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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 March 22, 2026 has been entered.
Response to Amendment
The following is in response to the amendment filed on October 21st, 2025.
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 -7 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
MPEP 2106.04(a)(2)(III) “Accordingly, the "mental processes" abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, Judgments, and opinions.
Further, the MPEP recites “The courts do not distinguish between mental processes that are performed entirely in the human mind and mental processes that require a human to use a physical aid (e.g., pen and paper or a slide run) to perform the claim limitation.
MPEP 2106.04(a)(2)(I) “The mathematical concepts grouping is defined as mathematical relationships, mathematical formulas or equations, and mathematical calculations.”
With respect to claim 1:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Generating, with the processor, using a backbone neural network, a learning phase feature set including latent features based on the scene-specific training data; Generating a learning phase feature set from training data could be practically performed in the mind.
Generating, with the processor, using a classifier network a learning phase evaluation of a likelihood of each respective mode of a plurality of different modes for the future developments of the input scene based on the learning phase feature set, each different mode corresponding to a different set of possible future behaviors of the plurality of vehicles; Generating an evaluation based on a learning phase feature set could be practically performed in the mind.
Generating, with the processor, using a respective prediction model for each respective mode of the plurality of different modes for the future developments of the input scene a respective prediction for the future developments of the input scene determined by the scene-specific training data; Generating a prediction of future developments of an input scene could be practically performed in the mind.
Determining, with the processor, for each respective prediction model a deviation of the respective prediction from an actual development of the input scene and deriving from the deviation a realistic evaluation of the respective mode of the plurality of different mods; Determining a deviation from the prediction based on an actual development and deriving a realistic mode evaluation could be practically performed in the mind.
Step 2A, Prong 2:
The recited additional elements:
Receiving, with a processor, scene specific training data including aggregated scene-specific information of an input scene; (receiving data step adds insignificant extra-solution activity to the judicial exception, as discussed in MPEP 2106.05(g). This element represents generic computer functions)
Using a backbone neural network, using a classifier neural network, using a respective prediction model; Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level application of a backbone neural network, classifier neural network and prediction model)
Training, with the processor, at least one of (i) the backbone neural network to generate the latent features and (ii) the classifier neural network to evaluate the likelihood of each respective mode of a plurality of different modes, the training including modifying at least one of weights of the backbone neural network and weights of the classifier neural network such that the deviation between the learning phase evaluation and the realistic evaluation of the different modes is reduced. (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model by modifying weights.)
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.
Step 2B:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Receiving, with a processor, scene specific training data including aggregated scene-specific information of an input scene; (receiving data step is well understood routine and conventional, See MPEP 2106.05(d))
Using a backbone neural network, using a classifier neural network, using a respective prediction model; Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level application of a backbone neural network, classifier neural network and prediction model)
Training, with the processor, at least one of (i) the backbone neural network to generate the latent features and (ii) the classifier neural network to evaluate the likelihood of each respective mode of a plurality of different modes, the training including modifying at least one of weights of the backbone neural network and weights of the classifier neural network such that the deviation between the learning phase evaluation and the realistic evaluation of the different modes is reduced. (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model by modifying weights.)
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.
The claim is ineligible.
With respect to claim 2:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Each respective prediction model generates, as a prediction of the future development of the input scene, a deterministic and/or probabilistic prediction trajectory for each traffic participant in the input scene as the future development of the input scene; Generating as a prediction a trajectory for each traffic participant could be practically performed in the mind.
The deviations between the respective prediction trajectories and actual trajectories of the traffic participants from the input scene are respectively determined; and Determining deviations between a prediction and actual trajectory could be practically performed in the mind.
A realistic evaluation of the mode associated with the respective prediction models is derived based on the determined deviations. Deriving an evaluation based on deviations could be practically performed in the mind.
There are no additional elements in the claim, the claim is ineligible.
With respect to claim 3:
Step 2A Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
Generates a respective prediction for the future development of the input scene based on the training data. Generating a prediction for future development could be practically performed in the mind.
Step 2A Prong 2:
At least one of the respective prediction models is a pre-trained prediction neural network or a model-based prediction model. (mere instructions to apply the exception using a generic computer component).
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.
Step 2B:
At least one of the respective prediction models is a pre-trained prediction neural network or a model-based prediction model. (mere instructions to apply the exception using a generic computer component).
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.
The claim is ineligible.
With respect to claim 4:
Step 2A Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
The at least one previously untrained prediction neural network generates a network learning phase prediction for the future development of the input scene based on the training data and/or the learning phase feature set; Generating a prediction of future developments could be practically performed in the mind.
A deviation of the network learning phase prediction from the actual development of the input scene is determined and a realistic network evaluation of an associated mode is derived from the deviation; Determining a deviation from the actual development of the input scene and deriving a realistic evaluation could be practically performed in the mind.
Weights of the at least one previously untrained prediction neural network are modified such that a deviation between the network learning phase evaluation and the realistic network evaluation is reduced. Modifying weights to reduce deviation could be practically performed in the mind.
Step 2A Prong 2:
Training at least one previously untrained prediction neural network, wherein: (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model with previously determined data);
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.
Step 2B:
Training at least one previously untrained prediction neural network, wherein: (Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) – Examiner’s note: high level recitation of training a machine learning model with previously determined data);
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.
The claim is ineligible.
With respect to claim 5:
Step 2A, Prong 1:
A judicial exception is recited in this claim as it recites a mental process:
The weights of the backbone network and/or the weights of the classifier network and/or the weights of the at least one previously untrained prediction neural network are modified such that an entropy of the prediction of the prediction models is increased. Modifying weights could be practically performed in the mind.
There are no additional elements in the claim, the claim is ineligible.
Claims 6 and 7 are rejected according to claim 1.
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 1, 6 and 7 are rejected under 35 U.S.C. 103 as being unpatentable over Adetiloye (US PG Pub 2021/0020036) in view of Horvitz et al (KR 20060092909) and Hu et al (Probabilistic Future Prediction for Video Scene Understanding).
With respect to claim 1:
Adetiloye teaches:
A method comprising:
Receiving, with a processor, scene specific training data including aggregated scene-specific information of an input scene; (Paragraph [009], discloses collecting traffic image data from a plurality of cameras)
Generating, with the processor, using a backbone neural network, a learning phase feature set (textual data) including latent features (visual characteristics) based on the scene-specific training data; (Paragraph [037], discloses analyzing visual characteristics of the images to generate textual data)
Determining, with the processor, for each respective prediction model a deviation of the respective prediction from an actual development of the input scene and deriving from the deviation a realistic evaluation of the respective mode; (Paragraph [036], discloses optimizing the model by determining how accurately the prediction of the congestion is)
Training, with the processor, …, the training including modifying at least one of the weights of the backbone neural network and weights of the classifier neural network such that the deviation between the learning phase evaluation and the realistic evaluation of the different modes is reduced. (Paragraph [036], discloses training a neural network by finding optimum set of weights for images as to increase accuracy)
Adetiloye does not appear to explicitly disclose:
Generating, with the processor, using a classifier neural network, a learning phase evaluation of a likelihood of each respective mode of a plurality of different modes for the future developments of the input scene based on the learning phase feature set, each different mode corresponding to a different set of possible future behaviors of the plurality of vehicles;
Generating, with the processor, using a respective prediction model for each respective mode of the plurality of different modes for the future developments of the input scene a respective prediction for the future developments of the input scene determined by the scene-specific training data;
… at least one of (i) the backbone neural network to generate the latent features and (ii) the classifier neural network to evaluate the likelihood of each respective mode of a plurality of different modes,
Horvitz teaches:
Generating, with the processor, using a respective prediction model for each respective mode of the plurality of different modes for the future developments of the input scene a respective prediction for the future developments of the input scene determined by the scene-specific training data; (Page 8, Paragraph [004], discloses a plurality of prediction models. Page 14 Paragraph [007], discloses a prediction model being generated specifically for context data such as an unexpected event and using that model to predict a congestion point. This teaches a prediction model being specific for a context data)
It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to implement a method that utilized the teachings of Adetiloye and the teachings of Horvitz, both in the same field of invention. This would allow for providing users with advanced warning to take actions such as finding new alternatives (Horvitz, Page 4, Paragraph [001])
The combination of Adetiloye and Horvitz does not appear to explicitly disclose:Generating, with the processor, using a classifier neural network, a learning phase evaluation of a likelihood of each respective mode of a plurality of different modes for the future developments of the input scene based on the learning phase feature set, each different mode corresponding to a different set of possible future behaviors of the plurality of vehicles;
… at least one of (i) the backbone neural network to generate the latent features and (ii) the classifier neural network to evaluate the likelihood of each respective mode of a plurality of different modes, Hu teaches:Generating, with the processor, using a classifier neural network, a learning phase evaluation of a likelihood of each respective mode of a plurality of different modes for the future developments of the input scene based on the learning phase feature set, each different mode corresponding to a different set of possible future behaviors of the plurality of vehicles; (Pages 12 -13 and Figs. 3 and 4, disclose a model learning phase where entropy (read as the applications likelihood) for a plurality of futures (modes) is calculated, wherein each future is a different driving maneuver for the vehicles)
… at leats one of (i) the backbone neural network to generate the latent features and (ii) the classifier neural network to evaluate the likelihood of each respective mode of a plurality of different modes. (Pages 12 and 13 and Figs. 3 and 4 disclose using the trained model to calculate the entropy of a future of a plurality of possible futures) It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to implement a method that utilized the teachings of Adetiloye and Horvitz and the teachings of Hu, all in the same field of invention. This would generate visually diverse and plausible futures that substantially improve a learned autonomous driving policy. (Hu, Page 2, Introduction Section)
Claims 6 and 7 are rejected according to claim 1.
Claims 2-4 are rejected under 35 U.S.C. 103 as being unpatentable over Adetiloye (US PG Pub 2021/0020036) in view of Horvitz et al (KR 20060092909) and Hu et al (Probabilistic Future Prediction for Video Scene Understanding) and further in view of Janjos et al (Self-Supervised Action-Space Prediction for Automated Driving).
With respect to claim 2:
The combination of Adetiloye, Horvitz and Hu does not appear to explicitly disclose:
Each respective prediction model generates, as a prediction of the future development of the input scene, a deterministic and/or probabilistic prediction trajectory for each traffic participant in the input scene as the future development of the input scene;
The deviations between the respective prediction trajectories and actual trajectories of the traffic participants from the input scene are respectively determined; and a realistic evaluation of the mode associated with the respective prediction models is derived based on the determined deviations.
Janjos discloses:
Each respective prediction model generates, as a prediction of the future development of the input scene, a deterministic and/or probabilistic prediction trajectory for each traffic participant in the input scene as the future development (trajectories) of the input scene; (Page 1, Introduction section, discloses predicting trajectories for all vehicles in an input scene)
The deviations between the respective prediction trajectories and actual trajectories of the traffic participants from the input scene are respectively determined; and a realistic evaluation of the mode associated with the respective prediction models is derived based on the determined deviations. (Page 7, Performance section, teaches evaluating performance by calculating metrics and evaluating performance by calculating metrics and evaluating them against ground truth (actual trajectories))
It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to implement a method that utilized the teachings of the combination of Adetiloye, Horvitz and Hu and the teachings of Janjos, all in the same field of invention. This would allow for more accurate long term predictions (Janjos, Page 1, Introduction section)
With respect to claim 4:
The combination of Adetiloye, Horvitz and Hu teaches:
Weights of the at least one previously untrained prediction network are modified such that a deviation between the network learning phase evaluation and the realistic network evaluation is reduced. (Adetiloye, Paragraph [036], discloses weights can be modified to predict a more accurate prediction score)
The combination of Adetiloye, Horvitz and Hu does not appear to explicitly disclose:
Training at least one previously untrained prediction network, wherein:
The at least one previously untrained prediction network generates a network learning phase prediction for the future development of the input scene based on the training data and/or the learning phase feature set;
A deviation of the network learning phase prediction from the actual development of the input scene is determined and a realistic network evaluation of an associated mode is derived from the deviation.
Janjos teaches:
Training at least one previously untrained prediction network, wherein: the at least one previously untrained prediction network generates a network learning phase prediction for the future development of the input scene based on the training data and/or the learning phase feature set; (Page 7, training section, discloses training a model with training data and action reconstruction)
A deviation of the network learning phase prediction from the actual development of the input scene is determined and a realistic network evaluation of an associated mode is derived from the deviation. (Page 7, Performance Section, discloses evaluating performance against ground truth of the data sets)
Claims 5 is rejected under 35 U.S.C. 103 as being unpatentable over Adetiloye (US PG Pub 2021/0020036) in view of Horvitz et al (KR 20060092909), Hu et al (Probabilistic Future Prediction for Video Scene Understanding) and Janjos et al (Self-Supervised Action-Space Prediction for Automated Driving) and further in view of Tsou et al (US Patent 11,315,045).
With respect to claim 5:
The combination of Adetiloye, Horvitz, Hu and Janjos does not appear to explicitly disclose:
The weights of the backbone network and/or the weights of the classifier network and/or the weights of the at least one previously untrained prediction network are modified such that an entropy of the prediction of the prediction modules is increased.
Tsou teaches:
The weights of the backbone network and/or the weights of the classifier network and/or the weights of the at least one previously untrained prediction network are modified such that an entropy of the prediction of the prediction modules is increased. (Column 9 Lines 55-64, discloses using weights to increase the entropy of decision trees).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the present application to implement a method that utilized the teachings of Adetiloye, Horvitz, Hu, Janjos and the teachings of Tsou, all in the same field of invention. This would allow for random forests to be robust for decision purposes (Tsou, Column 9 Lines 63-65).
Response to Arguments
Claim Rejections - 35 USC § 101
With respect to the 101 rejection of claims 1-7:
Applicant argues “these features are integrated in a practical application by the non-abstract process of using training neural networks in a multi-stage architecture to predict future developments of a traffic scene”. Examiner respectfully disagrees. The claim does not reflect the improvement, the training step recites training a neural network such that a deviation is reduced but there is no recitation of using that trained neural network to predict future developments of a traffic scene.
Claim Rejections - 35 USC § 103
Applicant’s arguments have been considered but are moot in view of the new grounds of rejection.
Conclusion
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MARIELA D REYES whose telephone number is (571)270-1006. The examiner can normally be reached Monday-Friday, 7:30 am -5:00 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, David Wiley can be reached at (571) 272-3923. 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.
/Mariela Reyes/ Supervisory Patent Examiner, Art Unit 2142
Prosecution Insights