Detailed Action
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Amendments
This action is in response to amendments filed September 29th, 2025, in which Claims 1, 5, 7, 11, 13, & 16 have been amended. No new claims have been added, and claims 4 & 10 have been cancelled. The amendments have been entered, and Claims 1-3, 5-9, & 11-18 are currently pending.
Response to Arguments
Regarding the applicant’s traversal of the 35 U.S.C. 101 rejections for claims 13-18 regarding claims being directed to “software per se” of the previous office action, the applicant’s arguments filed September 29th, 2025 have been fully considered, and are unpersuasive.
Applicant asserts that the amendment “executing on a data processor of a system for deriving further actionable evaluations from environmental inputs…” has resulted in the recitation of the claim as a system/device, which is now implemented by execution of a data processor, and thus makes the claims fall within the statutory category of a machine.
The examiner respectfully submits that, the structure of the amended claims is that of a neural network parameter tuner which is executed on a data processor of a system, meaning that the system is not being claimed, but is merely recited as the environment in which the claimed “neural network parameter tuner” operates. In other words, the system/machine is not being claimed itself. The examiner respectfully suggests claiming a system with at least one processor which executes a neural network parameter tuner, thus concretely claiming a machine rather than describing it as an environment for executing the claimed tuner. Therefore, these rejections are maintained.
Regarding the applicant’s traversal of the 35 U.S.C. 101 rejections for claims 1-18 regarding claims being directed to “abstract ideas” of the previous office action, the applicant’s arguments filed September 29th, 2025 have been fully considered, and are unpersuasive.
Applicant submits that the claims do not recite mental processes because the claims are tied to specific technological environment and further asserting that the cited limitations cannot be practically performed in the human mind.
The examiner respectfully asserts that the cited limitations from claim 1 recite deriving a value from an input data stream based upon detected noise components, assigning a detection threshold for the input data stream derived from that value, and classifying components of the input data stream based upon that threshold. The examiner respectfully asserts that all three of these things can be done mentally, as it essentially equivalates to seeing something (an input data stream), detecting some kind of noise (which may be visible and thus, mentally detected), and then classifying the noise based on that (making a mental assumption about the noise). Further, as cited in MPEP 2106.05(f), the mere use of a computer to perform an abstract idea, such as a mental process, does not preclude the limitation from reciting an abstract idea.
Further, applicant submits that the abstract ideas are integrated into a practical application and therefore recite something significantly more than the abstract idea, further clarifying that the derivation of an ambient classification value from the input data stream with a neural network being used as a basis for assigning a threshold that in turn, controls the classification operations improves the classification accuracy and efficiency of the neural network.
The examiner respectfully submits that the cited improvement is seemingly directed toward the abstract ideas themselves, and not “additional elements”, as is required, which can be shown in the MPEP 2106.04 at Prong Two:
“Prong Two asks does the claim recite additional elements that integrate the judicial exception into a practical application? In Prong Two, examiners evaluate whether the claim as a whole integrates the exception into a practical application of that exception. If the additional elements in the claim integrate the recited exception into a practical application of the exception, then the claim is not directed to the judicial exception (Step 2A: NO) and thus is eligible at Pathway B. This concludes the eligibility analysis. If, however, the additional elements do not integrate the exception into a practical application, then the claim is directed to the recited judicial exception (Step 2A: YES), and requires further analysis under Step 2B (where it may still be eligible if it amounts to an ‘‘inventive concept’’). For more information on how to evaluate whether a judicial exception is integrated into a practical application.”
Considering the cited portion of the MPEP, the abstract ideas themselves may not be relied upon as evidence of integration into a practical application, as the improvement must be found within additional elements of the claims rather than the abstract limitations. Therefore, the rejections are maintained.
Regarding the applicant’s traversal of the 35 U.S.C. 102/103 rejections of the previous office action, the applicant’s arguments filed September 29th, 2025 have been fully considered, and are unpersuasive.
Applicant asserts that claim 1 is not anticipated by MAIJALA, further asserting that since the threshold by which a sound is classified is manually adjusted as opposed to being adaptively tuned automatically, using paragraph 2 of section 3.3 as evidence, the reference fails to anticipate the limitation as claimed.
The examiner respectfully asserts that the claim recites “…assigning a detection threshold for the input data stream in response to an update to the derived ambient classification value, the detection threshold corresponding to the ambient classification value…” As shown in Figure 4 of MAIJALA, a “target likelihood” is assigned, by the neural network, based on the derived ambient classification value (the annotations). This value is then used to classify the input data stream based on the assigned detection threshold (“target likelihood”), as in the claim. This “target likelihood” value which is defined from the ambient classification value, determining how the data is classified, meets, under broadest reasonable interpretation, “a detection threshold” as claimed. Regardless of the fact that MAIJALA has a separate predetermined threshold value that is manually adjusted, the “target likelihood” value itself, in this case, is being cited as what is effectively acting as the “detection threshold” which corresponds to the ambient classification value and is adaptively determined by the model rather than being manually tuned. Therefore, while the examiner respectfully recognizes the difference in the invention described in the specification and MAIJALA, the claims as formed do not adequately represent this beyond other interpretations. Therefore, the rejections are maintained.
Independent claims 7 & 13 recite similar amended limitations as claim 1 and therefore, their rejections are maintained under the same rationale. Further, all dependent claims are dependent upon one of these three claims and their rejections have been subsequently maintained.
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 13-18 are further rejected under 35 U.S.C. 101 because they are directed to non-statutory subject matter (software per se).
Regarding claim 13, in Step 1 of the 101-analysis set forth in MPEP 2106, the claimed invention is directed to non-statutory subject matter. The claim does not fall within at least one of the four categories of patent eligible subject matter because a "neural network parameter tuner" is not explicitly defined in the specification, and under the broadest reasonable interpretation, is directed to software per se, with no structural basis and thus, does not qualify as one of the statutory categories of invention.
Regarding claims 14-18, they are dependent of claim 13 and are thereby directed to the same non-statutory subject matter with additional limitations.
Claims 1-3, 5-9, & 11-18 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea (mental process) without significantly more.
Regarding claim 1, in Step 1 of the 101-analysis set forth in MPEP 2106, the claim recites a method for adaptively tuning parameters for a neural network. A method is one of the four statutory categories of invention.
In Step 2a Prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
“deriving… an ambient classification value from the input data stream based upon detected noise components therein” (a person can mentally make a determination of a value simply by evaluating data, and making a judgement of how to set it (MPEP 2106).)
“assigning a detection threshold for the input data stream in response to an update to the derived ambient classification value, the detection threshold corresponding to the ambient classification value” (a person can mentally make a determination of a value simply by evaluating another value, and making a judgement of how to set it (MPEP 2106).)
“classifying… the signal components in the input data stream based upon the assigned detection threshold” (a person can mentally label (classify) something by evaluating data in comparison to a predetermined understanding (a threshold) and making a judgement of how to label it (MPEP 2106).)
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. According, the claim “recites” an abstract idea.
In Step 2a Prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
“receiving an input data stream derived from readings from a sensor device converted to digital data, the input data stream including signal components and noise components associated with ambient conditions” (Adding insignificant extra-solution activity (mere data gathering) to the judicial exception (MPEP 2106.05(g))
“feeding the input data stream to a neural network executing on a data processor of the system” (Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)))
“…with the neural network…” (uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))
Since the claim does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In Step 2b of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional element (v) recites generally linking the use of the judicial exception to a particular technological environment or field of use, which is not indicative of significantly more. Additional element (iv) recites an insignificant extra solution activity (mere data gathering). Further, element (iv) recites steps that store and retrieve information in memory, which has been determined by the courts to recite a well understood, routine and conventional activity which is not indicative of significantly more (see MPEP 2106.05(d)(II)). Additional element (vi) recites the application of a computer tool, which is not indicative of significantly more. Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Regarding claim 2, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 2 recites the following additional mental processes.
“deriving… a subsequent ambient classification value from a different part of the input data stream based upon detected noise components therein” (a person can mentally make a determination of a value simply by evaluating data, and making a judgement of how to set it (MPEP 2106).)
“updating the detection threshold for the input data stream for the subsequently derived ambient classification value” (A person can mentally update a value by evaluating data and another value, to make a judgement that it should be updated. (MPEP 2106).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claim 3, it is dependent upon claim 2, and thereby incorporates the limitations of, and corresponding analysis applied to claim 2. Further, claim 3 recites the following additional mental process.
“reclassifying… the signal components in the input data stream based upon the assigned updated detection threshold.” (a person can mentally label (classify) something by evaluating data in comparison to a predetermined understanding (a threshold) and making a judgement of how to label it (MPEP 2106).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claim 5, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 5 recites “wherein the input data stream is representative of audio, the sensor device being a microphone” (In step 2A prong 2, this recites Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)) which is not indicative of integration into a practical application. In step 2B, a general technological environment is not indicative of significantly more.)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claim 6, it is dependent upon claim 1, and thereby incorporates the limitations of, and corresponding analysis applied to claim 1. Further, claim 6 recites “wherein the neural network is a classification neural network” (In step 2A prong 2, this recites Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)) which is not indicative of integration into a practical application. In step 2B, a general technological environment is not indicative of significantly more.)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claim 7, it comprises similar limitations to claim 1 and is rejected under the same rationale, with the following addition: “an auxiliary neural network” (In step 2A prong 2, this recites Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)) which is not indicative of integration into a practical application. In step 2B, a general technological environment is not indicative of significantly more.)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claims 8-9, & 11-12, they are dependent upon claim 7, and thereby incorporate the limitations of, and corresponding analysis applied to claim 7. Further, claims 8-9, & 11-12 recite similar limitations to claims 2-3, & 5-6, respectively, and are rejected under the same rationale.
Regarding claim 13, In Step 2a Prong 1 of the 101-analysis set forth in the MPEP 2106, the examiner has determined that the following limitations recite a process that, under the broadest reasonable interpretation, covers a mental process but for recitation of generic computer components:
“…periodically deriving an ambient classification value from the input data stream based upon the noise components detected therein” (a person can mentally make a determination of a value simply by evaluating data, and making a judgement of how to set it (MPEP 2106).)
“the signal components therein being classified… based upon an assigned detection threshold corresponding to the ambient classification value” (a person can mentally label (classify) something by evaluating data in comparison to a predetermined understanding (a threshold) and making a judgement of how to label it (MPEP 2106).)
If claim limitations, under their broadest reasonable interpretation, covers performance of the limitations as a mental process but for the recitation of generic computer components, then it falls within the mental process grouping of abstract ideas. According, the claim “recites” an abstract idea.
In Step 2a Prong 2 of the 101-analysis set forth in MPEP 2106, the examiner has determined that the following additional elements do not integrate this judicial exception into a practical application:
“an auxiliary neural network executing on the data processor and receptive to an input data stream derived from readings from a sensor device converted to digital data, the input data stream including signal components and noise components associated with ambient conditions” (Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)))
“the auxiliary neural network” periodically deriving… (uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))
“a primary neural network executing on the data processor and receptive to the input data stream” (Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)))
classified “by the primary neural network” (uses a computer as a tool to perform an abstract idea (MPEP 2106.05(f))
Since the claim does not contain any other additional elements that are indicative of integration into a practical application, the claim is “directed” to an abstract idea.
In Step 2b of the 101-analysis set forth in the 2019 PEG, the examiner has determined that the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
As discussed above, additional elements, (iii) and (v), recite generally linking the use of the judicial exception to a particular technological environment or field of use, which is not indicative of significantly more. Additional elements, (iv) and (vi) recite the application of a computer tool, which is not indicative of significantly more. Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
Regarding claims 14-15, they are dependent upon claim 13, and thereby incorporate the limitations of, and corresponding analysis applied to claim 13. Further, they recite similar additional limitations to claim 2 and are rejected under the same rationale.
Regarding claim 16, it is dependent upon claim 13, and thereby incorporates the limitations of, and corresponding analysis applied to claim 13. Further, the claim cites the input device “…feeding the input data stream to the auxiliary neural network and the primary neural network” (In step 2A prong 2, this recites insignificant extra solution activity of mere data gathering, which is not indicative of integration into a practical application. In step 2B, this recites receiving data over a network which is a well-understood, routine and conventional activity (MPEP 2106.05(d)), which is not indicative of significantly more).
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claim 17, it is dependent upon claim 16, and thereby incorporates the limitations of, and corresponding analysis applied to claim 16. Further, claim 17 recites “wherein the input device is an audio transducer” (In step 2A prong 2, this recites Generally linking the use of the judicial exception to a particular technological environment or field of use (MPEP 2106.05(h)) which is not indicative of integration into a practical application. In step 2B, a general technological environment is not indicative of significantly more.)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible.
Regarding claim 18, it is dependent upon claim 13, and thereby incorporates the limitations of, and corresponding analysis applied to claim 13. Further, claim 18 comprises similar limitations to claim 6 and is rejected under the same rationale.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 1-3, & 5-6 are rejected under 35 U.S.C. 102(a)(1) as being clearly anticipated by Maijala, Panu, et al. “Environmental noise monitoring using source classification in sensors” Available on January 1 2018 (hereafter, MAIJALA).
Regarding claim 1, MAIJALA teaches “A method for adaptively tuning parameters for a neural network of a system for deriving further actionable evaluations from environmental inputs”:
([Abstract] “Environmental noise monitoring systems continuously measure sound levels without assigning these measurements to different noise sources in the acoustic scenes, therefore incapable of identifying the main noise source. In this paper a feasibility study is presented on a new monitoring concept in which an acoustic pattern classification algorithm running in a wireless sensor is used to automatically assign the measured sound level to different noise sources (automatically assigning sound levels to various noise sources is equivalent to adaptively adjusting parameters). A supervised noise source classifier is learned from a small amount of manually annotated recordings and the learned classifier is used to automatically detect the activity of target noise source in the presence of interfering noise sources” (based upon previously adapted parameters.))
Further, MAIJALA teaches “receiving an input data stream… the input data stream including signal components and noise components associated with ambient conditions”:
([3. Automatic detection of target sources, paragraph 1] “Sounds propagating from the target sources belong to a target class, whereas interfering noises as well as silence belong to a background class. Examples of possible target sounds are plant noise and aircraft noise. Possible background noises may be caused by e.g. traffic, wind, rain, thunder, and birds (noise components associated with ambient conditions). The activity of the target sources is detected by analysing continuous audio input and making binary classification between the background and the target.” Continuous audio input is an “input data stream” that contains “signal and noise components”, as audio data is well understood as such, as pointed out by the specification of the present application ([0006] “Conventional data processor devices have the capability of performing numerous operations in a short period of time, and so are well suited for capturing real-time signals from an environment and converting the same to a stream of digital data. For instance, audio signal captured by a microphone and transduced to an analog electrical signal thereby may be converted to a sequence of data that correspond to numerical voltage values of such analog electrical signal at discrete time intervals. These digital audio data streams may be readily transferred from one device to another, replayed, and manipulated as desired with digital signal processing algorithms.”))
Further, MAIJALA teaches the input data stream being “derived from readings from a sensor device converted to digital data”:
([2. Noise monitoring] “The smart sensor consist of a measurement microphone and a single-board computer with a wireless transmission unit.” A microphone is a well-known device that reads real soundwaves and converts them to digital data, as previously pointed out in the present specification as well ([0006] “Conventional data processor devices have the capability of performing numerous operations in a short period of time, and so are well suited for capturing real-time signals from an environment and converting the same to a stream of digital data. For instance, audio signal captured by a microphone and transduced to an analog electrical signal thereby may be converted to a sequence of data that correspond to numerical voltage values of such analog electrical signal at discrete time intervals. These digital audio data streams may be readily transferred from one device to another, replayed, and manipulated as desired with digital signal processing algorithms.”))
Further, MAIJALA teaches “feeding the input data stream to a neural network executing on a data processor of the system”:
([3.2. Supervised Classifiers, paragraph 3] “An ANN is used to estimate a function that yields desired outputs with given inputs [35]. The parameters of an ANN are estimated using training examples. A training example consists of an input feature vector x and a target vector y. When an ANN is used as a classifier [21], the target output is typically a vector y with the size of C, the number of classes. Given the feature vector y from class i, the target vector entry yi is set to 1, whereas other elements in target vector y are set to 0. Thus, the output of an optimised ANN classifier is interpreted as the activity indications of C classes of sound events. The activity indication is later called likelihood, since it is used in the same way as estimated likelihood in the GMM, though the indication is not a probability measurement. In the proposed system the multilayer perceptron (MLP) [36], which is a basic type of ANN, was used.” Here it is noted that one embodiment utilizes the artificial neural network as the acoustic model which handles classification. Further, the input data can be seen to be sent to the acoustic model for classification in Figure 4.
Further, MAIJALA teaches “deriving, with the neural network, an ambient classification value from the input data stream based upon detected noise components therein; assigning a detection threshold for the input data stream in response to an update to the derived ambient classification value, the detection threshold corresponding to the ambient classification value; and classifying, with the neural network, the signal components in the input data stream based upon the assigned detection threshold”:
(Figure 5)
It can be seen in Figure 5 that the audio data is assigned an ambient classification value from the input data stream based upon noise components therein (e.g. alarm, crusher, traffic, etc.) and it is assigned a likelihood threshold between 0.0 and 1.0 (a detection threshold from that value. Finally, it can be seen, that the target sounds in this case “alarm and crusher” classify as active in the end while “traffic” is identified as background noise and thus, “inactive”, thus classifying “the signal components in the input data stream based upon the assigned detection threshold”)
Regarding claim 2, MAIJALA teaches the limitations of claim 1. Further, MAIJALA teaches “deriving, with the neural network, a subsequent ambient classification value from a different part of the input data stream based upon detected noise components therein” and “updating the detection threshold for the input data stream for the subsequently derived ambient classification value.”:
([Figure 5])
The continuous audio input is shown to be detecting and classifying multiple sounds (e.g. alarm, crusher, traffic) which correlates to “a subsequent ambient classification value from a different part of the input data stream based on detected noise components therein” as well as the target applying various thresholds to each and updating the overall detection threshold to “active” or “inactive” based on the average, as cited in ([3.3. Training and monitoring, paragraph 2] “At the monitoring stage, a detection is made in one second non-overlapping segments. For each class, a score is computed as the sum of log-likelihoods (the logarithm of the likelihoods) of each frame in the corresponding second. The target likelihood in Fig. 5 is calculated as the score of target class divided by the sum score from all classes. The target sound source is detected as being active when the target likelihood is over a threshold (default value 0.5), otherwise inactive.”)
Regarding claim 3, MAIJALA teaches the limitations of claim 2. Further, MAIJALA teaches “reclassifying, with the network, the signal components in the input data stream based upon the assigned updated detection threshold.”:
(Figure 5)
It can be seen in the annotated figure above that both the crusher and traffic were classified at specific intervals and reclassified based on those values to be active in the end.
Regarding claim 5, MAIJALA teaches the limitations of claim 1. Further, MAIJALA teaches “wherein the input data stream is representative of audio, the sensor device being a microphone”:
([2. Noise monitoring] “The smart sensor consist of a measurement microphone and a single-board computer with a wireless transmission unit.”)
Regarding claim 6, MAIJALA teaches the limitations of claim 1. Further, MAIJALA teaches “wherein the neural network is a classification neural network”:
([3.2. Supervised Classifiers, paragraph 3] “An ANN is used to estimate a function that yields desired outputs with given inputs [35]. The parameters of an ANN are estimated using training examples. A training example consists of an input feature vector x and a target vector y. When an ANN is used as a classifier [21], the target output is typically a vector y with the size of C, the number of classes. Given the feature vector y from class i, the target vector entry yi is set to 1, whereas other elements in target vector y are set to 0. Thus, the output of an optimised ANN classifier is interpreted as the activity indications of C classes of sound events. The activity indication is later called likelihood, since it is used in the same way as estimated likelihood in the GMM, though the indication is not a probability measurement. In the proposed system the multilayer perceptron (MLP) [36], which is a basic type of ANN, was used. A neural network that is used to classify data is a classification neural network.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 7-9 & 11-18 are rejected under 35 U.S.C. 103 as being unpatentable over MAIJALA as applied to claims above, and further in view of Brownlee, Jason. “Ensemble Learning Methods for Deep Learning Neural Networks” Available on August 6 2019 (hereafter, BROWNLEE)
Regarding claim 7, it comprises similar limitations to claim 1 and is rejected under the same rationale, with the following exception: MAIJALA fails to explicitly teach the tasks being split between a “primary” neural network and an “auxiliary” neural network. However, analogous art about the benefits of ensemble learning (using multiple neural networks), BROWNLEE, does teach this.:
([Reduce Variance Using an Ensemble of Models, sentences 7-9] “This approach belongs to a general class of methods called “ensemble learning” that describes methods that attempt to make the best use of the predictions from multiple models prepared for the same problem. Generally, ensemble learning involves training more than one network on the same dataset, then using each of the trained models to make a prediction before combining the predictions in some way to make a final outcome or prediction. In fact, ensembling of models is a standard approach in applied machine learning to ensure that the most stable and best possible prediction is made.”)
It would be obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to combine the base reference of MAIJALA (methods for adaptively tuning parameters of a neural network using continuous audio input) with the teachings of BROWNLEE (The benefits of using more than one neural network in a single application) because the methods in MAIJALA use the neural network for multiple tasks, and the teachings of BROWNLEE show that multiple networks trained on the same data can work in tandem.
One of ordinary skill in the art would be motivated to do so because, as BROWNLEE teaches, ([Reduce Variance Using an Ensemble of Models, sentence 9] “ensembling of models is a standard approach in applied machine learning to ensure that the most stable and best possible prediction is made.”)
Regarding claim 8, MAIJALA in view of BROWNLEE teaches the limitations of claim 7. Further, MAIJALA teaches “deriving, with the… neural network, a subsequent ambient classification value from a different part of the input data stream based upon detected noise components therein” and “updating the detection threshold for the input data stream for the subsequently derived ambient classification value.”
([Figure 5])
The continuous audio input is shown to be detecting and classifying multiple sounds (e.g. alarm, crusher, traffic) which correlates to “a subsequent ambient classification value from a different part of the input data stream based on detected noise components therein” as well as the target applying various thresholds to each and updating the overall detection threshold to “active” or “inactive” based on the average, as cited in ([3.3. Training and monitoring, paragraph 2] “At the monitoring stage, a detection is made in one second non-overlapping segments. For each class, a score is computed as the sum of log-likelihoods (the logarithm of the likelihoods) of each frame in the corresponding second. The target likelihood in Fig. 5 is calculated as the score of target class divided by the sum score from all classes. The target sound source is detected as being active when the target likelihood is over a threshold (default value 0.5), otherwise inactive.”)
Regarding claim 9, MAIJALA in view of BROWNLEE teaches the limitations of claim 8. Further, MAIJALA teaches “reclassifying, with the… network, the signal components in the input data stream based upon the assigned updated detection threshold”:
(Figure 5)
It can be seen in the annotated figure above that both the crusher and traffic were classified at specific intervals and reclassified based on those values to be active in the end.
Regarding claim 11, MAIJALA in view of BROWNLEE teaches the limitations of claim 7. Further, MAIJALA teaches “wherein the input data stream is representative of audio,
the sensor device being a microphone”:
([2. Noise monitoring] “The smart sensor consist of a measurement microphone and a single-board computer with a wireless transmission unit.”)
Regarding claim 12, MAIJALA in view of BROWNLEE teaches the limitations of claim 7. Further, MAIJALA teaches “wherein the neural network is a classification neural network”:
([3.2. Supervised Classifiers, paragraph 3] “An ANN is used to estimate a function that yields desired outputs with given inputs [35]. The parameters of an ANN are estimated using training examples. A training example consists of an input feature vector x and a target vector y. When an ANN is used as a classifier [21], the target output is typically a vector y with the size of C, the number of classes. Given the feature vector y from class i, the target vector entry yi is set to 1, whereas other elements in target vector y are set to 0. Thus, the output of an optimised ANN classifier is interpreted as the activity indications of C classes of sound events. The activity indication is later called likelihood, since it is used in the same way as estimated likelihood in the GMM, though the indication is not a probability measurement. In the proposed system the multilayer perceptron (MLP) [36], which is a basic type of ANN, was used. A neural network that is used to classify data is a classification neural network.
Regarding claim 13, MAIJALA teaches “A neural network parameter tuner executing on a data processor of a system for deriving further actionable evaluations from environmental inputs”:
([Abstract] “Environmental noise monitoring systems (systems with data processors for deriving further actionable evaluations from environmental inputs) continuously measure sound levels without assigning these measurements to different noise sources in the acoustic scenes, therefore incapable of identifying the main noise source. In this paper a feasibility study is presented on a new monitoring concept in which an acoustic pattern classification algorithm running in a wireless sensor is used to automatically assign the measured sound level to different noise sources (automatically assigning sound levels to various noise sources is equivalent to adaptively adjusting parameters). A supervised noise source classifier is learned from a small amount of manually annotated recordings and the learned classifier is used to automatically detect the activity of target noise source in the presence of interfering noise sources” (based upon previously adapted parameters.))
Further, MAIJALA teaches “an… neural network executing on the data processor and receptive to an input data stream… including signal components and noise components…”:
(Figure 4)
the input data can be seen to be sent to the acoustic model for classification in Figure 4.
In Figure 4, the input data can be seen as audio recordings, which as cited in the present application’s specification, are known to contain signal components and noise components. ([0006] “Conventional data processor devices have the capability of performing numerous operations in a short period of time, and so are well suited for capturing real-time signals from an environment and converting the same to a stream of digital data. For instance, audio signal captured by a microphone and transduced to an analog electrical signal thereby may be converted to a sequence of data that correspond to numerical voltage values of such analog electrical signal at discrete time intervals. These digital audio data streams may be readily transferred from one device to another, replayed, and manipulated as desired with digital signal processing algorithms.”))
Further, MAIJALA teaches the input data stream being “derived from readings from a sensor device converted to digital data”:
([2. Noise monitoring] “The smart sensor consist of a measurement microphone and a single-board computer with a wireless transmission unit.” A microphone is a well-known device that reads real soundwaves and converts them to digital data, as previously pointed out in the present specification as well ([0006] “Conventional data processor devices have the capability of performing numerous operations in a short period of time, and so are well suited for capturing real-time signals from an environment and converting the same to a stream of digital data. For instance, audio signal captured by a microphone and transduced to an analog electrical signal thereby may be converted to a sequence of data that correspond to numerical voltage values of such analog electrical signal at discrete time intervals. These digital audio data streams may be readily transferred from one device to another, replayed, and manipulated as desired with digital signal processing algorithms.”))
Further, MAIJALA teaches those components being“…associated with ambient conditions”:
([3. Automatic detection of target sources, paragraph 1] “Sounds propagating from the target sources belong to a target class, whereas interfering noises as well as silence belong to a background class. Examples of possible target sounds are plant noise and aircraft noise. Possible background noises may be caused by e.g. traffic, wind, rain, thunder, and birds (associated ambient conditions). The activity of the target sources is detected by analysing continuous audio input and making binary classification between the background and the target.”)
Further, MAIJALA teaches “the… neural network periodically deriving an ambient classification value from the input data stream based upon the noise components detected therein; and a… neural network receptive to the input data stream, the signal components therein being classified by the… neural network based upon an assigned detection threshold corresponding to the ambient classification value.”:
(Figure 5)
It can be seen in Figure 5 that the audio data is assigned an ambient classification value from the input data stream based upon noise components therein (e.g. alarm, crusher, traffic, etc.) and it is assigned a likelihood threshold between 0.0 and 1.0 (a detection threshold from that value. Finally, it can be seen, that the target sounds in this case “alarm and crusher” classify as active in the end while “traffic” is identified as background noise and thus, “inactive”, thus classifying “the signal components in the input data stream based upon the assigned detection threshold”)
MAIJALA fails to explicitly teach the tasks being split between a “primary” neural network and an “auxiliary” neural network. However, analogous art about the benefits of ensemble learning (using multiple neural networks), BROWNLEE, does teach this. ([Reduce Variance Using an Ensemble of Models, sentences 7-9] “This approach belongs to a general class of methods called “ensemble learning” that describes methods that attempt to make the best use of the predictions from multiple models prepared for the same problem. Generally, ensemble learning involves training more than one network on the same dataset, then using each of the trained models to make a prediction before combining the predictions in some way to make a final outcome or prediction. In fact, ensembling of models is a standard approach in applied machine learning to ensure that the most stable and best possible prediction is made.”)
It would be obvious to one of ordinary skill in the art prior to the effective filing date of the claimed invention to combine the base reference of MAIJALA (methods for adaptively tuning parameters of a neural network using continuous audio input) with the teachings of BROWNLEE (The benefits of using more than one neural network in a single application) because the methods in MAIJALA use the neural network for multiple tasks, and the teachings of BROWNLEE show that multiple networks trained on the same data can work in tandem.
One of ordinary skill in the art would be motivated to do so because, as BROWNLEE teaches, ([Reduce Variance Using an Ensemble of Models, sentence 9] “ensembling of models is a standard approach in applied machine learning to ensure that the most stable and best possible prediction is made.”)
Regarding claims 14-15, MAIJALA in view of BROWNLEE teaches the limitations of claim 13. Further, claims 14-15 comprise similar limitations to claim 8 and are rejected under the same rationale.
Regarding claim 16, MAIJALA in view of BROWNLEE teaches the limitations of claim 13. Further, MAIJALA teaches “an input device connected to the data processor and feeding the input data stream to the… neural network”:
([3.2. Supervised Classifiers, paragraph 3] “An ANN is used to estimate a function that yields desired outputs with given inputs [35]. The parameters of an ANN are estimated using training examples. A training example consists of an input feature vector x and a target vector y. When an ANN is used as a classifier [21], the target output is typically a vector y with the size of C, the number of classes. Given the feature vector y from class i, the target vector entry yi is set to 1, whereas other elements in target vector y are set to 0. Thus, the output of an optimised ANN classifier is interpreted as the activity indications of C classes of sound events. The activity indication is later called likelihood, since it is used in the same way as estimated likelihood in the GMM, though the indication is not a probability measurement. In the proposed system the multilayer perceptron (MLP) [36], which is a basic type of ANN, was used.” Here it is noted that one embodiment utilizes the artificial neural network as the acoustic model which handles classification. Further, the input data can be seen to be sent to the acoustic model for classification in Figure 4.
Regarding claim 17, MAIJALA in view of BROWNLEE teaches the limitations of claim 16. Further, MAIJALA teaches “wherein the input device is an audio transducer”:
([2. Noise monitoring] “The smart sensor consist of a measurement microphone and a single-board computer with a wireless transmission unit.” A microphone is known to be an audio transducer, as pointed out in the present specification as well ([0006] “For instance, audio signal captured by a microphone and transduced to an analog electrical signal thereby may be converted to a sequence of data that correspond to numerical voltage values of such analog electrical signal at discrete time intervals.”))
Regarding claim 18, MAIJALA in view of BROWNLEE teaches the limitations of claim 13. Further, claim 18 comprises similar limitations to claim 12 and is rejected under the same rationale.
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 nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW LEE LEWIS whose telephone number is (571)272-1906. The examiner can normally be reached Monday: 12:00PM - 4:00PM and Tuesday - Friday: 12:00PM - 9PM.
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, Tamara Kyle can be reached at (571)272-4241. 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.
/Matthew Lee Lewis/Examiner, Art Unit 2144
/TAMARA T KYLE/Supervisory Patent Examiner, Art Unit 2144
Prosecution Insights