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 .
Claims 1-32 are presented for examination.
Continued Examination under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on April 29, 2026 has been entered.
Response to Amendment
Applicant’s amendment has obviated the rejections under 35 USC § 101 and the specification objections. Therefore, those objections and rejections are withdrawn.
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Claim Objections
Claim 21 is objected to because of the following informalities: “measuring membrane potentials” should be “electrodes for measuring membrane potentials”. Claims 22-27 are objected to for dependency on claim 21. Appropriate correction is required.
Claim Rejections - 35 USC § 103
Claims 1-14, 16-20, and 28-32 are rejected under 35 U.S.C. 103 as being unpatentable over Kasabov, “NeuCube: A Spiking Neural Network Architecture for Mapping, Learning, and Understanding of Spatio-Temporal Brain Data,” in 52 Neural Networks 62-76 (2014) (“Kasabov”) in view of Yanagisawa et al. (US 20210128053) (“Yanagisawa”) and further in view of further in view of Afzali-Ardakani et al. (US 20190164597) (“Afzali”).
Regarding claim 1, Kasabov discloses “[a] method of mapping a natural neural network into an electronic neural network device (computational architecture for the creation of efficient computational models from spatiotemporal brain data is offered [i.e., the method is executed using a processor] – Kasabov, sec. 1.3, last paragraph), the method comprising:
measuring membrane potentials of a plurality of biological neurons of the natural neural network by measuring, for a first timing interval, first membrane potentials and second membrane potentials using … electrodes and measuring, for a second timing interval, the first membrane potentials and the second membrane potentials using same electrodes (electroencephalography is the recording of electrical signals from the brain [over time, implying the existence of two timing intervals] by attaching surface electrodes to the subject’s scalp; these electrodes [i.e., the same set of electrodes] record brain waves; EEGs allow researchers to track electrical potentials [plural, i.e., first and second potentials] across the surface of the brain – Kasabov, sec. 1.2, first paragraph);
constructing a neural network map representing a connection structure and synaptic weights1 of a natural neural network based on the measured membrane potentials (spatiotemporal brain [natural neural network] data can be integrated in a machine learning model – Kasabov, sec. 1.1, last paragraph; EEG channels are spatially distributed on the scalp; the spatially defined positions of the channels can be mapped into a spiking neural network reservoir (SNNr) of a NeuCube model [electronic neural network] using different approaches – id. at sec. 4.1, first paragraph; EEGs allow researchers to track electrical potentials [membrane potentials] across the surface of the brain [biological neurons] – id. at sec. 1.2, first paragraph [since membrane potentials are from neuron to neuron, measuring the potentials measures the connection structure]; brain data are entered into relevant areas of a spiking neural network over time, and unsupervised learning is performed to modify the initially set connection weights – id. at sec. 5.3, first paragraph), where the first membrane potentials and the second membrane potentials correspond to at least two different respective forms of membrane potentials (NeuCube model can learn spatio-temporal patterns in the form of short-term memory, represented as changes of the post-synaptic potentials, and long-term memory, represented as long-term potentiation (LTP) and long-term depression (LTD) – Kasabov, sec. 5.3, first two paragraphs; see also sec. 5.4 and Table 1 (disclosing that the post-synaptic potentials modeled include fast excitation PSPs, slow excitation PSPs, fast inhibition PSPs, and slow inhibition PSPs [different forms of membrane potentials])); and
mapping the neural network map to the electronic neural network device (EEG channels are spatially distributed on the scalp; the spatially defined positions of the channels can be mapped into a spiking neural network reservoir using different approaches – Kasabov, sec. 4.1) by configuring, based on the construed neural network map, either one or both of:
circuit layers of the electronic neural network device configured to perform voltage amplification corresponding to either one or both of accumulation and firing of one or more of the membrane potentials of the biological neurons; and
[memory layers] of the electronic neural network device … corresponding to the synaptic weights in the neural network map (STBD is entered into relevant areas of the SNNr over time; the initially set connection weights are modified through unsupervised learning; a NeuCube model can learn spatio-temporal patterns in its three forms of memory [i.e., the weights are stored in memory and the mapping therefore involves configuring memory layers of the electronic device] – Kasabov, sec. 5.3, first two paragraphs).”
Kasabov appears not to disclose explicitly the further limitations of the claim. However, Yanagisawa discloses “measuring, for a first timing interval, first membrane potentials and second membrane potentials using different electrodes (EEG electrodes are placed on a subject’s head; the electrodes sense voltages [potentials] at the surface of the subject’s skin, and the EEG receiver outputs respective time series information for each of the electrodes over the duration of the measurement [i.e., there is a measurement from one electrode at one time step and a measurement from another electrode at another time step] – Yanagisawa, paragraphs 18-19) ….”
Yanagisawa and the instant application both relate to potential measurements of biological brains and are analogous. 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 Kasabov to measure potentials for different time intervals using different electrodes, as disclosed by Yanagisawa, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would give a more complete picture of the subject’s brain activity than would be the case if the same electrodes were used. See Yanagisawa, paragraphs 18-19.
Neither Kasabov nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays of the electronic neural network device comprising crosspoints comprising memory elements having resistances corresponding to the synaptic weights (synaptic crossbar memory array may have precisely programmed resistance values that may represent synaptic weights – Afzali, paragraph 6; see also paragraph 16 (disclosing that the crossbar comprises crosspoints)) ….”
Afzali and the instant application both relate to neuromorphic devices and are analogous. 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 Kasabov/Yanagisawa to perform learning with a crossbar array, as disclosed by Afzali, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system to provide error-tolerant, massive parallelism, thereby easing the input of complex information. See Afzali, paragraph 2.
Claim 28 is a device claim corresponding to method claim 1 and is rejected for the same reasons as given in the rejection of that claim.
Regarding claim 2, Kasabov/Yanagisawa/Afzali discloses that “the constructing of the neural network map and the mapping of the neural network map are achieved based on respective information of first measured membrane potentials interacting with respective information of second measured membrane potentials for respective pre- and/or post-synaptic relationships among pre-synaptic biological neurons and post-synaptic biological neurons of the natural neural network (different types of STBD have been collected; one of the most common types is an EEG, which is the recording of electrical signals from the brain by attaching surface electrodes to the subject’s scalp; these electrodes record brain waves; EEGs allow researchers to track electrical potentials across the surface of the brain [i.e., pre-/post-synaptic relationships among pre-synaptic and post-synaptic biological neurons] and observe changes taking place over a few milliseconds – Kasabov, sec. 1.2, first paragraph; see also Table 1 (disclosing that there are multiple types of action potential modeled, e.g., a fast excitation PSP [first form of membrane potential] and a fast inhibition PSP [second form of membrane potential]), sec. 4.1 (disclosing the mapping of the EEG signals to the SNNr)), and …
the first measured membrane potentials correspond to a first form of membrane potential of the at least two different respective forms of membrane potentials, and the second measured membrane potentials correspond to a different second form of membrane potential of the at least two different respective forms of membrane potentials (different types of STBD have been collected; one of the most common types is an EEG, which is the recording of electrical signals from the brain by attaching surface electrodes to the subject’s scalp; these electrodes record brain waves; EEGs allow researchers to track electrical potentials across the surface of the brain and observe changes taking place over a few milliseconds – Kasabov, sec. 1.2, first paragraph; see also Table 1 (disclosing that there are multiple types of action potential modeled, e.g., a fast excitation PSP [first form of membrane potential] and a fast inhibition PSP [second form of membrane potential])).”
Claim 32 is a device claim corresponding to method claim 2 and is rejected for the same reasons as given in the rejection of that claim.
Regarding claim 3, Kasabov/Yanagisawa/Afzali discloses that “the constructing comprises:
identifying the connection structure among the plurality of biological neurons (EEG channels are spatially distributed on the scalp; 64 EEG channels are positioned on a human head; the spatially defined positions of the channels [i.e., the connection structure of the neurons measured by the EEG] is mapped onto an SNNr [after identification] – Kasabov, sec. 4.1, first paragraph); and
estimating the synaptic weights for connections between multiple biological neurons of the plurality of biological neurons (in unsupervised learning, the STBD [i.e., connections between biological neurons] is entered into relevant areas of the SNNr over time; unsupervised learning is performed to modify [estimate] the initially set connection weights – Kasabov, sec. 5.3, first paragraph).”
Claim 29 is a device claim corresponding to method claim 3 and is rejected for the same reasons as given in the rejection of that claim.
Regarding claim 4, Kasabov/Yanagisawa/Afzali discloses that “the estimating of the synaptic weights is based on a result of the identifying of the connection structure (in unsupervised learning, the STBD [i.e., connections between biological neurons] is entered into relevant areas of the SNNr over time; unsupervised learning is performed to modify [estimate] the initially set connection weights – Kasabov, sec. 5.3, first paragraph).”
Regarding claim 5, Kasabov/Yanagisawa/Afzali discloses “measuring membrane potentials of the plurality of biological neurons over time (EEGs allow researchers to track [measure] electrical potentials [membrane potentials] across the surface of the brain [biological neurons] and observe changes taking place over a few milliseconds [time] – Kasabov, sec. 1.1, last paragraph);
extracting action potentials (APs), of the plurality of biological neurons, from action potential results of the measuring of the membrane potentials (Kasabov Table 1 shows neuronal action potential parameters in a computational neuro-genetic model of a spiking neuron; first paragraph of sec. 5.4 discloses that the neuro-genetic model is used within the NeuCube architecture to integrate different types of STBD [i.e., the action potentials in the model are derived from their counterparts in the brain]); and
extracting post-synaptic potentials (PSPs), of the plurality of biological neurons, from post-synaptic potential results of the measuring of the membrane potentials (Kasabov Table 1 shows neuronal action potential parameters, including multiple types of post-synaptic potential parameters, in a computational neuro-genetic model of a spiking neuron; first paragraph of sec. 5.4 discloses that the neuro-genetic model is used within the NeuCube architecture to integrate different types of STBD [i.e., the action potentials in the model are derived from their counterparts in the brain]).”
Claim 31 is a device claim corresponding to method claim 5 and is rejected for the same reasons as given in the rejection of that claim.2
Regarding claim 6, Kasabov/Yanagisawa/Afzali discloses that “the measuring of the membrane potentials of the plurality of biological neurons includes measuring intracellular membrane potentials of the plurality of biological neurons using intracellular electrodes (EEG is the recording of electrical signals in the brain by attaching surface electrodes to the subject’s scalp; these electrodes record brain waves which are electrical signals naturally produced by the brain; EEGs allow researchers to track electrical potentials across the surface of the brain and observe changes taking place over a few milliseconds – Kasabov, sec. 1.2, first paragraph [note that the measurement of brain waves across the brain includes measuring potentials within each cell]).”
Regarding claim 7, Kasabov/Yanagisawa/Afzali discloses that “the identifying of the connection structure comprises identifying the connection structure among the plurality of biological neurons based on respective timings of the APs and respective timings of the PSPs (STDP learning rule utilizes Hebbian plasticity in the form of LTP and LTD; efficiency of synapses is strengthened or weakened based on the timing of post-synaptic action potential in relation to the pre-synaptic spike; presynaptic activity [AP] that precedes post-synaptic firing [PSP] can induce long-term potentiation; reversing this temporal order causes long-term depression – Kasabov, sec. 2.2, second paragraph; SNNr has 1471 neurons and the coordinates of these neurons correspond directly to the Talairach template coordinates with a resolution of 1 cm3 [identifying the coordinates of the neurons = identifying connection structure among the biological neurons] – id. at sec. 3.3, last paragraph; spiking sequences representing EEG channels are entered into the correspondingly located neurons before STDP is applied [i.e., the connection structure includes the timing constraints of STDP] – id. at sec. 4.1).”
Regarding claim 8, Kasabov/Yanagisawa/Afzali discloses that “the identifying of the connection structure comprises determining pre- and/or post-synaptic relationships among pre-synaptic neurons and post-synaptic neurons of the plurality of biological neurons (in order to integrate different types of STBD, a new model of a spiking neuron can be used in the NeuCube architecture, known as a probabilistic neurogenetic model; this model incorporates different types of action potential, including fast and slow excitation and fast and slow inhibition PSPs; total PSP is calculated based on a formula that includes a term that is 1 if a spike has been emitted from a given neuron [i.e., a spike from a pre-synaptic to a post-synaptic neuron, or a relationship between the two] and 0 otherwise – Kasabov, sec. 5.4, second through fourth paragraphs and Table 1).”
Regarding claim 9, Kasabov/Yanagisawa/Afzali discloses that “the estimating of the synaptic weights comprises estimating the synaptic weights for connections between the pre-synaptic neurons and the post-synaptic neurons based on respective PSPs of the post-synaptic neurons and respective APs of the pre-synaptic neurons (EEG is the recording of electrical signals from the brain by attaching surface electrodes to the subject’s scalp; EEGs allow researchers to track electrical potentials [i.e., PSPs and APs] across the surface of the brain and observe changes taking place over a few milliseconds – Kasabov, sec. 1.2, first paragraph; STBD are entered into relevant areas of the SNNr over time, and the initially set connection weights are modified – id. at sec. 5.3, first paragraph [i.e., the estimation/modification of the weights is based on the STBD including the potentials measured by the EEG]; see also Table 1 (disclosing action potentials and post-synaptic potentials)).”
Regarding claim 10, Kasabov/Yanagisawa discloses that “the mapping comprises:
mapping the plurality of biological neurons to the circuit layers of the electronic neural network device (64 EEG channels [measuring biological neurons] are spatially distributed on a scalp, and the spatially defined positions of the channels can be mapped onto an SNNr [circuit layers of electronic NN device] – Kasabov, sec. 4.1, first paragraph); and
mapping the synaptic weights and corresponding connectivities among the plurality of biological neurons to the [memory layers] of the electronic neural network device (STBD is entered into relevant areas of the SNNr over time; the initially set connection weights are modified through unsupervised learning [note that the weights also contain information about connectivity because nonzero weights imply a connection and zero weights imply no connection]; a NeuCube model can learn spatio-temporal patterns in its three forms of memory [i.e., the weights are stored in memory] – Kasabov, sec. 5.3, first two paragraphs).”
Neither Kasabov nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays”, as shown in the rejection of claim 1. 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 the combination of Kasabov and Yanagisawa to employ crossbar arrays, as disclosed by Afzali, for substantially the same reasons as given in the rejection of claim 1.
Claim 30 is a device claim corresponding to method claim 10 and is rejected for the same reasons as given in the rejection of that claim.
Regarding claim 11, Kasabov/Yanagisawa/Afzali discloses that “the constructing of the neural network map and the mapping of the neural network map implement learning of the electronic neural network device (64 EEG channels are positioned on a human head, and the spatially defined positions of the channels are mapped onto a SNNr; neurons in SNNr are located following the same coordinates of the Talairach template of the brain and EEG channels are mapped according to the standard mapping; the spiking sequences that represent EEG channels are entered into the correspondingly located neurons before a training [learning] procedure is applied – Kasabov, sec. 4.1), and …
the method further comprises:
obtaining an input or stimuli (after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr – Kasabov, sec. 3.4, first paragraph);
activating the learned electronic neural network device, provided the obtained input or stimuli, to perform neural network operations (after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr [i.e., the SNNr is activated by the data] and the state of the SNNr is measured for each pattern and an output classifier is trained to recognize this state in a predefined output class for this input pattern [recognizing the state = neural network operation] – Kasabov, sec. 3.4, first paragraph); and
generating a neural network result for the obtained input or stimuli based on a result of the activated learned electronic neural device (after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr and the state of the SNNr is measured for each pattern and an output classifier is trained to recognize this state in a predefined output class [neural network result] for this input pattern – Kasabov, sec. 3.4, first paragraph).”
Regarding claim 12, Kasabov/Yanagisawa/Afzali discloses “[a] non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to implement the method of claim 1 (computational architecture for the creation of efficient computational models from spatiotemporal brain data is offered [i.e., the method is executed using a processor] – Kasabov, sec. 1.3, last paragraph; NeuCube model can learn spatiotemporal patterns in three forms of memory [storage medium storing instructions] – id. at sec. 5.3, second paragraph).”
Regarding claim 13, Kasabov discloses “[a] method for generating a neural network result, by an electronic device, using a learned electronic neural network device with learned synaptic connections and synaptic weights having characteristics of the learned electronic neural network device having been mapped from a natural neural network based on respective information of measured action potentials (APs) interacting with respective information of measured post-synaptic potentials (PSPs) for respective pre- and/or post-synaptic relationships among pre-synaptic biological neurons and post-synaptic biological neurons of the natural neural network (spatially defined positions of channels on an EEG [from a brain/natural neural network] can be mapped to an SNNr [electronic neural network device] – Kasabov, sec. 4.1, first paragraph; STBD is entered into relevant areas of the SNNr over time; unsupervised learning is performed to modify the initially set connection weights [i.e., the system has learned synaptic connections/weights that characterize the device] – id. at sec. 5.3, first bullet point; prior bioinformatics and neuroinformatics knowledge can be used when designing neurogenetic NeuCube models; for example, post-synaptic AMPA-type glutamate receptors mediate fast excitatory synaptic transmissions and are crucial for many aspects of brain functioning [i.e., information of post-synaptic potentials of the brain are used in mapping the brain to the ANN] – id. at second full paragraph on right-hand column of p. 72; probabilistic parameters of the model have their biological analogues; for example, the probability of a synapse to contribute to the post-synaptic potential after it has received a spike from a pre-synaptic neuron may be affected by different proteins [i.e., the system models pre-/post-synaptic relationships within the brain, or the mapping is based on these relationships] – id. at p. 73, second full paragraph; see also Table 1 (characterizing the PSPs as types of action potential)), the method comprising:
obtaining an input or stimuli (after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr – Kasabov, sec. 3.4, first paragraph) by measuring, for a respective first timing interval, the APs and the PSPs using … electrodes and measuring, for a respective different second timing interval, the APs and the PSPs using same electrodes (electroencephalography is the recording of electrical signals from the brain [over time, implying the existence of two timing intervals] by attaching surface electrodes to the subject’s scalp; these electrodes [i.e., the same set of electrodes] record brain waves; EEGs allow researchers to track electrical potentials [plural, i.e., first and second potentials] across the surface of the brain – Kasabov, sec. 1.2, first paragraph);
activating the learned electronic neural network device, provided the obtained input or stimuli, to perform neural network operations (after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr [i.e., the SNNr is activated by the data] and the state of the SNNr is measured for each pattern and an output classifier is trained to recognize this state in a predefined output class for this input pattern [recognizing the state = neural network operation] – Kasabov, sec. 3.4, first paragraph); and
generate, using either one or both of circuit layers of the electronic neural network device configured to perform voltage amplification corresponding to either one or both of accumulation and firing of one or more of the measured APs and the measured PSPs and [memory layers] of the learned electronic neural network device … corresponding to the synaptic weights in the neural network map configured based on the mapping, the neural network result for the obtained input or stimuli based on a result of the activated learned electronic neural device (after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr and the state of the SNNr is measured for each pattern and an output classifier is trained to recognize this state in a predefined output class [neural network result] for this input pattern – Kasabov, sec. 3.4, first paragraph; STBD is entered into relevant areas of the SNNr over time; the initially set connection weights are modified through unsupervised learning; a NeuCube model can learn spatio-temporal patterns in its three forms of memory [i.e., the weights are stored in memory and the mapping therefore involves configuring memory layers of the electronic device] – id. at sec. 5.3, first two paragraphs).”
Kasabov appears not to disclose explicitly the further limitations of the claim. However, Yanagisawa discloses “measuring, for a respective first timing interval, the [potentials] … using different electrodes (EEG electrodes are placed on a subject’s head; the electrodes sense voltages [potentials] at the surface of the subject’s skin, and the EEG receiver outputs respective time series information for each of the electrodes over the duration of the measurement [i.e., there is a measurement from one electrode at one time step and a measurement from another electrode at another time step] – Yanagisawa, paragraphs 18-19) ….” 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 Kasabov to measure potentials for different time intervals using different electrodes, as disclosed by Yanagisawa, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would give a more complete picture of the subject’s brain activity than would be the case if the same electrodes were used. See Yanagisawa, paragraphs 18-19.
Neither Kasabov nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays of the learned electronic neural network device comprising crosspoints comprising memory elements having resistances corresponding to the synaptic weights (synaptic crossbar memory array may have precisely programmed resistance values that may represent synaptic weights – Afzali, paragraph 6; see also paragraph 16 (disclosing that the crossbar comprises crosspoints)) ….” 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 Kasabov/Yanagisawa to perform learning with a crossbar array, as disclosed by Afzali, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system to provide error-tolerant, massive parallelism, thereby easing the input of complex information. See Afzali, paragraph 2.
Regarding claim 14, the rejection of claim 13 is incorporated. Kasabov further discloses “measuring, using first plural electrodes, the APs (EEG is the recording of electrical signals from the brain by attaching surface electrodes to the subject’s scalp; these electrodes record brain waves; EEGs allow researchers to track electrical potentials across the surface of the brain and observe changes taking place over a few milliseconds – Kasabov, sec. 1.2, first paragraph; see also Table 1 (giving APs and PSPs as examples of potentials that can be measured [i.e., in general, two or more electrodes can measure the APs and two or more electrodes can measure the PSPs]));
measuring, using second plural electrodes, the PSPs (EEG is the recording of electrical signals from the brain by attaching surface electrodes to the subject’s scalp; these electrodes record brain waves; EEGs allow researchers to track electrical potentials across the surface of the brain and observe changes taking place over a few milliseconds – Kasabov, sec. 1.2, first paragraph; see also Table 1 (giving APs and PSPs as examples of potentials that can be measured [i.e., in general, two or more electrodes can measure the APs and two or more electrodes can measure the PSPs])); and
performing learning of the electronic neural network device by constructing, by the electronic neural network device, a neural network map of the natural neural network based on respective information of the measured APs interacting with respective information of the measured PSPs (EEGs record brain waves by attaching surface electrodes to a subject’s scalp and allow researchers to track electrical potentials [i.e., APs and PSPs] across the surface of the brain – Kasabov, sec. 1.2, first paragraph; see also sec. 4.1 (disclosing the mapping of the EEG channels onto an SNNr [electronic neural network device]), sec. 5.3, first paragraph (disclosing the modification of the weights [i.e., learning] of the SNNr using the STBD), Table 1 (disclosing APs and PSPs)) ….”
Kasabov/Yanagisawa appears not to disclose explicitly the further limitations of the claim. However, Afzali discloses “performing learning of the electronic neural network device … using corresponding crosslinks of a crossbar (synaptic crossbar or crosspoint memory array includes multiple lower electrode lines in one direction, multiple upper electrode lines in an orthogonal direction, and an electronic synapse at each intersection [crosslinks] of the electrode lines – Afzali, paragraph 16; resistance values may represent synaptic weights [for learning] – id. at paragraph 6).” 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 Kasabov/Yanagisawa to perform learning with a crossbar array, as disclosed by Afzali, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system to provide error-tolerant, massive parallelism, thereby easing the input of complex information. See Afzali, paragraph 2.
Regarding claim 16, Kasabov discloses “[a] non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to perform the method of claim 13 (computational architecture for the creation of efficient computational models from spatiotemporal brain data is offered [i.e., the method is executed using a processor] – Kasabov, sec. 1.3, last paragraph; NeuCube model can learn spatiotemporal patterns in three forms of memory [storage medium storing instructions] – id. at sec. 5.3, second paragraph).”
Regarding claim 17, Kasabov discloses “[a] method of mapping a natural neural network into an electronic neural network device, the method comprising:
measuring membrane potentials of a plurality of biological neurons of a natural neural network by measuring, for a first timing interval, first membrane potentials and second membrane potentials using … electrodes and measuring, for a second timing interval, the first membrane potentials and the second membrane potentials using same electrodes (electroencephalography is the recording of electrical signals from the brain [over time, implying the existence of two timing intervals] by attaching surface electrodes to the subject’s scalp; these electrodes [i.e., the same set of electrodes] record brain waves; EEGs allow researchers to track electrical potentials [plural, i.e., first and second potentials] across the surface of the brain – Kasabov, sec. 1.2, first paragraph);
considering, using a plurality of neuron modules of the electronic neural network device, the measured first membrane potentials and the measured second membrane potentials that are different respective forms of membrane potentials (spatiotemporal brain [natural neural network comprising biological neurons] data can be integrated in a machine learning model – Kasabov, sec. 1.1, last paragraph; EEG channels are spatially distributed on the scalp; the spatially defined positions of the channels can be mapped into a spiking neural network reservoir of a NeuCube model [electronic neural network comprising neuron modules] using different approaches – id. at sec. 4.1, first paragraph; EEGs allow researchers to track electrical potentials [membrane potentials] across the surface of the brain – id. at sec. 1.2, first paragraph; NeuCube model can learn spatio-temporal patterns in the form of short-term memory, represented as changes of the post-synaptic potentials, and long-term memory, represented as long-term potentiation (LTP) and long-term depression (LTD) – id. at sec. 5.3, first two paragraphs; see also sec. 5.4 and Table 1 (disclosing that the post-synaptic potentials modeled include fast excitation PSPs, slow excitation PSPs, fast inhibition PSPs, and slow inhibition PSPs [different forms of membrane potentials])); and
configuring either one or both of circuit layers of the electronic neural network device configured to perform voltage amplification corresponding to either one or both of accumulation and firing of one or more of the membrane potentials of the biological neurons and [memory layers] of the electronic neural network device … corresponding to synaptic weights in a neural network map using the neural network map representing a connection structure and the synaptic weights of the natural neural network in the electronic neural network device constructed, based on the considering, to cause the electronic neural network device to mimic the natural neural network (EEG channels are spatially distributed on the scalp [i.e., the map is constructed based on the consideration of multiple types of potentials in the EEG signal]; the spatially defined positions of the channels can be mapped into a spiking neural network reservoir using different approaches – Kasabov, sec. 4.1; spatiotemporal information processing can be observed in the brain and integrated into [i.e., mimicked by] a machine learning model – id. at sec. 1.1, last paragraph; after the SNNr is trained on the STBD in an unsupervised model, the data are propagated again through the SNNr and the state of the SNNr is measured for each pattern and an output classifier is trained to recognize this state in a predefined output class for this input pattern – id. at sec. 3.4, first paragraph; STBD is entered into relevant areas of the SNNr over time; the initially set connection weights are modified through unsupervised learning [i.e., the connection structure and synaptic weights are learned]; a NeuCube model can learn spatio-temporal patterns in its three forms of memory [i.e., the weights are stored in memory and the mapping therefore involves configuring memory layers of the electronic device] – id. at sec. 5.3, first two paragraphs).”
Kasabov appears not to disclose explicitly the further limitations of the claim. However, Yanagisawa discloses “measuring, for a first timing interval, first membrane potentials and second membrane potentials using different electrodes (EEG electrodes are placed on a subject’s head; the electrodes sense voltages [potentials] at the surface of the subject’s skin, and the EEG receiver outputs respective time series information for each of the electrodes over the duration of the measurement [i.e., there is a measurement from one electrode at one time step and a measurement from another electrode at another time step] – Yanagisawa, paragraphs 18-19) ….” 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 Kasabov to measure potentials for different time intervals using different electrodes, as disclosed by Yanagisawa, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would give a more complete picture of the subject’s brain activity than would be the case if the same electrodes were used. See Yanagisawa, paragraphs 18-19.
Neither Kasabov nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays of the electronic neural network device comprising crosspoints comprising memory elements having resistances corresponding to synaptic weights (synaptic crossbar memory array may have precisely programmed resistance values that may represent synaptic weights – Afzali, paragraph 6; see also paragraph 16 (disclosing that the crossbar comprises crosspoints)) ….” 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 Kasabov/Yanagisawa to perform learning with a crossbar array, as disclosed by Afzali, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system to provide error-tolerant, massive parallelism, thereby easing the input of complex information. See Afzali, paragraph 2.
Regarding claim 18, Kasabov/Yanagisawa/Afzali discloses that “the considering includes considering interactions between respective information of measured action potentials (APs) and respective information of measured post-synaptic potentials (PSPs), for respective pre- and/or post-synaptic relationships among pre-synaptic biological neurons and post-synaptic biological neurons of the natural neural network (prior bioinformatics and neuroinformatics knowledge can be used when designing neurogenetic NeuCube models; for example, post-synaptic AMPA-type glutamate receptors mediate fast excitatory synaptic transmissions and are crucial for many aspects of brain functioning [i.e., information of post-synaptic potentials of the brain are used in mapping the brain to the ANN] – Kasabov, second full paragraph on right-hand column of p. 72; probabilistic parameters of the model have their biological analogues; for example, the probability of a synapse to contribute to the post-synaptic potential after it has received a spike from a pre-synaptic neuron may be affected by different proteins [i.e., the system models pre-/post-synaptic relationships within the brain, or the mapping is based on these relationships] – id. at p. 73, second full paragraph; see also Table 1 (characterizing the PSPs as types of action potential)).”
Regarding claim 19, Kasabov/Yanagisawa/Afzali discloses that “the constructing comprises:
identifying the connection structure among the plurality of neuron modules (EEG channels are spatially distributed on the scalp; 64 EEG channels are positioned on a human head; the spatially defined positions of the channels [connection structure] can be mapped onto a SNNr [neuron modules] – Kasabov, sec. 4.1, first paragraph); and
updating the synaptic weights for connectivities between different neuron modules of the plurality of neuron modules (spiking sequences that represent EEG channels are entered into the correspondingly located neurons of the SNNr [divisible into portions comprising neuron modules] before a training [weight update] procedure (e.g., STDP) is applied – Kasahov, sec. 4.1, second paragraph).”
Regarding claim 20, Kasabov/Yanagisawa/Afzali discloses “[a] non-transitory computer-readable storage medium storing instructions that, when executed by a processor, cause the processor to implement the method of claim 17 (computational architecture for the creation of efficient computational models from spatiotemporal brain data is offered [i.e., the method is executed using a processor] – Kasabov, sec. 1.3, last paragraph; NeuCube model can learn spatiotemporal patterns in three forms of memory [storage medium storing instructions] – id. at sec. 5.3, second paragraph).”
Claims 21-24 and 26-27 are rejected under 35 U.S.C. 103 as being unpatentable over Kasabov in view of Yanagisawa and Afzali and further in view of Yamazaki et al. (US 20220375529) (“Yamazaki”) and Chang et al. (US 20200026981) (“Chang”).
Regarding claim 21, Kasabov discloses “[a]n electronic neural network device, comprising:
[electrodes for] measuring membrane potentials of a plurality of biological neurons of a natural neural network by measuring, for a first timing interval, first membrane potentials and second membrane potentials using … electrodes and measuring, for a second timing interval, the first membrane potentials and the second membrane potentials using same electrodes (electroencephalography is the recording of electrical signals from the brain [over time, implying the existence of two timing intervals] by attaching surface electrodes to the subject’s scalp; these electrodes [i.e., the same set of electrodes] record brain waves; EEGs allow researchers to track electrical potentials [plural, i.e., first and second potentials] across the surface of the brain – Kasabov, sec. 1.2, first paragraph);
one or more [memory layers] configured to, using … memory elements having … synaptic weights in a neural network map, store the neural network map, of the natural neural network, for a plurality of neuron modules of the electronic neural network device (EEG channels are spatially distributed on the scalp [i.e., are connected to a brain/natural neural network]; spatially defined positions of the channels can be mapped into an SNNr [electronic neural network device containing neuron modules] – Kasabov, sec. 4.1, first paragraph; NeuCube model can learn spatio-temporal patterns in its three forms of memory [i.e., the electronic neural network comprises memory layers to store the map] – id. at sec. 5.3, second paragraph; STBD is entered into relevant areas of the SNNr over time; unsupervised learning is performed to modify the initially set connection weights [i.e., the system has learned synaptic connections/weights that characterize the device] – id. at sec. 5.3, first bullet point); [and]
one or more circuit layers configured to activate each of multiple neuron modules, of the plurality of neuron modules, in response to stimuli or an input signal to the electronic neural network device, perform signal transmissions among the multiple neuron modules (different learning rules for spiking neural networks [running on circuit layers] have been introduced; STDP learning rule utilizes Hebbian plasticity in the form of long-term potentiation and depression; efficacy of synapses is strengthened or weakened [activated] based on the timing of a post-synaptic action potential in relation to the pre-synaptic spike [post-synaptic and pre-synaptic neurons = neuron modules; spike = signal transmissions among the modules in response to input signal] – Kasabov, sec. 2.2, second paragraph; spiking sequences that represent EEG channels are entered into correspondingly located neurons before STDP is applied – id. at sec. 4.1, second paragraph) …; ...
wherein either one or both of the one or more [memory layers] and the one or more circuit layers are configured based on the stored neural network map (STBD is entered into relevant areas of the SNNr over time; the initially set connection weights are modified through unsupervised learning; a NeuCube model can learn spatio-temporal patterns in its three forms of memory [i.e., the weights are stored in memory and the mapping therefore involves configuring memory layers of the electronic device] – Kasabov, sec. 5.3, first two paragraphs).”
Kasabov appears not to disclose explicitly the further limitations of the claim. However, Yamazaki discloses “connectors configured to connect the one or more [memory layers] and the one or more circuit layers (memory device includes a layer 10, layers 20_1 to 20-l, and wirings EW [connectors] – Yamazaki, paragraph 92; circuit OSC [circuit layer] is provided in the layer 10 and a memory cell portion [memory layer] is provided in the layers 20 – id. at paragraph 97; see also Fig. 1A (showing wiring EW connecting the layers)).”
Yamazaki and the instant application both relate to memory devices and are analogous. 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 Kasabov to include connectors between the memory layers and the circuit layers, as disclosed by Yamazaki, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would shorten the data transfer distance, and therefore the data transfer time, between the memory and the processor, rendering the system more efficient. See Yamazaki, paragraph 120.
Kasabov/Yamazaki appears not to disclose explicitly the further limitations of the claim. However, Yanagisawa discloses “measuring, for a first timing interval, first membrane potentials and second membrane potentials using different electrodes (EEG electrodes are placed on a subject’s head; the electrodes sense voltages [potentials] at the surface of the subject’s skin, and the EEG receiver outputs respective time series information for each of the electrodes over the duration of the measurement [i.e., there is a measurement from one electrode at one time step and a measurement from another electrode at another time step] – Yanagisawa, paragraphs 18-19) ….” 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 Kasabov/Yamazaki to measure potentials for different time intervals using different electrodes, as disclosed by Yanagisawa, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would give a more complete picture of the subject’s brain activity than would be the case if the same electrodes were used. See Yanagisawa, paragraphs 18-19.
Neither Kasabov, Yamazaki, nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays … using crosspoints comprising memory elements having resistances corresponding to synaptic weights (synaptic crossbar memory array may have precisely programmed resistance values that may represent synaptic weights – Afzali, paragraph 6; see also paragraph 16 (disclosing that the crossbar comprises crosspoints)) ….” 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 Kasabov/Yamazaki/Yanagisawa to perform learning with a crossbar array, as disclosed by Afzali, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system to provide error-tolerant, massive parallelism, thereby easing the input of complex information. See Afzali, paragraph 2.
Neither Kasabov, Yamazaki, Yanagisawa, nor Afzali appears to disclose explicitly the further limitations of the claim. However, Chang discloses “perform[ing] voltage amplification corresponding to either one or both of accumulation and firing of one or more membrane potentials of the neuron modules (dynamic threshold is updated each time after neuron B satisfies an equation including a term representing the membrane potential of the B neuron after accumulating [amplifying] all of the input voltages at the current time step – Chang, paragraph 63) ….”
Chang and the instant application both relate to neuromorphic devices and are analogous. 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 Kasabov/Yamazaki/Yanagisawa/Afzali to perform voltage amplification corresponding to accumulation of membrane potentials of neurons, as disclosed by Chang, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system properly to determine when to fire by comparing the accumulated voltage with a threshold. See Chang, paragraph 63.
Regarding claim 22, Kasabov, as modified by Yamazaki/Yanagisawa/Afzali/Chang, discloses that “a neural network result of the stored neural network map of the natural neural network is generated dependent on the performing of the signal transmissions (NeuCube architecture consists of functional modules that include an output function (classification) module – Kasabov, sec. 3.1, first paragraph; after the SNNr [containing the map] is trained on STBD, the same input data are propagated again through the SNNr [i.e., signals are transmitted], the state of the SNNr is measured for each pattern, and an output classifier is trained to recognize this state in a predefined output class [neural network result] for this input pattern – id. at sec. 3.4, first paragraph).”
Regarding claim 23, the rejection of claim 21 is incorporated. Kasabov further discloses that “when the electronic neural network device is a learned electronic neural network device, information in the one or more [memory layers] and information in the one or more circuit layers have characteristics of the electronic neural network device (NeuCube model can learn spatio-temporal patterns in its three forms of “memory,” similar to the brain memory [i.e., information stored in the memory layers is characteristic of the NeuCube model/electronic neural network device] – Kasabov, sec. 5.3, second paragraph; spatially defined positions of the EEG channels can be mapped onto an SNNr [circuit layers of electronic neural network device processing characteristic information] – id. at sec. 4.1, first paragraph) having been mapped from the natural neural network based on respective information of measured action potentials (APs) interacting with respective information of measured post-synaptic potentials (PSPs) for respective pre- and/or post-synaptic relationships among pre-synaptic biological neurons and post-synaptic biological neurons of the natural neural network (prior bioinformatics and neuroinformatics knowledge can be used when designing neurogenetic NeuCube models; for example, post-synaptic AMPA-type glutamate receptors mediate fast excitatory synaptic transmissions and are crucial for many aspects of brain functioning [i.e., information of post-synaptic potentials of the brain are used in mapping the brain to the ANN] – Kasabov, second full paragraph on right-hand column of p. 72; probabilistic parameters of the model have their biological analogues; for example, the probability of a synapse to contribute to the post-synaptic potential after it has received a spike from a pre-synaptic neuron may be affected by different proteins [i.e., the system models pre-/post-synaptic relationships within the brain, or the mapping is based on these relationships] – id. at p. 73, second full paragraph; see also Table 1 (characterizing the PSPs as types of action potential)).”
Neither Kasabov, Yamazaki, Chang, nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays”, as shown in the rejection of claim 21. 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 the combination of Kasabov, Yamazaki, Chang, and Yanagisawa to employ crossbar arrays, as disclosed by Afzali, for substantially the same reasons as given in the rejection of claim 21.
Regarding claim 24, Kasabov, as modified by Yamazaki/Yanagisawa/Chang, discloses that “the connectors comprise at least one of:
through-silicon vias (TSVs) penetrating through respective [memory layers] of the one or more [memory layers] and respective circuit layers of the one or more circuit layers (memory device includes a layer 10, layers 20_1 to 20-l, and wirings EW [connectors] – Yamazaki, paragraph 92; circuit OSC [circuit layer] is provided in the layer 10 and a memory cell portion [memory layer] is provided in the layers 20 – id. at paragraph 97; through electrode may be used to connect the circuit to the package substrate, and a TSV can be used as the through electrode – id. at paragraph 559; see also Fig. 1A (showing wiring EW connecting the layers)); and
micro bumps connecting the respective crossbar arrays and the respective circuit layers.” 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 Kasabov/Yanagisawa/Chang to employ TSVs as the connectors, as disclosed by Yamazaki, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would shorten the data transfer distance, and therefore the data transfer time, between the memory and the processor, rendering the system more efficient. See Yamazaki, paragraph 120.
Neither Kasabov, Yamazaki, Chang, nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays”, as shown in the rejection of claim 21. 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 the combination of Kasabov, Yamazaki, Chang, and Yanagisawa to employ crossbar arrays, as disclosed by Afzali, for substantially the same reasons as given in the rejection of claim 21.
Regarding claim 26, neither Kasabov, Yanagisawa, Chang, nor Yamazaki appears to disclose explicitly the further limitations of the claim. However, Afzali discloses that “respective synaptic weights in the neural network map are stored in respective crosspoints of the one or more crossbar arrays (synaptic crossbar or crosspoint memory array [memory layer] includes multiple lower electrode lines in one direction, multiple upper electrode lines in an orthogonal direction, and an electronic synapse at each intersection [crosspoint] of the electrode lines – Afzali, paragraph 16; each synapse includes a resistor with one-time-alterable resistance in series with a diode – id. at paragraph 17; resistance values may represent synaptic weights – id. at paragraph 6).” 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 the combination of Kasabov, Yanagisawa, Chang, and Yamazaki to employ a crossbar array with synaptic weights stored in the crosspoints, as disclosed by Afzali, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow the system to provide error-tolerant, massive parallelism, thereby easing the input of complex information. See Afzali, paragraph 2.
Regarding claim 27, Kasabov, as modified by Yamazaki/Yanagisawa/Chang, discloses that “the one or more [memory layers] and the one or more circuit layers are three-dimensionally stacked (data processing device including a first layer and a second layer is provided; an arithmetic processing device [circuit] is provided in the first layer, and a memory cell portion [memory] is provided in the second layer; at least part of the second layer is stacked over the first layer [i.e., three-dimensionally] – Yamazaki, paragraph 15).” 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 Kasabov/Yanagisawa to stack the memory and circuit layers three-dimensionally, as disclosed by Yamazaki, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would shorten the data transfer distance, and therefore the data transfer time, between the memory and the processor, rendering the system more efficient. See Yamazaki, paragraph 120.
Neither Kasabov, Yamazaki, Chang, nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays”, as shown in the rejection of claim 21. 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 the combination of Kasabov, Yamazaki, Chang, and Yanagisawa to employ crossbar arrays, as disclosed by Afzali, for substantially the same reasons as given in the rejection of claim 21.
Claim 25 is rejected under 35 U.S.C. 103 as being unpatentable over Kasabov in view of Yamazaki, Yanagisawa, Chang, and Afshali and further in view of Tschirhart (US 20200242452) (“Tschirhart”).
Regarding claim 25, the rejection of claim 21 is incorporated. Kasabov further discloses that “a neural network result of the stored neural network map of the natural neural network is generated dependent on the performing of the signal transmissions (NeuCube architecture consists of functional modules that include an output function (classification) module – Kasabov, sec. 3.1, first paragraph; after the SNNr [containing the map] is trained on STBD, the same input data are propagated again through the SNNr [i.e., signals are transmitted], the state of the SNNr is measured for each pattern, and an output classifier is trained to recognize this state in a predefined output class [neural network result] for this input pattern – id. at sec. 3.4, first paragraph) ….”
Neither Kasabov, Yanagisawa, Chang, nor Yamazaka appears to disclose explicitly the further limitations of the claim. However, Tschirhart discloses that “the one or more circuit layers are further configured to activate corresponding neuron modules, for the generating of the neural network result, by reading synaptic weights corresponding to connectivities among the corresponding neuron modules from the one or more [memory layers] in response to the stimuli or input signal (input pulse representing an action potential [stimuli] is received by a simulated neuron; a synapse weight value is accessed [read] from a superconducting digital array memory [memory layers] based on the input signal; synapse weight values accessed during a time period are accumulated; output signal is emitted as an output pulse [i.e., the neuron is activated by circuit layers to generate a result] based on a comparison of the accumulated signal to a threshold – Tschirhart, paragraph 8; see also paragraph 28 (disclosing that the weights represent synaptic connections between neurons)).”
Tschirhart and the instant application both relate to neuromorphic devices and are analogous. 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 the combination of Kasabov, Yanagisawa, and Yamazaki to activate neurons from a circuit layer to generate a result by reading weights from a memory layer, as disclosed by Tschirhart, and an ordinary artisan could reasonably expect to have done so successfully. Doing so would allow for the engineering of new computational platforms by mimicking neuro-biological architectures. See Tschirhart, paragraph 3.
Neither Kasabov, Yamazaki, Tschirhart, Chang, nor Yanagisawa appears to disclose explicitly the further limitations of the claim. However, Afzali discloses “crossbar arrays”, as shown in the rejection of claim 21. 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 the combination of Kasabov, Yamazaki, Tschirhart, Chang, and Yanagisawa to employ crossbar arrays, as disclosed by Afzali, for substantially the same reasons as given in the rejection of claim 21.
Allowable Subject Matter
Claim 15 would be allowable if rewritten to overcome the rejection(s) under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), 2nd paragraph, set forth in this Office action and to include all of the limitations of the base claim and any intervening claims.
Response to Arguments
Applicant’s arguments with respect to the claims have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument. In particular, Applicant’s argument that the previously cited references do not teach the amended portion of the independent claims is moot by the use of Anagisawa, Chang, and Afzali to teach the disputed limitations.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RYAN C VAUGHN whose telephone number is (571)272-4849. The examiner can normally be reached M-R 7:00a-5:00p ET.
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 at 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 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.
/RYAN C VAUGHN/ Primary Examiner, Art Unit 2125
1 To Examiner’s knowledge, there is nothing directly analogous to the artificial neural network concept of “synaptic weights” in biological brains. Therefore, any map that is used to produce a weighted artificial neural network will be construed to read on the claim.
2 This remains true notwithstanding the difference in dependency, as claim 31 is directly dependent on independent claim 28 whereas claim 5 is dependent on equivalent independent claim 1 through claim 3. Since claim 3 was rejected under § 102 over Kasabov, the ground of rejection is substantially the same.
Prosecution Insights