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 .
Remarks
This Office Action is responsive to Applicants' Amendment filed on February 9, 2026, in which claims 1-4, 11, 12, 14, 15, 17, 26, 27, and 29 are currently amended. Claims 1-15, 17-21, 26-27, and 29-31 are currently pending.
Response to Arguments
Applicant’s arguments with respect to rejection of claims 1-15, 17-21, 26-27, and 29-31 under 35 U.S.C. 101 based on amendment have been considered, however, are not persuasive.
With respect to Applicant's arguments on p. 10 of the Remarks submitted 2/9/2026 that explicit "means for" language "should not be interpreted as being directed to software alone, and any subsequent analysis premised on a "software per se" interpretation is based on an improper claim construction" and most importantly that “these claims are directed to physical structure”, Applicant’s arguments are persuasive. The software-per-se rejection is withdrawn in view of Applicant’s arguments.
With respect to Applicant's arguments on pp. 10-12 of the Remarks submitted 2/9/2026 that "when considered as a whole, claim 1 is not directed to a mental process", Examiner respectfully disagrees. The claims as a whole are directed towards receiving a classification, identifying an uncertainty metric, calculating normalized values based on the uncertainty metric, and selectively assigning the classification based on the calculated values. Each of these steps can be readily performed entirely in the mind with or without the assistance of tools such as pen and paper. The mere addition of generic computer components "at least one memory; machine readable instructions; and processor circuitry to at least one of instantiate or execute the machine readable instructions to" amounts to mere instructions to apply the judicial exception and does not integrate the judicial exception into a practical application (MPEP 2106.07(a)(II) "employing well-known computer functions to execute an abstract idea, even when limiting the use of the idea to one particular environment, does not integrate the exception into a practical application"). With respect to Applicant's arguments on p. 11 of the Remarks submitted 2/9/2026 that the Ex parte DesJardins decision "criticized approaches that "essentially equated any machine learning with an unpatentable 'algorithm'", Examiner notes that the instant claims do not even recite active use of a machine learning algorithm but rather analyzing and evaluating data broadly related to a machine learning algorithm. For at least these reasons Examiner asserts that it is reasonable and appropriate to maintain the subject matter eligibility rejection under 35 USC 101.
Applicant’s arguments with respect to rejection of claims 1-15, 17-21, 26-27, and 29-31 under 35 U.S.C. 103 based on amendment have been considered, however, are not persuasive.
With respect to Applicant’s arguments on p. 13 of the Remarks submitted 2/9/2026 that Rashid “does not teach or suggest controlling whether a predicted classification is assigned”, Examiner respectfully disagrees. First, Examiner would like to note that the actual instant claim language requires “selectively assign the first predicted classification to the first input when the first pairwise conflict value satisfies a threshold”. Examiner has mapped the pairwise conflict value to the vagueness in Rashid which is explicitly thresholded by top-L, and this threshold is used to selectively assign the predicted classification ([p. 24] “the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted”).
With respect to Applicant’s arguments on p. 13 of the Remarks submitted 2/9/2026 that Rashid does not teach or suggest “processor circuitry to calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN”, Examiner notes that Rashid is not used alone but in combination with Kandemir which provides the architectural details not explicitly provided by Rashid (Kandemir [¶0015] "the training signal may include a contribution for evidential classifier training and a contribution for density estimation. Thus, the model may determine the class probability values based on concentration parameters that are also trained to let the generative model reproduce the sensor data, leading to improved calibration." [¶0139] "The executable code may be stored in a transitory or non-transitory manner. Examples of computer readable mediums include memory devices, optical storage devices, integrated circuits, servers, online software, etc. FIG. 8 shows an optical disc 800. Alternatively, the computer readable medium 800 may comprise transitory or non-transitory data 810 representing model data representing a classification model, for example, trained according to a method described herein"). Rashid discloses “to calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN” (Rashid [p. 4] "A Posterior network [3] does not require arbitrary target prior distributions and access to OOD training data and instead, uses Normalizing Flows to predict a Dirichlet distribution over the predicted singleton class probabilities, from which the probabilistic uncertainties can be inferred." [p. 13 §3.6] "Belief mass assigned to a composite value indicates that there is evidence that fails to discriminate between specific singletons, thus resulting in vagueness. Given a hyper-opinion ω, the vague belief mass of an element yj ∈ R(Y) is given by [See Eqn. 3.16] [...] The total vague belief mass in an opinion ω is defined as the sum of belief masses on composite values yj ∈ C(Y) given by [See Eqn. 3.18]" [p. 16] "we introduce Grouped Dirichlet distribution in section 3.5, which is a special case of the Hyper Dirichlet distribution and has a closed-form solution for the normalization constant" [p. 23 §3.5] "We, therefore, focus on Grouped Dirichlet distribution (GDD) instead, which has a closed-form solution for the normalization constant allowing us to exactly calculate the expected value of p").
With respect to Applicant’s arguments on p. 13 of the Remarks submitted 2/9/2026 that Rashid does not teach or suggest “calculate a first pairwise conflict value based on the normalized values”, Examiner respectfully disagrees. Rashid discloses “to calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN” (Rashid [p. 4] "A Posterior network [3] does not require arbitrary target prior distributions and access to OOD training data and instead, uses Normalizing Flows to predict a Dirichlet distribution over the predicted singleton class probabilities, from which the probabilistic uncertainties can be inferred." [p. 13 §3.6] "Belief mass assigned to a composite value indicates that there is evidence that fails to discriminate between specific singletons, thus resulting in vagueness. Given a hyper-opinion ω, the vague belief mass of an element yj ∈ R(Y) is given by [See Eqn. 3.16] [...] The total vague belief mass in an opinion ω is defined as the sum of belief masses on composite values yj ∈ C(Y) given by [See Eqn. 3.18]" [p. 16] "we introduce Grouped Dirichlet distribution in section 3.5, which is a special case of the Hyper Dirichlet distribution and has a closed-form solution for the normalization constant" [p. 23 §3.5] "We, therefore, focus on Grouped Dirichlet distribution (GDD) instead, which has a closed-form solution for the normalization constant allowing us to exactly calculate the expected value of p") where vagueness is interpreted as the first pairwise conflict value.
For at least these reasons and those further detailed below, Examiner asserts that it is reasonable and appropriate to maintain the rejection in view of the combination of Rashid and Kandemir.
Specification
The specification is objected to as failing to provide proper antecedent basis for the claimed subject matter. See 37 CFR 1.75(d)(1) and MPEP § 608.01(o). Correction of the following is required: The instant specification does not define or provide support for a "conflict value" or "aggregated conflict value".
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The claims in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
This application includes one or more claim limitations that do not use the word “means,” but are nonetheless being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, because the claim limitation(s) uses a generic placeholder that is coupled with functional language without reciting sufficient structure to perform the recited function and the generic placeholder is not preceded by a structural modifier. Such claim limitation(s) is/are:
“means for identifying” in claim 26
“means for calculating” in claim 26
“means for assigning” in claim 26
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof. For example, the instant specification describes means for identifying as ([¶0036] “some or all of the circuitry of FIG. 2 may be implemented by microprocessor circuitry executing instructions to implement one or more virtual machines and/or containers” [¶0037] “ the uncertainty vector identification circuitry 200 can identify an uncertainty metric”), similarly, the means for calculating is described ([¶0040] “the prediction certification circuitry 114 includes means for calculating a dissonance score”), as well as the means for assigning ([¶0041] “the classification circuitry 204 assigns the predicted classification to the first input of the input data 104”). Applicant explicitly confirms on p. 10 of the Remarks submitted 2/9/2026 that the intended structure is physical.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
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 1-15, 17-21, 26-27, and 29-31 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.
Regarding claims 1, 14, and 26, "calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN" does not contain support in the instant specification. The instant specification does not mention "normalizing", "normalization", "normal", or any variation thereof that would support "calculat[ing] normalized values" as recited in the claims, much less to "calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN".
The remaining claims are rejected with respect to their dependence on the rejected claims.
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 1-15, 17-21, 26-27, and 29-31 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.
Regarding claims 1, 14, and 26, "indicating relative support for respective predicted classifications" is indefinite. The claim language attempts to add narrowing meaning "relative support", however, fails to provide any objective basis for relative comparison. The term “relative support” in claims 1, 14, and 26 is a relative term which renders the claim indefinite. The term “relative support” is not defined by the claim, the specification does not provide a standard for ascertaining the requisite degree, and one of ordinary skill in the art would not be reasonably apprised of the scope of the invention. In the interest of further examination the claim is interpreted as "based on respective predicted classifications".
The remaining claims are rejected with respect to their dependence on the rejected claims.
Claim Rejections - 35 USC § 101
101 Rejection
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 1-15, 17-21, 26, 27, and 29-31 are rejected under 35 USC § 101 because the claimed invention is directed to non-statutory subject matter.
Regarding Claim 1: Claim 1 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1 Analysis: Claim 1 is directed to an non-transitory computer-readable recording medium having stored therein a program that causes a computer to execute a process, which is directed to a product, one of the statutory categories.
Step 2A Prong One Analysis: Claim 1 under its broadest reasonable interpretation is a series of mental processes. For example, but for the generic computer components language, the above limitations in the context of this claim encompass neural network processing, including the following:
identify a first uncertainty metric associated with the EVDL NN, the first uncertainty metric corresponding to the first input of the EVDL NN (observation, evaluation, and judgement),
calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN (observation, evaluation, and judgement)
Calculate a first pairwise conflict value based on the normalized values; (observation, evaluation, and judgement)
Selectively assign the first predicted classification to the first input when the first pairwise conflict value satisfies a threshold (observation, evaluation, and judgement)
Therefore, claim 1 recites an abstract idea which is a judicial exception.
Step 2A Prong Two Analysis: Claim 1 recites additional elements “An apparatus comprising: at least one memory; machine readable instructions; and processor circuitry to at least one of instantiate or execute the machine readable instructions to”. However, these additional features are computer components recited at a high-level of generality, such that they amount to no more than mere instructions to apply the judicial exception using a generic computer component. An additional element that merely recites the words “apply it” (or an equivalent) with the judicial exception, or merely includes instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea, does not integrate the judicial exception into a practical application (See MPEP 2106.05(f)). Claim 1 also recites additional elements “receive a first predicted classification of a first input of an evidential deep learning neural network (EVDL NN)” which amounts to gathering and outputting data which is insignificant extra-solution activity (See MPEP 2106.05(g)). Therefore, claim 1 is directed to a judicial exception.
Step 2B Analysis: Claim 1 does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the lack of integration of the abstract idea into a practical application, the additional elements recited in claim 1 amount to no more than mere instructions to apply the judicial exception using a generic computer component and insignificant extra-solution activity. The gathering and outputting of data is considered well-understood, routine, and conventional in the art (See MPEP 2106.05(d)(II)(i)).
For the reasons above, claim 1 is rejected as being directed to non-patentable subject matter under §101. This rejection applies equally to independent claims 14 and 26, which recite a non-transitory computer readable media and an apparatus, respectively, as well as to dependent claims 2-13, 15, 17-21, 27, and 29-31. The additional limitations of the dependent claims are addressed briefly below:
Dependent claims 2, 15, and 27 recite additional observation, evaluation, and judgement “when the first pairwise conflict value does not satisfy the threshold: calculate a second pairwise conflict value based on the first pairwise conflict value and the first predicted classification; calculate a summed dissonance score based on the first and second pairwise conflict values; and when the aggregated pairwise conflict value satisfies the threshold, assign the first predicted classification to the first input.”
Dependent claim 3 recites additional instructions to apply the judicial exception using generic computer components “the EVDL NN is a recurrent model including one or more stages, each stage having one or more predicted classifications.” (a recurrent neural network comprising one or more stages is considered a generic computer component)
Dependent claims 4, 17, and 29 recite additional observation, evaluation, and judgement “identify a second uncertainty metric associated with the EVDL NN, the second uncertainty metric corresponding to a second input of the EVDL NN, the second input associated with a second predicted classification, the second predicted classification determined by the EVDL NN, the second predicted classification different from the first predicted classification; calculate a second pairwise conflict value based on the second uncertainty metric; and when the second pairwise conflict value satisfies the threshold, assign the second predicted classification to the second input”.
Dependent claims 5, 18, and 30 recite additional insignificant extra-solution activity “wherein the first input includes a first frame of a video and the second input includes a second frame of the video” which amounts to selection of a data type (see MPEP 2106.05(g)) which is well-understood, routine, and conventional in the art (See Cortivo ([Abstract] “We test our proposed methodologies on the well-known Vision dataset, which collects almost 2000 video sequences belonging to different devices” Vision dataset is a machine learning dataset comprising video frames for input into machine learning models).
Dependent claims 6, 19, and 31 recite additional observation, evaluation, and judgement “wherein the first predicted classification corresponds to a first action in the first frame and the second predicted classification corresponds to a second action in the second frame, the first action different from the second action”
Dependent claims 7 and 20 recite additional instructions to apply the judicial exception using generic computer components “wherein the EVDL NN is trained on second inputs, the second inputs different from the first input”
Dependent claims 8 and 21 recite additional observation, evaluation, and judgement “wherein the first uncertainty metric is a Dirichlet distribution”
Dependent claim 9 recites additional observation, evaluation, and judgement “the Dirichlet distribution includes a simplex, the simplex including at least two vertices”
Dependent claim 10 recites additional observation, evaluation, and judgement “wherein ones of the at least two vertices correspond to different predicted classifications”
Dependent claim 11 recites additional observation, evaluation, and judgement “wherein the first pairwise conflict value can include a value between 0 and 1”
Dependent claim 12 recites additional observation, evaluation, and judgement “wherein the first pairwise conflict value satisfies the threshold when the first pairwise conflict value is less than the threshold”
Dependent claim 13 recites additional observation, evaluation, and judgement “wherein the first predicted classification is determined by the EVDL NN”
Therefore, when considering the elements separately and in combination, they do not add significantly more to the inventive concept. Accordingly, claims 1-15, 17-21, 26, 27, and 29-31 are rejected under 35 U.S.C. § 101.
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 1-4, 7-15, 17, 20, 21, 26, 27, and 29 are rejected under U.S.C. §103 as being unpatentable over the combination of Rashid (“HYPER EVIDENTIAL NEURAL NETWORK”, 2022) and Kandemir (US20240071048A1).
Regarding claim 1, Rashid teaches receive a first predicted classification of a first input of an evidential deep learning neural network (EVDL NN);([p. 7] "HENN parameterized by θ that takes xi as input" [p. 16 §4] "In section 4.1, we introduce our proposed framework. An HENN takes an observation’s features as inputs and learns to output the observation evidences of singleton and non-overlapping composite values, which are then used to parameterize the corresponding GDD" [p. 24 §5.3] "the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples. A sample is a true positive if both its ground-truth and predicted label is a composite element" HENN interpreted as an EVDL NN. True/False positive interpreted as classifications of a first input of the EVDL NN.)
identify a first uncertainty metric associated with the EVDL NN, the first uncertainty metric corresponding to the first input of the EVDL NN;([p. 13 §3.6] "a(yj |yl) is the relative base rate" [See Eqn. 13] Relative base rate of vagueness interpreted as uncertainty metric. Rashid first takes the raw network input and calculates an evidence vector ri which is then used to calculate a belief mass (Eqn. 3.5) which is explicitly correlated with the relative base rate (Eqn. 3.16).)
calculate, based on the first uncertainty metric, normalized values indicating relative support for respective predicted classifications of the first input of the EVDL NN; and calculate a first pairwise conflict value based on the normalized values;([p. 4] "A Posterior network [3] does not require arbitrary target prior distributions and access to OOD training data and instead, uses Normalizing Flows to predict a Dirichlet distribution over the predicted singleton class probabilities, from which the probabilistic uncertainties can be inferred." [p. 13 §3.6] "Belief mass assigned to a composite value indicates that there is evidence that fails to discriminate between specific singletons, thus resulting in vagueness. Given a hyper-opinion ω, the vague belief mass of an element yj ∈ R(Y) is given by [See Eqn. 3.16] [...] The total vague belief mass in an opinion ω is defined as the sum of belief masses on composite values yj ∈ C(Y) given by [See Eqn. 3.18]" [p. 16] "we introduce Grouped Dirichlet distribution in section 3.5, which is a special case of the Hyper Dirichlet distribution and has a closed-form solution for the normalization constant" [p. 23 §3.5] "We, therefore, focus on Grouped Dirichlet distribution (GDD) instead, which has a closed-form solution for the normalization constant allowing us to exactly calculate the expected value of p" Vagueness interpreted as first pairwise conflict value.)
and selectively assign the first predicted classification to the first input when the first pairwise conflict value satisfies a threshold([p. 24 §5.3] "the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples. A sample is a true positive if both its ground-truth and predicted label is a composite element, and it is a false positive if its predicted label is a composite element but its ground-truth label is not. The larger the AUROC, the better the uncertainty is in discriminating between true composite and true singleton samples" HENN uncertainty measure is the vagueness ([p. 26] "Figure 5.3 shows the performances of HENN’s vagueness") which is explicitly used to compare against a Top-L threshold to determine false positives (assign the first predicted classification to the first input). If a first vagueness score isn't included in the Top-L of a first iteration L is increased until L equals N and all vagueness scores meet the threshold to form the ROC curve (interpreted as selectively assigned).).
However, Rashid does not explicitly teach An apparatus comprising: at least one memory; machine readable instructions; and processor circuitry to at least one of instantiate or execute the machine readable instructions to:.
Kandemir, in the same field of endeavor, teaches An apparatus comprising: at least one memory; machine readable instructions; and processor circuitry to at least one of instantiate or execute the machine readable instructions to:([¶0015] "the training signal may include a contribution for evidential classifier training and a contribution for density estimation. Thus, the model may determine the class probability values based on concentration parameters that are also trained to let the generative model reproduce the sensor data, leading to improved calibration." [¶0139] "The executable code may be stored in a transitory or non-transitory manner. Examples of computer readable mediums include memory devices, optical storage devices, integrated circuits, servers, online software, etc. FIG. 8 shows an optical disc 800. Alternatively, the computer readable medium 800 may comprise transitory or non-transitory data 810 representing model data representing a classification model, for example, trained according to a method described herein").
Rashid as well as Kandemir are directed towards evidential neural networks. Therefore, Rashid as well as Kandemir are analogous art in the same field of endeavor. It would have been obvious before the effective filing date of the claimed invention to combine the teachings of Rashid with the teachings of Kandemir by using a recurrent neural network as the evidential neural network on a computer system with non-transitory machine readable media. Kandemir provides as additional motivation for combination ([¶0022] "The recurrent model may receive the overall classification of the time series as an overall input, or may receive class probabilities at a respective time point as inputs for determining the probability distribution parameters at that time point, for example.") and it would be obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use a computer to implement the neural network which is explicitly reinforced by Kandemir.
Regarding claim 2, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the processor circuitry is to: when the pairwise conflict value does not satisfy the threshold:(Rashid [p. 24 §5.3] "In each case, the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples" See also FIG. 5.3. If a first vagueness score isn't included in the Top-L of a first iteration L is increased until L equals N and all vagueness scores meet the threshold to form the ROC curve.)
calculate a second pairwise conflict value based on the first pairwise conflict value and the first predicted classification;(Rashid [p. 13 §3.6] "Belief mass assigned to a composite value indicates that there is evidence that fails to discriminate between specific singletons, thus resulting in vagueness. Given a hyper-opinion ω, the vague belief mass of an element yj ∈ R(Y) is given by [See Eqn. 3.16] [...] The total vague belief mass in an opinion ω is defined as the sum of belief masses on composite values yj ∈ C(Y) given by [See Eqn. 3.18]" [p. 13 §3.6] " Total vague belief mass interpreted as second dissonance score. In Rashid the number of composite elements each having unique dissonance scores is calculated by 2^K-2-K where K is given as 3 in at least one example ([p. 13 §3.6] "an opinion has pluri-vagueness when several composite values have belief masses assigned to them. For example, let a = h1/3, 1/3, 1/3i and b = h0, 0, 0, 0, 0, 0.8i for K = 3"))
calculate an aggregated pairwise conflict value based on the first and second pairwise conflict values; and(Rashid [p. 13 §3.6] "The total vague belief mass in an opinion ω is defined as the sum of belief masses on composite values yj ∈ C(Y) given by [See Eqn. 3.18]")
when the aggregated pairwise conflict value satisfies the threshold, assign the first predicted classification to the first input.(Rashid [p. 24 §5.3] "the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples. A sample is a true positive if both its ground-truth and predicted label is a composite element, and it is a false positive if its predicted label is a composite element but its ground-truth label is not. The larger the AUROC, the better the uncertainty is in discriminating between true composite and true singleton samples" HENN uncertainty measure is the vagueness ([p. 26] "Figure 5.3 shows the performances of HENN’s vagueness") which is explicitly used to compare against a Top-L threshold to determine false positives (assign the first predicted classification to the first input).).
Regarding claim 3, the combination of Rashid and Kandemir teaches The apparatus of claim 2, wherein the EVDL NN is a recurrent model including one or more stages, each stage having one or more predicted classifications.(Kandemir [¶0024] "the recurrent generative model may be used at a time point t to predict a location of a nearest vehicle at one or more future time points t+1, t+2, by rolling out the generative model further in time." time points interpreted as stages.).
Regarding claim 4, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the processor circuitry is to: identify a second uncertainty metric associated with the EVDL NN, the second uncertainty metric corresponding to a second input of the EVDL NN, the second input associated with a second predicted classification, the second predicted classification determined by the EVDL NN, the second predicted classification different from the first predicted classification;(Rashid [p. 5 §2] "The input is a set of direct observations for different singleton and composite labels, and the number of observations for each singleton and composite label in this set corresponds to the amount of observation evidence available for that label" [p. 24 §5.3] "the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples. A sample is a true positive if both its ground-truth and predicted label is a composite element, and it is a false positive if its predicted label is a composite element but its ground-truth label is not. The larger the AUROC, the better the uncertainty is in discriminating between true composite and true singleton samples" Rashid explicitly anticipates a plurality of inputs and a plurality of uncertainty metrics corresponding to each respective input. True positive and false positive are different classifications.)
calculate a second pairwise conflict value based on the second uncertainty metric; and(Rashid [p. 13 §3.6] "Belief mass assigned to a composite value indicates that there is evidence that fails to discriminate between specific singletons, thus resulting in vagueness. Given a hyper-opinion ω, the vague belief mass of an element yj ∈ R(Y) is given by [See Eqn. 3.16] [...] The total vague belief mass in an opinion ω is defined as the sum of belief masses on composite values yj ∈ C(Y) given by [See Eqn. 3.18]" [p. 13 §3.6] " Total vague belief mass interpreted as second dissonance score. In Rashid the number of composite elements each having unique dissonance scores is calculated by 2^K-2-K where K is given as 3 in at least one example ([p. 13 §3.6] "an opinion has pluri-vagueness when several composite values have belief masses assigned to them. For example, let a = h1/3, 1/3, 1/3i and b = h0, 0, 0, 0, 0, 0.8i for K = 3"))
when the second pairwise conflict value satisfies the threshold, assign the second predicted classification to the second input.(Rashid [p. 24 §5.3] "the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples. A sample is a true positive if both its ground-truth and predicted label is a composite element, and it is a false positive if its predicted label is a composite element but its ground-truth label is not. The larger the AUROC, the better the uncertainty is in discriminating between true composite and true singleton samples" HENN uncertainty measure is the vagueness ([p. 26] "Figure 5.3 shows the performances of HENN’s vagueness") which is explicitly used to compare against a Top-L threshold to determine false positives (assign the first predicted classification to the first input).).
Regarding claim 7, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the EVDL NN is trained on second inputs, the second inputs different from the first input.(Rashid [p. 22] "we add n copies of each of its datapoints to the training set (where n is equal to the number of constituent singleton classes of the composite element) […] The testing and validation sets").
Regarding claim 8, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the first uncertainty metric is a Dirichlet distribution.(Rashid [p. 6 §3] "A more suitable PDF for inferring the singleton class probabilities is the Hyper Dirichlet distribution [9], as introduced in section 3.4, that accounts for any overlapping present between the singleton and composite values. However, the normalization constant of a Hyper Dirichlet PDF is intractable, so we introduce Grouped Dirichlet distribution in section 3.5, which is a special case of the Hyper Dirichlet distribution and has a closed-form solution for the normalization constant.").
Regarding claim 9, the combination of Rashid and Kandemir teaches The apparatus of claim 8, wherein the Dirichlet distribution includes a simplex, the simplex including at least two vertices.(Rashid See Eqn. 3.2 where the vector p is described as the hyper-probability distribution over the hyper-domain R(Y) which means pj is greater than or equal to 0 and the sum of pj=1, i.e. it lies on the (k-1)-dimensional simplex which has at least two vertices for all k>1. In Rashid K=3 (three vertices).).
Regarding claim 10, the combination of Rashid and Kandemir teaches The apparatus of claim 9, wherein ones of the at least two vertices correspond to different predicted classifications.(Rashid See Eqn. 3.2 where the vector p is described as the hyper-probability distribution over the hyper-domain R(Y) which means pj is greater than or equal to 0 and the sum of pj=1, i.e. it lies on the (k-1)-dimensional simplex which has at least two vertices for all k>1. In Rashid K=3 (three vertices). Each vertex corresponds to different classifications (See FIG. 3.2 where each corner of the triangle is a classification corresponding to a true/false positive classification).).
Regarding claim 11, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the first pairwise conflict value can include a value between 0 and 1.(Rashid [p. 8 §3.2] "u ∈ [0, 1] is the uncertainty mass").
Regarding claim 12, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the first pairwise conflict value satisfies the threshold when the first pairwise conflict value is less than the threshold.(Rashid [p. 24 §5.3] "In each case, the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples" See also FIG. 5.3. If a first vagueness score isn't included in the Top-L of a first iteration L is increased until L equals N and all vagueness scores meet the threshold to form the ROC curve.).
Regarding claim 13, the combination of Rashid and Kandemir teaches The apparatus of claim 1, wherein the first predicted classification is determined by the EVDL NN.(Rashid [p. 7] "HENN parameterized by θ that takes xi as input" [p. 16 §4] "In section 4.1, we introduce our proposed framework. An HENN takes an observation’s features as inputs and learns to output the observation evidences of singleton and non-overlapping composite values, which are then used to parameterize the corresponding GDD" [p. 24 §5.3] "the samples are ranked by their uncertainty measures in descending order and the top-L samples are taken as the predicted composite samples, where L ∈ {1, ..., N} with N being the total number of test samples. A sample is a true positive if both its ground-truth and predicted label is a composite element").
Regarding claims 14, 15, 17, 20, and 21, claims 14, 15, 17, 20, and 21 are directed towards a non-transitory computer readable medium for performing the methods of claims 1, 2, 4, 7, and 8, respectively. Therefore, the rejections applied to claims 1, 2, 4, 7, and 8 also apply to claims 14, 15, 17, 20, and 21.
Regarding claims 26, 27, and 29, claims 26, 27, and 29 are directed towards an apparatus for performing the methods of claims 1, 2, and 4, respectively. Therefore, the rejections applied to claims 1, 2, and 4 also apply to claims 26, 27, and 29.
Claims 5, 6, 18, 19, 30, and 31 are rejected under U.S.C. §103 as being unpatentable over the combination of Rashid and Kandemir and Chen (“Dual-Evidential Learning for Weakly-supervised Temporal Action Localization”, 2022).
Regarding claim 5, the combination of Rashid and Kandemir teaches The apparatus of claim 4.
However, the combination of Rashid and Kandemir doesn't explicitly teach, wherein the first input includes a first frame of a video and the second input includes a second frame of the video..
Chen, in the same field of endeavor, teaches The apparatus of claim 4, wherein the first input includes a first frame of a video and the second input includes a second frame of the video.("Fig. 3: Overall framework of the proposed DELU. After obtaining the snippet level evidence, we aggregate them to generate the video-level evidence by selecting the top-k snippets according to the attention scores. The video-level evidence and uncertainty are used to generalize the EDL paradigm for WS-TAL, and the snippet-level uncertainty is employed to generate dynamic weights for progressive learning. Note that we omit the regular classification loss Lcls (Section 3.2) in this figure for simplicity" See FIG. 3).
The combination of Rashid and Kandemir as well as Chen are directed towards evidential deep neural networks. Therefore, the combination of Rashid and Kandemir as well as Chen are analogous art in the same field of endeavor. It would have been obvious before the effective filing date of the claimed invention to combine the teachings of the combination of Rashid and Kandemir with the teachings of Chen by applying the recurrent evidential deep neural network to video data. Chen provides as additional motivation for combination ([p. 3 §1] "we find that it is desirable to tackle the action-background ambiguity by considering the uncertainty of classification results in both video and snippet levels. Recently, Evidential Deep Learning (EDL) [44, 36], which can collect the evidence of each category and quantify the predictive uncertainty, has received extensive attention and achieved impressive performance in a few computer vision tasks"). This motivation for combination also applies to the remaining claims which depend on this combination.
Regarding claim 6, the combination of Rashid, Kandemir, and Chen teaches The apparatus of claim 5, wherein the first predicted classification corresponds to a first action in the first frame and the second predicted classification corresponds to a second action in the second frame, the first action different from the second action.(Chen [p. 7 §3.2] "superscript (i) is used to indicate the sample index, i = 1, ..., N, and subscript j is used to indicate the category index. Note that in the following, for simplicity, the superscript (i) has been omitted when there is no ambiguity. Given an untrimmed video V and its corresponding multi-hot action category label y ∈ {0, 1} C+1, where C is the action category number, and C + 1 represents the non-action background class. The action instances in video V detected by WS-TAL methods can be formulated as a set of ordered quadruplets {cm, ts m, te m, ϕm}M m=1, where M is the number of action instances in V , cm denotes the action category, t s m and t e m denote the start and end timestamps, and ϕm denotes the confidence score").
Regarding claims 18 and 19, claims 18 and 19 are directed towards a non-transitory computer readable media for performing the methods of claims 5 and 6, respectively. Therefore, the rejections applied to claims 5 and 6 also apply to claims 18 and 19.
Regarding claims 30 and 31, claims 30 and 31 are directed towards an apparatus for performing the methods of claims 5 and 6, respectively. Therefore, the rejections applied to claims 5 and 6 also apply to claims 30 and 31.
Conclusion
THIS ACTION IS MADE FINAL. 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 extension fee 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 SIDNEY VINCENT BOSTWICK whose telephone number is (571)272-4720. The examiner can normally be reached M-F 7:30am-5:00pm EST.
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, Miranda Huang can be reached on (571)270-7092. 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.
/SIDNEY VINCENT BOSTWICK/Examiner, Art Unit 2124
/MIRANDA M HUANG/Supervisory Patent Examiner, Art Unit 2124
Prosecution Insights