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 .
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 01/20/2026 has been entered.
Response to Amendment
The office action is responsive to the amendment filed on 01/20/2026. As directed by the amendments claims 1-2, 8-11,17-19,26-29, and 35-36 are amended. Claims 1-36 are pending for examination.
Response to Arguments
Regarding the 35 U.S.C § 101 Rejection:
Applicant's further arguments see pg. 10-11 filed 01/20/2026, have been fully considered but they are not persuasive.
APPLICANT ARGUMENT:
Applicant argues, “As described in para. [0059], [0067] and [0073] of the specification, learning contact-based tasks include numerous challenges for machine learning. A neural network architecture that includes a contact model to model contact dynamics for the autonomous device with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force serves to avoid discontinuities and accelerate training. FIGs. 4 and 5 illustrate experimental results showing improvements over other approaches. Accordingly, Applicant's claim reflect an improvement to computer technology and are thus patent eligible. The recent precedential PTAB Decision Ex Parte Desjardins et al., Appeal 2024-000567 (ARP Sept. 26, 2025) (precedential) is one point. Moreover, the specific neural network architecture that includes a dynamics model to model articulated dynamics including both a contact model to model contact dynamics with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force, and a joint limit model employing a continuous penalty-based force, serves to avoid discontinuities and accelerate training which is significantly more than an abstract idea. Moreover, controlling an autonomous device is a practical application. Accordingly, Applicant's claims are patent eligible for numerous reasons. Withdrawal of this rejection is respectfully requested”.
EXAMINER RESPONSE: Examiner respectfully disagree, applicant argument is not persuasive. The claimed contact model as disclosed is a mathematical function as the claimed model is a friction model that calculates coulombs friction with linear step function using damping force (see specification [0073] ) and is a mathematical calculation for calculating friction which is an abstract idea. In addition, the claimed “joint limit model” as disclosed is also a mathematical function as the “joint limit model” employs a “continuous penalty-based force” which is considered a mathematical function. Accordantly, amended claim 1 does not recite "An improvement in the functioning of a computer, or an improvement to other technology or technical field" the claims does not specify the improvement nor is disclosed in the claim. Rather, the claimed improvement itself is an abstract idea and not a technical improvement as argued by applicant.
Furthermore, applicant also argues that the claimed contact model is used to control an autonomous device which is a practical application however the claimed invention as disclosed do not have any recitation of controlling rather approximating contacts using a neural network which as disclosed broadly is mental process in combination of mathematical function.
Lastly, the claims presented for examination are not analogues to the Ex Parte Desjardins, Appeal No. 2024-000567 (hereinafter Ex Parte Desjardins) as the claims do not recite "an improvement in the functioning of a computer, or an improvement to other technology or technical field", do not specify the improvement nor is disclosed in the claim. Rather the claim as presented for examination, do not specify the technical features or details for how the computer functionality is being improved, in the Ex Parte Desjardins, the claims were found to be patent eligible because when evaluating the claim as whole, the limitation of independent claim 1, constituted an improvement to how the machine learning model itself operated, and not, for example, the identified mathematical calculation. However, the claims of the instant applicant do not reflect such improvement to the machine learning model, instead independent claims 1, 10, 18 and 28 recites claim language in a generic manner which appear to be an improvements of an abstract idea. The following paragraph from the USPTO Memo regarding the Ex Parte Desjardins, recites:
“The Appeals Review Panel (ARP) overall credited benefits including reduced storage, reduced system complexity and streamlining, and preservation of performance attributes associated with earlier tasks during subsequent computational tasks as technological improvements that were disclosed in the patent application specification. Specifically, the ARP upheld the Step 2A Prong One finding that the claims recited an abstract idea (i.e., mathematical concept). In Step 2A Prong Two, the ARP then determined that the specification identified improvements as to how the machine learning model itself operates, including training a machine learning model to learn new tasks while protecting knowledge about previous tasks to overcome the problem of “catastrophic forgetting” encountered in continual learning systems. Importantly, the ARP evaluated the claims as a whole in discerning at least the limitation “adjust the first values of the plurality of parameters to optimize performance of the machine learning model on the second machine learning task while protecting performance of the machine learning model on the first machine learning task” reflected the improvement disclosed in the specification. Accordingly, the claims as a whole integrated what would otherwise be a judicial exception instead into a practical application at Step 2A Prong Two, and therefore the claims were deemed to be outside any specific, enumerated judicial exception (Step 2A: NO)”. (emphasis added, see MPEP § 2106.04(d), subsection III).
Per contra, independent claims 1, 10, 18 and 28 when view as whole do not appear to recites features which provide concrete benefits, for example, the claim does not provide details in how for example the machine learning models are being optimize, how deployment times are being decreased or how lower computational cost for training in a target domain are being achieved. Further, applicant is reminder that while “the claim itself does not need to explicitly recite the improvement described in the specification (e.g., “thereby increasing the bandwidth of the channel”)”, if the specification sets forth an improvement in technology or a technical field, the claims must “includes the components or steps of the invention that provide the improvement described in the specification” (see MPEP § 2106.04(d)(1)).
Therefore for the above reason, claims 1-36 are not directed to patent-eligible
subject matter under 35 U.S.C § 101.
Regarding the 35 U.S.C § 102(a)(1):
Applicant's further arguments see pg. 10 filed 01/20/2026 have been fully considered but they are not persuasive.
APPLICANT ARGUMENT:
Applicant argues, the cited reference “Anonymous has not been shown to be prior art to the present application. The document includes dates of 29 Sept 2021 and 15 Nov 2021. However, there is no indication the document was published or otherwise publicly available on these dates. The referenced conference is a 2022 conference. At best, it appears Anonymous would not have been published or otherwise publicly available until the ICLR conference in 2022 after the December 15, 2021 filing date of the present application. Accordingly, Anonymous does not appear to be prior art to the present application. Applicant note that Office has the burden to establish prior art. Accordingly, aprimafacie rejection has not been establish with respect to Anonymous. Withdrawal of this rejection is respectfully requested”.
EXAMINER RESPONSE: Examiner respectfully disagree, applicant argument is not persuasive. As shown in the non-patent literature (NPL) document ANONYMOUS, “Accelerated Policy Learning with Parallel Differentiable Simulation," cited in the information disclosure statement (IDS) dated 08/14/2025 the document was retrieved from OpenRewiew and was available to “Everyone” to review prior to the International Conference on Learning Representations (ICLR) 2022 Conference. The NPL document includes reviews dates of 29 September 2021, thus the NPL document clearly show the document was released to the public for review prior to the effective filling date of the instant application. As shown in the “How to Guides” of OpenRewiew below submission of documents set to “Everyone” means the paper are released to the public.
PNG
media_image1.png
672
758
media_image1.png
Greyscale
In addition, “The wayback Machine” digital archive of the World Wide Web which lets users view past versions of website shows the article was publish for the public to view on October 17, 2021 . See also highlighted from “The wayback Machine”:
PNG
media_image2.png
1075
1569
media_image2.png
Greyscale
PNG
media_image3.png
577
1887
media_image3.png
Greyscale
Further, current NPL document published on OpenReview the first review was publish as early as early as October 18 2021. See section below:
PNG
media_image4.png
195
1060
media_image4.png
Greyscale
Accordingly, claims 1-36 are not patent eligible under 35 U.S.C § 102(a)(1) in view of NPL ANONYMOUS, “Accelerated Policy Learning with Parallel Differentiable Simulation," .
Regarding the 35 U.S.C § 103:
Applicant’s arguments with respect to claims 1-36 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.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 03/03/2026 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1-36 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Independent Claim 1 recites:
A processor, comprising:
one or more circuits to use one or more neural networks to indicate one or more actions of an articulated autonomous device and to indicate how the one or more actions will affect one or more subsequent actions of the articulated autonomous device, wherein the one or more neural networks include a dynamics model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including:
a contact model to model contact dynamics with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force; and
a joint limit model employing a continuous penalty-based force (emphasis added).
However, the specification of the instant application contains no mentions of “the one or more neural networks” including “a dynamics model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits”, the “contact damping force” being incorporated into the “contact model” and a “joint limit model” being employing a continuous penalty-based force. Rather paragraph [0073] of the instant application teaches to model contact a frictional contact model is used, damping force is incorporated into a dynamics model and teaches in order to “model joints limits a continuous penalty-based force is applied instead of enforcing limits as hard constraints”. Further, paragraph [0149] teaches “the one or more neural networks may include an actor network and a critic network”. Thus, the specification does not teach or suggest the newly added limitation of claim 1.
Independent claims 10, 19, and 28 recites similar limitation to those of claim 1, and thus are rejected for reasons set forth in the rejection of claim 1.
Claims 2-9,11-18,20-26, and 29-36 are dependent on claims 1, 10, 19, and 28, and thus are rejected for reasons set forth in the rejection of claim claims 1, 10, 19, and 28.
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.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-36 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites the limitation “the model” in line 9. There is insufficient antecedent basis for this limitation in the claim. For examination purposes, examiner is interpreting the limitations as “the dynamics model”.
Independent claims 10, 19, and 28 recites similar limitation to those of claim 1, and thus are rejected for reasons set forth in the rejection of claim 1.
Claims 2-9,11-18,20-26, and 29-36 are dependent on claims 1, 10, 19, and 28, and thus are rejected for reasons set forth in the rejection of claim claims 1, 10, 19, and 28.
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 1-36 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
STEP 1
Claims 1-9 are directed to a processor, claims 10-18 are directed to a computer-implemented method, claims 19 – 27 are directed to a computer system, claims 28 – 36 are directed to a non-transitory computer-readable storage medium. Therefore, claims 1-36 are directed to either a process, machine, manufacture or composition of matter.
Regarding claim 1: 2A Prong 1:
...to indicate one or more actions of an articulated autonomous device...(mental process – of indicate one or more actions of an articulated autonomous device can be performed by the human mind with the help of pen and paper. For example, this step for indicating actions of an autonomous device is practically implementable in the human mind by analyzing a given state and deciding the best course of action for the device. (e.g., evaluation )).
...and to indicate how the one or more actions will affect one or more subsequent actions of the articulated autonomous device,... (mental process – of indicate how the one or more actions will affect one or more subsequent actions of the articulated autonomous device can be performed by the human mind with the help of pen and paper. For example, this step for indicating the effect of actions is practically implementable in the human mind with the aid of pen and paper, for example by writing down possible actions and rating them by usefulness to infer future behavior (e.g., judgment)).
a contact model to model contact dynamics with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force; and (mathematical concept – for calculating contact friction. Paragraph [0073] of the instant application states in order to model contact a frictional content model is used, which approximate Coulomb friction with linear step function and in addition, contact damping force formulation is incorporates into the dynamic model to provide better smoothness of non-interpretation contact dynamics (e.g., mathematical function)).
a joint limit model employing a continuous penalty-based force (mathematical concept – for calculating penalty-based force. Paragraph [0073] of the instant application states in order to model joint limits, a continuous penalty-based force is applied (e.g., mathematical function)).
2A Prong 2: This judicial exception is not integrated into a practical application.
Additional elements:
A processor, comprising: one or more circuits to use one or more neural networks...(This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
...wherein the one or more neural networks include a dynamics model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
The additional elements as disclosed above alone or in combination do not integrate the judicial exception into practical application as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above.
2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
A processor, comprising: one or more circuits to use one or more neural networks...(This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
...wherein the one or more neural networks include a dynamics model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
The additional elements as disclosed above in combination of the abstract idea are not sufficient to amount to significantly more than the judicial exception as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above.
Regarding claim 10: is rejected under the same rational of claim 1. Claim 10 only recites the additional elements of A computer-implemented method... which is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f).
Regarding claim 19: is rejected under the same rational of claim 1. Claim 19 only recites the additional elements of A computer system comprising one or more processors and non-transitory computer-readable memory to store executable instruction... which is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f).
Regarding claim 28: is rejected under the same rational of claim 1. Claim 28 only recites the additional elements of A non-transitory computer-readable memory storing executable instructions that, as a result of being executed by one or more processors of a computer system, cause the computer system to... which is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f).
Regarding claim 2, 11, 20, 29:
2A Prong 1:
(mental process – of that indicates an action to control the autonomous device can be performed by the human mind with the help of pen and paper. For example, this step for indicating actions of an autonomous device involves a judgment or decision that can practically be performed in the human mind by analyzing a given state and deciding the best course of action for the device. (e.g., judgement )).
(mathematical concept – of estimates a value of a current state to achieving a result. Paragraph [0087] and function 6 of the instant application teach the mathematical function used to estimate state values (e.g., mathematical calculation)).
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
a first neural network... (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
and a second neural network... (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
Regarding claim 3, 12, 21, 30: 2A Prong 1:
(mental process – of accomplish a task by dividing the task into a plurality of action sequences can be performed by the human mind with the help of pen and paper. For example, this step for dividing a task is practically implementable in the human mind by taking into account the goal and then listing steps to attain it. (e.g., evaluation )).
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
the one or more neural networks are trained to accomplish a task... (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
and training the one or more neural networks over each of the action sequences individually (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
Regarding claim 4, 13, 22, 31: 2A Prong 1:
wherein the one or more actions are indicated based on a change to policy loss caused by a change to one or more inputs of a simulation (mental process – the steps for indicating actions can be performed by the human mind with the help of pen and paper (e.g., evolution )).
2A Prong 2 and 2B: None.
Regarding claim 5, 14, 23, 32: 2A Prong 1: None.
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
one or more neural networks are trained by training independently for each of a plurality of time intervals (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
Regarding claim 6, 15, 24, 33: 2A Prong 1:
wherein training of the one or more neural networks is based at least in part on a gradient of policy loss produced by a differentiable simulation (mathematical concept – of calculating gradients. Paragraph [0070] of the instant application disclose the backward pass used to calculate the gradient of a loss function (e.g., mathematical function)).
2A Prong 2 and 2B: None.
Regarding claim 7, 16, 25, 34: 2A Prong 1: None.
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the one or more neural networks are trained (This is understood to be insignificant extra-solution activity to the judicial exception - see MPEP 2106.05(g)).
(This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the one or more neural networks are trained (This step is directed to training neural networks in parallel, which is understood to be insignificant extra-solution activity and is well understood, routine and conventional activity as evidenced by the survey [“Parallel approaches to machine learning – A comprehensive survey”, abstract, lines 1-3, published in: Journal of Parallel and Distributed computing, Volume 73, Issue 3, March 2013]: “Literature has always witnessed efforts that make use of parallel algorithms / parallel architecture to improve performance; machine learning space is no exception. In fact, a considerable effort has gone into this area in the past fifteen years.” Specifically, on parallel training of neural networks, it says: [section 2.2.1, paragraph 13, lines 1-3]: “Lastly, the work in [23], talks about a parallel implementation for a cluster system. This used a network parallel training approach to implement the same.” Network in this context is referring to neural networks, as the paragraph covers the contribution in the neural network domain: [paragraph 13, line 1]: “A significant contribution is observed in the area of Neural Networks too.”)
(This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
The additional elements as disclosed above in combination of the abstract idea are not sufficient to amount to significantly more than the judicial exception as they are mere insignificant extra solution activity in combination of generic computer functions being implemented with generic computer elements in a high level of generality to perform the disclosed abstract idea above.
Regarding claim 8, 17, 26, 35: 2A Prong 1:
wherein the one or more neural networks includes an actor (mental process – of estimating a value can be performed by the human mind with the help of pen and paper (e.g., evaluation )).
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the one or more neural networks includes an actor network that learns a policy (This is directed to using computers or other machinery merely as a tool to perform an existing process. See MPEP 2106.05(f)).
Regarding claim 9:
2A Prong 1:
The claim recites no abstract idea, however inherits an abstract idea from claim 1. Therefore, the claim is not patentable under the same rationale.
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the autonomous device is a robot, a vehicle, a muscle-tendon network, a drone, a simulated character (This is directed to restricting the abstract idea to a particular technological environment. See MPEP 2106.05(h)).
Regarding claim 18: 2A Prong 1:
The claim recites no abstract idea, however inherits an abstract idea from claim 10. Therefore, the claim is not patentable under the same rationale.
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the autonomous device is a robot, a vehicle, a muscle-tendon network, a drone, a simulated character (This is directed to restricting the abstract idea to a particular technological environment. See MPEP 2106.05(h)).
Regarding claim 27: 2A Prong 1:
The claim recites no abstract idea, however inherits an abstract idea from claim 19. Therefore, the claim is not patentable under the same rationale.
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the autonomous device is a robot, a vehicle, a muscle-tendon network, a drone, a simulated character (This is directed to restricting the abstract idea to a particular technological environment. See MPEP 2106.05(h)).
Regarding claim 36: 2A Prong 1:
The claim recites no abstract idea, however inherits an abstract idea from claim 28. Therefore, the claim is not patentable under the same rationale.
2A Prong 2 and 2B: The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the autonomous device is a robot, a vehicle, a muscle-tendon network, a drone, a simulated character (This is directed to restricting the abstract idea to a particular technological environment. See MPEP 2106.05(h)).
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.
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, 2, 5, 8, 9 - 11, 14, 17 - 20, 23, 26, 27 - 29, 32, 35 are rejected under 35 U.S.C. 103 as being unpatentable over Clavera et al. MODEL-AUGMENTED ACTOR-CRITIC:BACKPROPAGATING THROUGH PATHS (hereinafter Clavera), in view of Geilinger et al. ADD: Analytically Differentiable Dynamics for Multi-Body Systems with Frictional Contact (hereinafter Geilinger) as cited in the information disclosure statement (IDS) dated 08/17/2022 in further view of Kurutach et al. Model-Ensemble Trust-Region Policy Optimization (hereinafter Kurutach) as cited in IDS dated 08/17/2022.
Regarding claim 1:
A processor, (Clavera [Page 6, Algorithm I] The presence of the algorithm implies the presence of hardware capable of its execution)
comprising: one or more circuits to use one or more neural networks... (According to the specifications of the instant application, in paragraph [0149], ”one or more neural networks may include an actor network and a critic network as describe above.” In Clavera, the actor-critic setup is based on neural networks: [page 2, section 2, paragraph differentiable planning, lines 8-9] “Our method, instead learns a neural network policy in an actor-critic fashion aided with a learned model.”)
to indicate one or more actions of an articulated autonomous device... (Clavera [page 3, section 3.1, paragraph “actor-critic”] A policy pi is learned that represents the actor and their subsequent actions. This policy is also present in the model-augmented setup: [page 3, section 4, line 2 and page 4, line 1] “Our approach, model-augmented actor-critic (MAAC), exploits the learned model by computing the analytic gradient of the returns with respect to the policy.” And [page 1, paragraph 4, lines 3]: “For instance, we achieve a 10k return in the half-cheetah environment in just 50 trajectories.” The half-cheetah environment features a 2-dimensional robot for simulation purposes. The broadest reasonable interpretation of ‘autonomous device’ includes a simulated robot/character).
and to indicate how the one or more actions will affect one or more subsequent actions of the articulated autonomous device, (Clavera [page 3, section 3.1, paragraph “Actor-Critic”, lines 1-2]: “In actor-critic methods, we learn a function Qˆ (critic) that approximates the expected return conditioned on a state s and action a. […] Then, the learned Q-function is used to optimize a policy π (actor).” The Q-function (critic), is optimized in the model-augmented actor-critic setup as well [page 6, section “Q-function learning”, lines 8 - 9]:” we maximally make use of the model by not only using it for the policy gradient step, but also for training the Q-function.” The expected return is the evaluation of the proposed action given a state, provided by the critic, shaping the policy pi of the actor, affecting the future actions of the actor. Further, [page 1, paragraph 4, lines 3]: “For instance, we achieve a 10k return in the half-cheetah environment in just 50 trajectories.” The half-cheetah environment features a 2-dimensional robot for simulation purposes. The broadest reasonable interpretation of ‘autonomous device’ includes a simulated robot/character).
Clavera does not teach wherein the one or more neural networks include a dynamics model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including: a contact model to model contact dynamics with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force; and a joint limit model employing a continuous penalty-based force.
Nonetheless, Geilinger teaches the following:
a model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including: (Geilinger Fig. 1 teaches modeling articulated dynamics for the multi-body systems (i.e., articulated autonomous device) without discontinuities with respect to contact and joint limits and pg. 4, right col., sec: 4 Differentiable Frictional Contact Model, para 1, teaches “formulate a smooth frictional contact model that approaches, in the limit, the discontinuous nature of physical contacts).
a contact model to model contact dynamics with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force; and (Geilinger pg. 2, left col., para. 2 and 3 teaches a contact model to model contact dynamics with different types of object in an environment. Further, pg. 5, sec: 4.2 Penalty methods, para. 3-4 and function 17, teaches the contact model by approximate contact friction using a linear step function. In particular, Geilinger teaches “we can formulate Coulomb friction as a clamped linear penalty force with corresponding penalty factor tangential”, thus this teach the contact model (i.e., function 17) calculates friction using a linear penalty force (i.e., linear step function). Moreover, Geilinger pg. 9, left col., para. 1-2 teaches explicitly including a “damping force” ).
a joint limit model employing a continuous penalty-based force (Geilinger pg. 5, sec: 4.2 Penalty methods, para. 1-2 and function 16, teaches a penalty model that introduces continuous (i.e., smooth) penalty forces through a piece wise linear penalty function
f
n
=
n
k
n
max
-
g
,
0
where
k
n
is a penalty factor that must be chosen large enough to sufficiently enforce the normal constraint”).
Geilinger is also in the same field of endeavor as Clavera (machine learning). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of a model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including a contact model and a joint limit model as being disclosed and taught by Geilinger, in the system taught by Calvera to yield the predictable results of enabling “differentiability and gradient-based optimization” (see Geilinger pg. 13, right col., para. 5).
Neither Clavera and Geilinger teaches wherein the one or more neural networks include a dynamics model.
However, Kurutach teaches the following:
wherein the one or more neural networks include a dynamics model... (Kurutach pg. 5, para. 2, teaches the one or more neural networks include a set of dynamics models
f
ϕ
1
…
f
ϕ
k
, termed a model ensemble).
Kurutach is also in the same field of endeavor as Clavera and Geilinger (machine learning). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of ensemble dynamics models as being disclosed and taught by Clavera and Geilinger, in the system taught by Clavera and Geilinger to yield the predictable results of improve the policy of the model (Kurutach see pg. 5 Algorithm 1).
Regarding Claim 2:
Clavera, Geilinger and Kurutach teach The processor of claim 1. Clavera specifically teaches wherein the one or more neural networks includes: a first neural network that indicates an action to control the articulated autonomous device; ([page 3, section 3.1, paragraph “actor-critic”] A policy pi is learned that represents the actor and their subsequent actions. And [page 9, section 6 Conclusion, lines 7-8]: “For future work, it would be enticing to deploy the presented algorithm on a real-robotic agent.” Examiner understands a real-robotic agent to be an autonomous device. Further, [page 1, paragraph 4, lines 3]: “For instance, we achieve a 10k return in the half-cheetah environment in just 50 trajectories.” The half-cheetah environment features a 2-dimensional robot for simulation purposes. The broadest reasonable interpretation of ‘autonomous device’ includes a simulated robot/character).
and a second neural network that estimates a value of a current state to achieving a result (In the broadest reasonable interpretation, the value of a state to achieving a result may be based on the output of a critic. [page 3, section 3.1, paragraph “Actor-Critic”, lines 1-2]: “In actor-critic methods, we learn a function Qˆ (critic) that approximates the expected return conditioned on a state s and action a. […] Then, the learned Q-function is used to optimize a policy π (actor).” The expected return is the evaluation of the proposed action given a state, provided by the critic, shaping the policy pi of the actor, affecting the future actions of the actor).
Regarding Claim 5:
Clavera, Geilinger and Kurutach teach The processor of claim 1. Clavera specifically teaches wherein the one or more neural networks are trained by training independently for each of a plurality of time intervals ([Page 3, section “Model-Based RL”, lines 4-5]: “The models are trained via maximum likelihood”. The formula in line 5 shows st, the state at time step t, among others, therefore distinct time intervals for training are given. The model can be neural networks: [Page 2, section “Differentiable Planning”, lines 8-9] “Our method, instead learns a neural network policy in an actor-critic fashion aided with a learned model.” One or more neural networks are trained over time intervals in Clavera. They are trained independently: [page 5, paragraph Model learning, lines 1-3] “In order to prevent overfitting and overcome model-bias (Deisenroth & Rasmussen, 2011), we use a bootstrap ensemble of dynamics models […]. Each of the dynamics models parameterizes the mean and the covariance of a Gaussian distribution with diagonal covariance”. A bootstrap ensemble, also known as bootstrap aggregating or “bagging”, in the art of machine learning, is understood to be independently trained models using random samples.)
Regarding claim 8:
Clavera, Geilinger and Kurutach teach The processor of claim 1. Clavera specifically teaches wherein the one or more neural networks includes an actor network that learns a policy (In Clavera, the actor-critic setup is based on neural networks: [page 2, section 2, paragraph differentiable planning, lines 8-9] “Our method, instead learns a neural network policy in an actor-critic fashion aided with a learned model.” This policy is also present in the model-augmented setup: [page 3, section 4, line 2 and page 4, line 1] “Our approach, model-augmented actor-critic (MAAC), exploits the learned model by computing the analytic gradient of the returns with respect to the policy.”)
that controls the autonomous device ([page 1, paragraph 4, lines 3]: “For instance, we achieve a 10k return in the half-cheetah environment in just 50 trajectories.” The half-cheetah environment features a 2-dimensional robot for simulation purposes. The broadest reasonable interpretation of ‘autonomous device’ includes a simulated robot/character)
and a critic network that learns to estimate a value of a state of the autonomous device with respect to a result ([page 3, section 3.1, paragraph “Actor-Critic”, lines 1-2]: “In actor-critic methods, we learn a function Qˆ (critic) that approximates the expected return conditioned on a state s and action a. […] Then, the learned Q-function is used to optimize a policy π (actor).” The Q-function (critic), is optimized in the model-augmented actor-critic setup as well [page 6, section “Q-function learning”, lines 8 - 9]:”we maximally make use of the model by not only using it for the policy gradient step, but also for training the Q-function.” The expected return is the evaluation of the proposed action given a state, provided by the critic, shaping the policy pi of the actor, affecting the future actions of the actor. Further, [page 1, paragraph 4, lines 3]: “For instance, we achieve a 10k return in the half-cheetah environment in just 50 trajectories.” The half-cheetah environment features a 2-dimensional robot for simulation purposes. The broadest reasonable interpretation of ‘autonomous device’ includes a simulated robot/character).
Regarding Claim 9:
Clavera, Geilinger and Kurutach teach The processor of claim 1. Clavera specifically teaches wherein the articulated autonomous device is a robot, a vehicle, a muscle-tendon network, a drone, a simulated character (Further, [page 1, paragraph 4, lines 3]: “For instance, we achieve a 10k return in the half-cheetah environment in just 50 trajectories.” The half-cheetah environment features a 2-dimensional robot for simulation purposes. The broadest reasonable interpretation of ‘autonomous device’ includes a simulated robot/character.)
Regarding claims 10, 11, 14 and 17 – 18:
Claims 10, 11, 14 and 17 – 18 are claims directed to computer implemented methods, having similar limitations as the processor claims 1, 2, 5 and 8 – 9. Therefore, claims 10, 11, 14 and 17 – 18 are rejected under the same rationale as claims 1, 2, 5 and 8 – 9.
Regarding claim 19:
Clavera teaches: A computer system comprising one or more processors and non-transitory computer-readable memory to store executable instructions that, as a result of being executed by the one or more processors, cause the computer system to ([Page 6, Algorithm I] The presence of the algorithm implies the presence of hardware capable of its execution. Specifically, the presence of the steps in the algorithm implies the presence of hardware that can be called a processor, which is capable of storing and executing machine-readable instructions that, in the art of computing, are typically transmitted to it from non-transitory, computer readable media. It is also common in the art of computing to use multiple processors to execute a set of instructions.)
The remaining limitations in claim 19 are similar to the limitations in claim 1 and are therefore rejected under the same rationale.
Regarding claims 20, 23 and 26 – 27:
Claims 20, 23 and 26 – 27 are claims directed to a computer system, having similar limitations as the processor claims 2, 5 and 8 – 9. Therefore, claims 20, 23 and 26 – 27 are rejected under the same rationale as claims 2, 5 and 8 – 9.
Regarding claim 28:
Clavera teaches: A non-transitory computer-readable memory storing executable instructions that, as a result of being executed by one or more processors of a computer system, cause the computer system to ([Page 6, Algorithm I] The presence of the algorithm implies the presence of hardware capable of its execution. Specifically, the presence of the steps in the algorithm implies the presence of hardware that can be called a processor, which is capable of storing and executing machine-readable instructions that, in the art of computing, are typically transmitted to it from non-transitory, computer readable media. It is also common in the art of computing to use multiple processors to execute a set of instructions.)
The remaining limitations in claim 28 are similar to the limitations in claim 1 and are therefore rejected under the same rationale.
Regarding claims 29, 32 and 35 – 36:
Claims 29, 32 and 35 – 36 are claims directed to a non-transitory computer-readable memory, having similar limitations as the processor claims 2, 5 and 8 – 9. Therefore, claims 29, 32 and 35 - 36 are rejected under the same rationale as claims 2, 5 and 8 – 9.
Claims 3, 7, 12, 16, 21, 25, 30 and 34 are rejected under 35 U.S.C. 103 as being unpatentable over Clavera, Geilinger, Kurutach and further in view of Ganin et al. US 20210271968 A1( hereinafter Ganin).
Regarding Claim 3:
Clavera, Geilinge and Kurutach teach The processor of claim 1.
Clavera, Geilinger and Kurutach do not explicitly teach the one or more neural networks are trained to accomplish a task by dividing the task into a plurality of action sequences; and training the one or more neural networks over each of the action sequences individually.
However, Ganin teaches, in the same field of endeavor: the one or more neural networks are trained to accomplish a task by dividing the task into a plurality of action sequences; ([Para 0004] “Once trained the system may be used to provide a sequence of instructions to control the same or another agent to perform the task e.g. in the real world.” Specifically, the neural networks are trained to generate sequences of actions with regards to specific tasks: [Para 0050] “the reinforcement learning neural network subsystem learns to generate a sequence of actions, i.e. control commands or instructions, for controlling the simulator to produce a final output which matches a training data item. These may be, for example, control commands for controlling a CAD program to produce a design,” Producing a design can be understood to be the specified task.)
and training the one or more neural networks over each of the action sequences individually ([Figure 3] shows the policy learner unit 220 using the unrolled Pi 112 (the policy) to generate sequences of actions for to optimize the generator objective 222. Examiner understands this to be the learning/training process. Also see [para 0045]: “The reinforcement learning neural network subsystem 110 comprises an action selection neural network, in implementations a policy recurrent neural network (RNN) 112, for generating the sequence of actions.” and [para 0057]: “The policy learner computing unit 220 is trained off-policy using the generated trajectory of experiences and the reward(s), e.g. by optimizing a generator objective L.sub.G 222”. This process can be described as using the action sequences generated by RNN 112 to train “A generative adversarial neural network system to provide a sequence of actions for performing a task” [Abstract]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to incorporate the combined teaching of Clavera, Geilinger and Kurutach with the sequential learning techniques in Ganin. One would do this as training a neural network over small action sequences to train models with individual tasks can improve model training efficiency and enables addressing tasks in a more granular way by finding action sequences that can fit very specific tasks.
Regarding Claim 7:
Clavera, Geilinge and Kurutach teach The processor of claim 1.
Clavera, Geilinge and Kurutach do not explicitly teach: wherein the one or more neural networks are trained at least in part my sampling a plurality of possible actions in parallel using a simulator.
However, Ganin teaches: wherein the one or more neural networks are trained at least in part my sampling a plurality of possible actions in parallel using a simulator. ([para 0013]: “In some implementations the actors and learner(s) of the actor-critic reinforcement learning subsystem may be implemented by different workers in a distributed environment. Thus there may be a plurality of actors each comprising a copy of the policy recurrent neural network coupled to a respective simulator to generate a plurality of simulator outputs. The actors may pool their experience in an experience buffer. For example the actors may store trajectories from the simulator outputs in the experience buffer, each trajectory comprising a sequence of the one or more actions and corresponding simulator output. One or more learners may then employ off-policy learning to update the parameters of the policy recurrent neural network using the stored trajectories.” Training in distributed environments by multiple learners is understood to be parallel training.)
Regarding claims 12, 16, 21, 25, 30 and 34:
Claim 12 and 16 are directed to a computer-implemented method, claims 21 and 25 are directed to a computer system and claim 30 and 34 are directed to a non-transitory, computer-readable memory, having similar limitations as the processor claims 3 and 7 respectively. Therefore, claims 12, 16, 21, 25, 30 and 34 are rejected under the same respective rationales.
Claims 4, 6, 13, 15, 22 , 24, 31 and 33 are rejected under 35 U.S.C. 103 as being unpatentable over Clavera, Geilinge, Kurutach and further in view of Peng et al. US 20210258235 A1 (hereinafter Peng).
Regarding Claim 4:
Clavera, Geilinge and Kurutach teach The processor of claim 1.
Clavera, Geilinge and Kurutach do not explicitly teach wherein the one or more actions are indicated based on a change to policy loss caused by a change to one or more inputs of a simulation.
However, Li teaches, in the same field of endeavor: wherein the one or more actions are indicated based on a change to policy loss caused by a change to one or more inputs of a simulation ([para 0016]: “The model may be trained using a simulator that uses network traces of past real streaming sessions (e.g., communication sessions) of users. During training, the decisions of the model (e.g., the "agent") are then used to calculate a reward. The reward is then used to calculate a loss function which is used to train the model. For example, the model may be a neural network and the loss function may be used to retrain the weights applied to the inputs to the neurons in one or more layers of the neural network.” [para 0025]: “the agent may be an asynchronous actor-critic (A3C) reinforcement learning model” The network traces are interpreted as input to the simulation. The decisions of the agent are informed by the actor/policy, which are based on inputs [para 0018]. A loss for this is calculated and used to update the weights of the neural network, causing an update to policy, which will inform future actions of the neural network).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the combined teaching of Clavera, Geilinge and Kurutach with the change to policy loss based on inputs in Peng. One would do this because accounting for the change in inputs reflected in the change of loss is essential in deciding how to change the simulation inputs to optimize the policy.
Regarding claims 13, 22 and 31:
Claim 13 is directed to a computer implemented method, claim 22 is directed to a computer system and claim 31 is directed to a non-transitory computer-readable memory, having similar limitations as claim 4. Therefore, claims 13, 22 and 31 is rejected under the same rationale.
Claims 6, 15, 24 and 33 are rejected under 35 U.S.C. 103 as being unpatentable over Clavera, Geilinge, Kurutach and in further view of Degrave et al. A Differentiable Physics Engine for Deep Learning in Robotics (hereinafter Degrave).
Regarding Claim 6:
Clavera, Geilinge and Kurutach teach The processor of claim 1.
Clavera, Geilinge and Kurutach do not explicitly teach wherein training of the one or more neural networks is based at least in part on a gradient of policy loss produced by a differentiable simulation.
However, Degrave teaches, in the same field of endeavor: wherein training of the one or more neural networks is based at least in part on a gradient of policy loss produced by a differentiable simulation. (The simulations are differentiable as they take place in a differentiable physics engine, and neural networks are trained: [page 4, paragraph 2.2 Policy Search, lines 1-2]: “To evaluate the relevance of our differentiable physics engine, we use a neural network as a general controller for a robot”. A plurality of neural networks was considered in Degrave: [page 5, figure 2 text, lines 2-3]: “The neural networks [plural] gdeep and hdeep with weights W receive sensor signals st from the sensors on the robot and use these to generate motor signals”. The neural networks are trained: [page 5, paragraph 2, lines 1-3]: “We need to train our network Gdeep […]” The objective of the training is to find a policy based on a gradient of policy loss: [page 4, 2.2 policy search, lines 5-6] “If Xt is the trajectory of the state up to time t-1, the goal is to find a policy ut = pi(Xt) such that we minimize the loss Lpi.” The loss L subscript pi is understood to be a policy loss. The gradient of this policy loss with regards to weights is considered in training: [page 5, paragraph 1, lines1-2]: “The gradient required for that is the Jacobian dL/dW, which is found with automatic differentiation software.”)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the combined teaching of Clavera, Geilinge and Kurutach with the gradient of policy loss in differentiable simulations in Degrave. One would do this, because “deep learning experience has taught us that optimizing with a gradient is often faster and more efficient. This fact is especially true when there are a lot of parameters, as is common in deep learning.” [Degrave, page 1, Introduction, pagraph 2, lines 3-4].
Regarding claims 15, 24 and 33:
Claim 15 is directed to a computer implemented method, claim 24 is directed to a computer system and claim 33 is directed to a non-transitory computer-readable memory, having similar limitations as claim 6. Therefore, claims 15, 24 and 33 are rejected under the same rationale.
Claim Rejections - 35 USC § 102
Alternative Rejection based on IDS submitted on 8/14/2025
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1-36 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Anonymous (Accelerated Policy Learning with Parallel Differentiable Simulation by Anonymous, Cited in IDS filed on 8/14/2025).
Regarding claim 1:
Anonymous teaches A processor, comprising: one or more circuits to use one or more neural networks (page 6, “4.1 EXPERIMENT SETUP”; page 8-9, “Scalability to high-dimensional problems”) to indicate one or more actions of an articulated autonomous device and to indicate how the one or more actions will affect one or more subsequent actions of the articulated autonomous device, (page 1, “1 Introduction” par. 1 and 3; page 4, Section 3.3, par. 1-3 and figure 3; action and critic networks work together in the model where actions are performed and subsequent result of the action or the next state of the action is indicated as an effect of the action using the transition function)
wherein the one or more neural networks include a dynamics model to model articulated dynamics for the articulated autonomous device without discontinuities with respect to contact and joint limits, the model including: a contact model to model contact dynamics with one or more other objects in an environment at least by approximating contact friction using a linear step function and by including a contact damping force; and a joint limit model employing a continuous penalty-based force (page 1, Abstract; page 3, “3.1 DIFFERENTIABLE DYNAMICS SIMULATION” last paragraph; page 14, “A.1 DIFFERENTIABLE SIMULATION DETAILS” ).
Regarding claim 2:
Anonymous teaches wherein the one or more neural networks includes: a first neural network that indicates an action to control the articulated autonomous device; and a second neural network that estimates a value of a current state to achieving a result (page 4, Section 3.3, par. 1-3 and figure 3; action and critic network to estimate action and corresponding result as a transition function).
Regarding claim 3:
Anonymous teaches wherein: the one or more neural networks are trained to accomplish a task by dividing the task into a plurality of action sequences; and training the one or more neural networks over each of the action sequences individually (page 4, Section 3.3, par. 1; page 5, par. 1-3; task is divided in sub-window and trained in episodes).
Regarding claim 4:
Anonymous teaches wherein the one or more actions are indicated based on a change to policy loss caused by a change to one or more inputs of a simulation (page 4-5, Section 3.3; differentiable simulator is used to calculated the transition function and policy loss is calculated (Equation 4 page 5).
Regarding claim 5:
Anonymous teaches wherein the one or more neural networks are trained by training independently for each of a plurality of time intervals (page 4, Section 3.3; par. 1; page 6 Algorithm; networks are trained in one interval/episode at a time).
Regarding claim 6:
Anonymous teaches wherein training of the one or more neural networks is based at least in part on a gradient of policy loss produced by a differentiable simulation (page 5, par. 2).
Regarding claim 7:
Anonymous teaches wherein the one or more neural networks are trained at least in part my sampling a plurality of possible actions in parallel using a simulator (page 5, Figure 3 with description and paragraph 1)).
Regarding claim 8:
Anonymous teaches wherein the one or more neural networks includes an actor network that learns a policy that controls the autonomous device and a critic network that learns to estimate a value of a state of the articulated autonomous device with respect to a result (page 4, section 3.3, par. 1).
Regarding claim 9:
Anonymous teaches wherein the articulated autonomous device is a robot, a vehicle, a muscle-tendon network, a drone, or a simulated character (page 2, figure 1).
Regarding claim 10-18, 19-27 and 28-36:
Claims 10-18, 19-27 and 28-36 they are method, system and program product claims having similar limitations as of apparatus claims 1-9 above. Therefore they are rejected under the same rational.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GISEL G FACCENDA whose telephone number is (703)756-1919. The examiner can normally be reached Monday - Friday 8:00 am - 4:00 pm.
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, Abdullah Al Kawsar can be reached at (571) 270-3169. 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.
/G.G.F./Examiner, Art Unit 2127
/ABDULLAH AL KAWSAR/Supervisory Patent Examiner, Art Unit 2127
Prosecution Insights