Notice of Pre-AIA or AIA Status
Claims 21-39 are pending in this application. Claims 1-20 are canceled and claims 21-39 are new by applicant’s amendment filed 21 April 2026.
Drawings
The drawings are objected to because new figs. 7 and 8, filed 21 April 2026, add new matter that was not included in the original disclosure. These drawing should be deleted. 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 amendment filed 21 April 2026 is objected to under 35 U.S.C. 132(a) because it introduces new matter into the disclosure. 35 U.S.C. 132(a) states that no amendment shall introduce new matter into the disclosure of the invention. Points 47-50 of the remarks, which include additional paragraphs added to pages 44 and 45 of the specification.
Applicant is required to cancel the new matter in the reply to this Office Action.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 21-39 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claims do not fall within at least one of the four categories of patent eligible subject matter because they recite a neural network comprising neurons and connections among the neurons. The applicant’s specification does not describe how the neurons and connections are implemented, and it is known for neurons of a neural network to be implemented in software. Software is none of a process, machine, manufacture, or composition of matter.
Adding subject matter to the specification and drawings does not overcome the rejection under 35 U.S.C. 101. The examiner suggests amending the claims to limit the claimed invention to hardware embodiments of a neural network, i.e. with neurons that are implemented as physical devices.
Claim Objections
Claim 25 is objected to because of the following informalities: the present claim recites “the triplelet context”; the examiner assumes that this should recite “the triplet context.” In addition, the claim recites “a censor.” Claim 21 recites “one or more sensory areas” and “a sensory area type,” so it appears that the present claim should recite “a sensor” rather than “a censor.” Appropriate correction is required. The present claim
Claim Rejections - 35 USC § 103
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.
Claims 21-24, 28, and 30-39 are rejected under 35 U.S.C. 103 as being unpatentable over Weng, Juyang (“Brains as naturally emerging turing machines,” 2015 International Joint Conference on Neural Networks (IJCNN). IEEE, 2015; hereinafter “Weng”) in view of Islam, Md Monirul, et al. (“A new adaptive merging and growing algorithm for designing artificial neural networks,” IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics) 39.3 (2009): 705-722; hereinafter “Islam”).
Regarding Claim 21, Weng teaches a computer-implemented neural network which has three types of areas, sensory, motor, and hidden (section II and fig. 1b—the developmental Network {DN} has sensory areas X, hidden areas Y, and motor areas Z. The areas are further described in section III), wherein
an area of the sensory type connects to and from a sensor (section II and figs. 1 and 2—sensory area X connects to and from sensors that produce a sensory image),
an area of the motor type connects to and from a motor (also called effector or actuator) (section II and figs. 1 and 2—motor area Z connects to and from a motor/output),and
a hidden area connects to and from sensory areas, to and from motor areas, and to and from hidden areas (section II and figs. 1 and 2—hidden area Y connects to and from sensory areas X, and to and from motor areas Z), and
Weng further teaches at least one of the three areas grows while the network updates across time (section III—every area performs mitosis-equivalent{i.e. cell division that grows the network if it is needed during each iteration), but does not explicitly teach at least one of the three areas grows from a single neuron to a plurality of neurons while the network updates across time. However, Islam teaches at least one of three areas grows from a single neuron to a plurality of neurons while the network updates across time (section II. A. and fig. 1—a hidden layer initially contains only one neuron, and all layers {areas} grow while the network updates across time).
All of the claimed elements were known in Weng and Islam and could have been combined by known methods with no change in their respective functions. It therefore would have been obvious to a person of ordinary skill in the art at the time of filing of the applicant’s invention to combine the layer growing from a single neuron to a plurality of neurons of Islam with the areas growing across time of Weng to yield the predictable result of at least one of the three areas grows from a single neuron to a plurality of neurons while the network updates across time. One would be motivated to make this combination for the purpose of designing compact ANN architectures with good generalization ability compared to other algorithms (Islam, Abstract).
Regarding Claim 22, Weng/Islam teaches the neural network of Claim 21 which recursively maps from a triplelet context of sensory, hidden, and motor at each previous time instant to a triplelet context of sensor, hidden and motor to each current time instant and, therefore, is capable of learning any finite, emergent (vector-input), and incrementally taught Turing machine error-free (Weng, section V—the system can learn a Turing machine TM, and asserts that the controller of any TM is an FA, and a grounded DN can learn the FA perfectly. The algorithm described in section III operates on triplets of sensor, hidden, and motor).
Regarding Claim 23, Weng/Islam teaches the neural network of Claim 21 which has a number of pixels (receptors) wherein the pixels are not spatially uniform to emulate a biological retina or another biological sensor, or a number of motor neurons (muscles) wherein the motor neurons are not spatially uniform to emulate biological muscles or another biological effector (Weng, section II—the DN uses natural sensory input without special encoding, implying non-uniform pixels to emulate a biological retina. It also emulates natural effectors, thus implying motor neurons that are not spatially uniform.).
Regarding Claim 24, Weng/Islam teaches the neural network of Claim 21 which has at least one of following properties represented by two acronyms SACUT GENISAMA: Single brain cells to start, All lives reported, Contexts as motor-hidden-receptor triplelets, Unsupervised, Turing machines, Grounded, Emergent brain areas, Natural, Incremental, Skull-closed, Attentive, Motivated, Abstractive (Weng, section V—Turing machines, emergent brain areas, and context state. Section II describes natural sensory input and muscle actuation).
Regarding Claim 8, Weng/Islam teaches the neural network of Claim 21 which is free from world-symbols in some areas in the network, including motor areas, to enable motors to learn any world-consciousness at different levels of consciousness in a natural language that is spoken, written, or signed (Weng, section IV—the language acceptor takes regular language as input and produces outputs that are symbolic and thus free from world-symbols; the motor areas are therefore free from world-symbols).
Regarding Claim 30, Weng/Islam teaches wherein the network is free from motor-imposed training after birth so that postnatal learning is unsupervised or reinforcement or both (Weng, section VI—WWNs are versions of DNs in which learning does not need to be supervised, and may use reinforcement learning; they are thus free of motor-imposed training after birth).
Regarding Claim 31, Weng/Islam teaches the neural network of Claim 21 wherein an emergence of firing patterns in the motor area (or premotor area—inside the hidden area and near the motor area) represents a larger and higher context as consciousness (Weng, section VI—firing in the Z area {motor area} represents a context that meets the applicant’s definition of consciousness).
Regarding Claim 32, Weng/Islam teaches the neural network of Claim 21 wherein an on-the-fly consciousness learning process facilitates an acquisition of intelligence (Weng, section VI—the autonomous attention of the WWN is an on-the-fly consciousness learning process, and is a form of an acquisition of intelligence).
Regarding Claim 33, Weng/Islam teaches the neural network of Claim 21 wherein an autonomous imitation process is a general-purpose mechanism for both learning consciousness and acquiring intelligence (Weng, section VII—a DN interactively and incrementally learns a naturally emerging TM by imitating a teacher TM).
Regarding Claim 34, Weng/Islam teaches the neural network of Claim 21 wherein motor neurons automatically direct an attention on sensors (Weng, section VI).
Regarding Claim 35, Weng/Islam teaches the neural network of Claim 21 wherein an on-the-fly conscious learning process avoids static data sets (Weng, section II—a DN learns incrementally rather than relying on a static data set).
Regarding Claim 36, Weng/Islam teaches the neural network of Claim 21 wherein sensors are calibrated by the network autonomously through trial and error (Weng, section II—sensor input is naturally emergent; thus, it is calibrated by the network autonomously through trial and error).
Regarding Claim 37, Weng/Islam teaches the neural network of Claim 21 wherein a redundant or nonredundant limb (or effector) is calibrated by the network autonomously through trial and error (Weng, section II—output drives natural effectors using motor input that is emergent; the effector is therefore calibrated by the network autonomously through trial and error).
Regarding Claim 38, Weng/Islam teaches the neural network of Claim 21 wherein the network does not have a "central government" like controller such as convolution (Weng, section III—the algorithm describes a distributed update process, without any central controller or convolution).
Regarding Claim 39, Weng/Islam teaches the neural network of Claim 21 which solves at least one of following 20 "million-dollar" problems: an image annotation problem, a sensorimotor recurrence problem, a motor-supervision problem, a sensor calibration problem, an inverse kinematics problem, a government-free problem, a closed-skull problem, a nonlinear controller problem, a curse of dimensionality problem, an under- sample problem, a distributed vs. local representations problem, a symbol problem or a frame problem, a local minima problem, an abstraction problem, a compositionality problem, a smooth representations problem, a motivation problem, an optimality problem, an auto-programming for general purposes (APFGP) problem, a brain-thinking problem (Weng, sections IV and V describe solving a symbol problem; section IV describes attention that solves a brain thinking problem; and section III describes solving a closed-skull problem. Islam, section III describes solving an optimization problem).
Claim 25 is rejected under 35 U.S.C. 103 as being unpatentable over Weng in view of Islam, as applied to claim 1, above, and further in view of Zhang et al. (U.S. 2021/0197855, hereinafter “Zhang”).
Regarding Claim 25, Weng/Islam does not specifically teach the neural network of Claim 21 wherein the hidden area generates a hierarchy of features from the triplelet context of sensor, hidden, and motor, wherein features are concrete or local near a censor and are abstract or global near a motor. However, Zhang teaches a neural network where a hidden area generates a hierarchy of features wherein features are concrete or local near an input and are abstract or global near an output (fig. 7; ¶ [0220]—a task abstraction network includes a hidden layer that extracts features from an input, rendering them more abstract on the output, thus rendering the features concrete or local near a sensor {input} and are abstract or global near a motor {output}).
All of the claimed elements were known in Weng/Islam and Zhang and could have been combined by known methods with no change in their respective functions. It therefore would have been obvious to a person of ordinary skill in the art at the time of filing of the applicant’s invention to combine the hidden layer extracting abstract features of Zhang with the hidden area triplet context of sensor, hidden, and motor of Weng/Islam to yield the predictable result of the neural network of Claim 21 wherein the hidden area generates a hierarchy of features from the triplelet context of sensor, hidden, and motor, wherein features are concrete or local near a censor and are abstract or global near a motor. One would be motivated to make this combination for the purpose of determining a better global policy for processing tasks more quickly (Zhang, ¶ [0008]).
Claims 26 and 29 are rejected under 35 U.S.C. 103 as being unpatentable over Weng in view of Islam, as applied to claim 1, above, and further in view of Weng et al. (U.S. 2019/0392321, hereinafter “Weng2”).
Regarding Claim 26, Weng/Islam does not specifically teach the neural network of Claim 21 which contains multiple glial cells and wherein each neuron has a 3D location indicated by glial cells and the 3D locations of generated neurons are changed by the activities of generated neurons to emulate brain patterning. However, Weng2 teaches a neural network which contains multiple glial cells and each neuron has a 3D location indicated by glial cells and the 3D locations of generated neurons are changed by the activities of generated neurons to emulate brain patterning (¶ [0132] – [0144]).
All of the claimed elements were known in Weng/Islam and Weng2 and could have been combined by known methods with no change in their respective functions. It therefore would have been obvious to a person of ordinary skill in the art at the time of filing of the applicant’s invention to combine the glial cells and neuron locations of Weng2 with the neural network of Weng/Islam to yield the predictable result of the neural network of Claim 21 which contains multiple glial cells and wherein each neuron has a 3D location indicated by glial cells and the 3D locations of generated neurons are changed by the activities of generated neurons to emulate brain patterning. One would be motivated to make this combination for the purpose of improving the neural network’s ability to learn by automatically and incrementally generating an internal hierarchy (Weng2, Abstract).
Regarding Claim 29, Weng/Islam/Weng2 teaches the neural network of Claim 21 wherein usage-based neuronal mitosis enables automatic recruitment of neuronal resources based on a competition and wherein a nerve growth factor simulates growth scheduling over time (Weng2, ¶ [0086] and [0132]—mitosis {splitting} involves the best matched neurons {i.e. competition} using calculated growth factors to simulate growth scheduling over time).
Claim 27 is rejected under 35 U.S.C. 103 as being unpatentable over Weng in view of Islam, as applied to claim 1, above, and further in view of Grbic, Djordje, and Sebastian Risi (“Safer reinforcement learning through transferable instinct networks,” Artificial Life Conference Proceedings 33. Vol. 2021. No. 1. One Rogers Street, Cambridge, MA 02142-1209, USA journals-info@ mit. edu: MIT Press, 2021; hereinafter “Grbic”).
Regarding Claim 7, Weng/Islam does not specifically teach the neural network of Claim 21 goes through a prenatal development using motor-imposed training to model biological development of innate behaviors. However, Grbic teaches a neural network that goes through a prenatal development using motor-imposed training to model biological development of innate behaviors (pp. 3-4, Phase 2: Instinct pre-training section—pre-training is a form of prenatal development for the neural network, and trains an artificial organism to avoid collisions with hazards {i.e. motor-imposed training} to model instinctive {i.e. innate} behaviors).
All of the claimed elements were known in Weng/Islam and Grbic and could have been combined by known methods with no change in their respective functions. It therefore would have been obvious to a person of ordinary skill in the art at the time of filing of the applicant’s invention to combine the prenatal training for innate behaviors of Grbic with the neural network of Weng/Islam to yield the predictable result of the neural network of Claim 21 goes through a prenatal development using motor-imposed training to model biological development of innate behaviors. One would be motivated to make this combination for the purpose of ensuring that agents can operate in safety-critical environments (Grbic, Abstract).
Response to Arguments
The amendments to the claims are accepted as overcoming the previous rejections under 35 U.S.C. 112(b). Note, however, that claim 25 still recites a term that was objected to in the first office action, and adds a typographical error.
Applicant’s arguments filed 21 April 2026 have been fully considered but they are not persuasive.
Regarding the rejections under 35 U.S.C. 101, amending the specification and drawings to show a hardware embodiment does not overcome the rejections of the claims as reciting non-statutory subject matter. These amendments in fact necessitate additional rejections because, despite the applicant’s assertions, they add new matter that was not present in the original disclosure. This new matter needs to be removed, as detailed above. To overcome the 101 rejections, the claims need to be amended to limit the claimed invention to a hardware implementation that falls under one of the statutory categories of invention.
Regarding the arguments with respect to the rejections under 35 U.S.C. 103, the applicant argues that in claim 21, “there are five places that requires ‘to and from’.” The examiner notes that the broadest reasonable interpretation of “to and from” does not require bidirectional data transfer, nor does in require two separate connections between two entities. The claim is simply not that detailed. The examiner also notes that prior art reference Weng teaches this limitation, as shown in fig. 2 and described in the associated text. Whether or not Weng starts the hidden area from a single neuron is not relevant to this limitation.
The applicant repeatedly argues that Islam does not teach “at least one of the areas grows from a single neuron to a plurality of neurons while the network updates across time,” because Islam uses a feed-forward ANN. The applicant asserts “without the feedforward constraint, Islam had no idea what it will result in.” However, the present limitation does not place any constraints on the resulting neural network other than growing from a single neuron. It does not recite what connections are created to other neurons, how the neurons function, or how decisions are made about how many neurons to create or the process by which they are created. It simply states that at least one area starts with a single neuron and grows to a plurality (at least two) with updates over time. This is a very broad limitation, and Islam teaches to this breadth, as detailed above.
To overcome the rejection, the examiner suggests amending claim 21 to include additional details about the process by which the area grows, using details described in the specification. For example, what conditions or calculations are used to decide when to add neurons and how many to add? How are the new neurons connected to other neurons or areas? As for the “to and from” connections, how do these function within the neural network? What data are transferred in each direction, and what operations are performed in the neurons using these data?
Regarding Claim 25, the applicant again argues about the use of Islam, but the examiner has not relied on Islam to teach the limitations of claim 25. The examiner also disagrees with the applicant’s assertion that the features in Zhang are hand-crafted. Zhang performs feature extraction using a neural network. The feature vectors are therefore generated by the neural network, not crafted by humans. As with the previously discussed claims, the present claim is very broad, so it does not recite any limitations that are not taught by Zhang (in combination with Weng and Islam).
Regarding Claim 28 (former claim 8), the claim does not define or limit “any world-consciousness.” Weng teaches producing outputs that are symbolic and thus free from world-symbols, so it can be said to learn any world-consciousness.
Regarding Claim 30 (former claim 10), the applicant argues that Weng and Islam do not contain prenatal development. However, the claim does not recite anything about prenatal development; it only recites postnatal learning, which is taught by Weng. The applicant is arguing for features that are not claimed.
Regarding Claim 31 (former claim 11), the applicant argues that Weng does not meet the applicant’s definition of consciousness because it does not conduct brain patterning. However, the applicant’s specification states “For our discussion, we refer to consciousness as awareness for “larger and higher contexts” (¶ [0002] of U.S. 2023/0306244). There is no mention here of brain patterning, and the claim does not recite anything about brain patterning. The examiner also point out that the applicant has not demonstrated that his invention achieves consciousness in a machine. Any claims to consciousness are therefore suspect, and can only be interpreted as the machine exhibiting behavior that appears conscious to an observer.
Regarding Claims 32, 33, 35, 37, and 39 (former claims 12, 13, 15, 17, and 19), the applicant repeatedly argues that Weng and Islam do not teach the limitations because they do not teach brain patterning. None of these claims recite brain patterning, not to mention the details about how brain patterning is performed. The rejections are therefore maintained.
Regarding applicant’s points 13-19 in the remarks, the comments about conscious learning are irrelevant to the rejections. They too are arguing for features that are not recited by the claims. Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). The claims do not recite details of the process of brain patterning; claim 21 merely states that at least one neuron area grows from a single neuron, but includes hardly any details about how it grows, how connections are chosen/formed, and how it is patterned after a biological brain. As far as the arguments about machine consciousness, most AI researchers do not make any assertions about their systems achieving consciousness because scientists do not agree on what consciousness is and there is no known way to prove that a machine has achieved consciousness. Unless the applicant can provide such proof, any claims to consciousness can only be interpreted as the machine giving the appearance or illusion of consciousness. Otherwise, the claimed invention will need to be rejected under 35 U.S.C. 101 as being inoperative or implausible.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HAL W SCHNEE whose telephone number is (571) 270-1918. The examiner can normally be reached M-F 7:30 a.m. - 6:00 p.m.
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, Michael Huntley can be reached at 303-297-4307. 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.
/HAL SCHNEE/ Primary Examiner, Art Unit 2129
Prosecution Insights