top or upper layer parts or components made of metal material having magnetic permeability and saturation, and/or a nickel-iron (NiFe) ferromagnetic alloy.
8 Paragraphs 35 and 37 of applicant’s specification state that parts are “made of Mu-metal (μ-metal), which is a nickel-iron soft ferromagnetic alloy.” and “bottom layer 33 can be identified as three parts including a bottom head (BH) 331, a bottom body (BB) 332 and a bottom tail (BT) 333. Each of the BH 331, the BB 332 and the BT 333 is made of a material with high magnetic permeability and low magnetic saturation. In an embodiment, each of the Each of the BH 331, the BB 332 and the BT 333 is made of Mu-metal (μ-metal).” Therefore, bottom layer parts “made of Mu-metal”, under the BRI, in light of the specification, are any bottom layer parts or components made of metal material having magnetic permeability and saturation, and/or a nickel-iron (NiFe) ferromagnetic alloy.
9 Paragraph 14 of applicant’s specification merely repeats the claim language and paragraphs 15, 40 and 44 provide general examples in stating “the signal differential module is a signal differentiator configured to direct signal to either the first pulsating direct current module or the second pulsating direct current module.” and “the magnetic effect artificial neuron 3 includes a signal differential module (SDM) 34” and “the SDM 34 is a signal differentiator, which includes but not exclusively electronic differential circuits configured to direct the signal to either PDC( +) 35 or PDC(-) 36. The SDM 34 differentiates the two input signals”. Therefore, “a signal differential module”/SDM, under the BRI, in light of the specification, is any module, unit, element or component that accepts, processes or differentiates input signals.
10 Paragraph 14 of applicant’s specification merely repeats the claim language and paragraphs 16, 40 and 45 provide general examples in stating “the first pulsating direct current module is a semiconductor diode bridge configured to generate a pulsating direct current with positive polarity, and the second pulsating direct current module is a semiconductor diode bridge configured to generate a pulsating
direct current with negative polarity.”, “a first pulsating direct current (PDC) module (denoted
by PDC(+)) 35, a second pulsating direct current (PDC) module (denoted by PDC(-))” and “the PDC( +) 35 is a semiconductor diode bridge configured to generate a pulsating DC with positive polarity, while the PDC( -) 36 is a semiconductor diode bridge configured to generate a pulsating DC with negative polarity.” Therefore, first and second “pulsating direct current”/PDC modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal.
11 As indicated above, the first and second “pulsating direct current” modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal and the “signal differential module”, under the BRI, in light of the specification, is any module, unit, element or component that accepts or processes signals
12 Paragraph 14 of applicant’s specification merely repeats the claim language and paragraphs 18, 40 and 46 provide general examples in stating “In an embodiment, the trigger unit is a threshold-exceeded
starter configured to conduct electrical current when an accumulated input signal reaches or exceeds a threshold voltage.”, “a trigger unit (TU) 37” and “In an embodiment, the TU 37 is a threshold-exceeded starter (TES), which includes but not exclusively a diode for alternating current (DIAC) as a trigger device when the strength of signals is accumulated to designated threshold level”. Therefore, “a trigger unit”/TU, under the BRI, in light of the specification is any unit, element or module that can be triggered, fired or activated
13 Paragraph 14 of applicant’s specification merely repeats the claim language and paragraphs 17, 40 and 46 provide general examples in stating “In an embodiment, the first magnetoresistance and
amplification unit and the second magnetoresistance and amplification unit are magnetoresistance amplifiers configured to measure magnetic field strength and work with corresponding resistance to generate signal gain.”, “first magnetoresistance and amplification (MRA) unit (denoted by MRA1) 38, and a second magnetoresistance and amplification (MRA) unit ( denoted by MRA2)” and “In an embodiment, the MRA unit is a magnetoresistance amplifier whose operating is based on changes of its electrical resistance value caused by an externally-applied magnetic field. … MRA1 38 and MRA2 39 are able to measure the magnetic field strength and work with corresponding resistance to generate signal gain.” Therefore, a first and second “magnetoresistance and amplification unit”, under the BRI, in light of the specification are any modules, units, elements or components that measure and amplify a signal (i.e., generate signal gain).
14 As indicated in the objection to this claim above, this recitation is grammatically incorrect and appears to be missing one or more words between “direct” and “signal”. As further indicated above, “a signal differential module”/SDM, under the BRI, in light of the specification, is any module, unit, element or component that accepts, processes or differentiates input signals” and the first and second “pulsating direct current” modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal and the “signal differential module” .
15 As indicated above, a first and second “magnetoresistance and amplification unit”, under the BRI, in light of the specification are any modules, units, elements or components that measure and amplify a signal (i.e., generate signal gain).
16 As indicated above, “a trigger unit”/TU, under the BRI, in light of the specification is any unit, element or module that can be triggered, fired or activated
17 As indicated above, the “input pre-processing unit”, under the BRI, in light of the specification, is any unit, element, module or component that accepts and processes, converts or transforms input signals or data.
18 As indicated above, “a signal differential module”/SDM, under the BRI, in light of the specification, is any module, unit, element or component that accepts, processes or differentiates input signals.
19 As indicated above, first and second “pulsating direct current”/PDC modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal.
20 As indicated above, the first and second “pulsating direct current” modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal and the “signal differential module”, under the BRI, in light of the specification, is any module, unit, element or component that accepts or processes signals
21 As indicated above, “a trigger unit”/TU, under the BRI, in light of the specification is any unit, element or module that can be triggered, fired or activated
22 As indicated above, a first and second “magnetoresistance and amplification unit”, under the BRI, in light of the specification are any modules, units, elements or components that measure and amplify a signal (i.e., generate signal gain).
23 As indicated above, “a trigger unit”/TU, under the BRI, in light of the specification is any unit, element or module that can be triggered, fired or activated.
24 As indicated above, “a signal differential module”/SDM, under the BRI, in light of the specification, is any module, unit, element or component that accepts, processes or differentiates input signals.
25 As indicated above, first and second “pulsating direct current”/PDC modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal.
26 As indicated above, the first and second “pulsating direct current” modules, under the BRI, in light of the specification, are any modules, units, elements or components related to generating a pulsing or pulsating positive or negative direct current/DC signal and the “signal differential module”, under the BRI, in light of the specification, is any module, unit, element or component that accepts or processes signals
27 As indicated above, “a trigger unit”/TU, under the BRI, in light of the specification is any unit, element or module that can be triggered, fired or activated
28 As indicated above, a first and second “magnetoresistance and amplification unit”, under the BRI, in light of the specification are any modules, units, elements or components that measure and amplify a signal (i.e., generate signal gain).
29 As indicated above, “a trigger unit”/TU, under the BRI, in light of the specification is any unit, element or module that can be triggered, fired or activated.