DETAILED ACTION
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 5 March 2026 has been entered.
Status
This Office Action is in response to the communication filed on 5 March 2026. Claims 6-7 and 16-17 have been cancelled currently or previously, claims 10 has been amended, and no new claims have been added. Therefore, claims 1-5, 8-15, and 18-20 are pending and presented for examination.
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 .
Response to Amendment
A summary of the Examiner’s Response to Applicant’s amendment:
Applicant’s amendment does not overcome the rejection(s) under 35 USC § 101; therefore, the Examiner maintains the rejection(s) while updating phrasing in keeping with current examination guidelines.
Applicant’s amendment does not overcome the prior art rejection(s) under 35 USC §§ 102 or 103, except as to claim 10; therefore, the Examiner maintains the rejection(s) as below.
Applicant’s amendment overcomes the rejection(s) under 35 USC §§ 102 and/or 103 as to claim 10; therefore, the Examiner places new grounds of rejection for claim 10.
Applicant’s arguments are found to be not persuasive; please see the Response to Arguments below.
Priority
Applicant’s claim for benefit under 35 U.S.C. 119(e) to U.S. Provisional Application No. 63/434,133, filed on 21 December 2022, is acknowledged.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 14 December 2023 was filed after the mailing date of the application on 14 December 2023. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Interpretation
The Examiner notes that “a skill embedding” is claimed, where Applicant’s specification does not define either a skill or an embedding, a “skill” is apparently the action(s) taken or applied (e.g., treatment(s)) by medical professionals as correlated to the patient status. See, e.g., Applicant ¶ 0019, “skill imitation system 108 learns skills applied by the medical professionals 102 in response to developing patient conditions. For example, the medical records 106 may include historical patient healthcare conditions (e.g., biometric information and a description of symptoms) and actions taken by the medical professionals 102 responsive to those conditions”. An “embedding” as best understood by the context of the specification is a mathematical representation of words or phrases – in this case, the “prototype” of states and actions.
Applicant argues (Remarks submitted 25 September 2025 at p. 7) the interpretation, indicating that “For example, Applicant respectfully asserts that a ‘skill’ need not be limited to medical actions”. However, the above indication is not limited to “medical actions”, the interpretation is inclusive of any “action(s) taken or applied” – i.e., both medical and non-medical actions; however, the light of the specification is that the skill(s) or action(s) would be performed by medical professionals in general.
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.
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.
Claim 10 is 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.
Claim 10 recites “wherein the treatment action includes an instruction to a treatment system to automatically administer an updated dialysis treatment to a patient by updating operational parameters of a dialysis machine in response to an undesirable albumin reading for the patient”. The Examiner has searched for the concept of albumin readings as the basis for administering dialysis or updating parameters for dialysis, but does not find it. The only indication of “albumin” is Applicant ¶ 0024 indicating “blood test measurements may be taken regularly, for example at a frequency of twice per month, and may measure such factors as albumin, glucose, and platelet count. The CTR may also be taken regularly, for example at a frequency of once per month. Dynamic information may also be recorded during the dialysis session, for example using sensors in the dialysis machine 204. The dynamic information may be modeled as time series over their respective frequencies”. Dialysis is discussed in the specification (at Applicant ¶¶ 0022-0024), and the specification indicates that “operational parameters of a dialysis machine 204 or any other system in a hospital or other healthcare facility many be monitored, along with a history of past events at the system, to predict events” (at Applicant ¶ 0025). However, there does not appear to be any indication that dialysis is administered based on any albumin reading, nor that there is or would be any undesirable, unhealthy, abnormal, or out-of-range albumin reading, nor that operational parameters of a dialysis machine are adjusted based on albumin readings. Albumin may be measured, and dialysis parameters may be monitored, but there is no nexus indicated so as to “automatically administer an updated dialysis treatment to a patient by updating operational parameters of a dialysis machine in response to an undesirable albumin reading for the patient”.
Claim 10 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 10 recites “wherein the treatment action includes an instruction to a treatment system to automatically administer an updated dialysis treatment to a patient by updating operational parameters of a dialysis machine in response to an undesirable albumin reading for the patient”. However, it is indefinite whether the “treatment action” being claimed is among the data used for training, or if this is somehow a selected result, or if claim 10 has any actual limiting affect. Parent claim 1 indicates “training a machine learning model”, where that training is “based on segments of the patient trajectory” and that trajectory “include[es] a prototype layer” learning classes of trajectory segments and “an imitation learning layer that learns a policy to select a treatment action”. So, it is the learned policy that “select[s] a treatment action”, and the “imitation learning layer that learns a policy” must then apparently have selected treatments as learning or training data so that it could learn the policy and select a treatment action.
If the instruction to administer an updated dialysis treatment is training data, it is just that: data itself, used to train the model.
If the instruction to administer an updated dialysis treatment is supposed to be a result, then the instruction is never indicated as sent, and if it were sent to “a treatment system”, that treatment system (i.e., apparently, a “dialysis machine”) is not part of the claimed system – claim 10 would apparently not offer any actual limitation to parent claim 1.
Therefore, the Examiner is completely uncertain what claim 10 does, or is intended to recite. Claim 10 is therefore indefinite as to whether the claim is data provided for training the machine learning model, or an intended result the model is supposed to arrive at, and also what limitation claim 10 would have for parent independent claim 1.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Please see the following Subject Matter Eligibility (“SME”) analysis:
For analysis under SME Step 1, the claims herein are directed to a method (claims 1-5 and 8-10) and system (claims 11-15 and 18-20), which would be classified under one of the listed statutory classifications (SME Step 1=Yes).
For analysis under revised SME Step 2A, Prong 1, independent claim 1 recites a computer-implemented method for training a healthcare treatment machine learning model, comprising: segmenting a patient trajectory, which includes a sequence of patient states and treatment actions; and training a machine learning model based on segments of the patient trajectory, including a prototype layer that learns prototype vectors representing respective classes of trajectory segments and an imitation learning layer that learns a policy to select a treatment action based on an input state and a skill embedding, wherein training the machine learning model includes minimizing a loss function that includes a clustering structure regularization term, a prototype segment evidence regularization term, a diversity regularization term, and an imitation learning term expressed as:
PNG
media_image1.png
59
339
media_image1.png
Greyscale
where m is a length of a segment, n is a number of segments, πE is an expert policy, at(j) is an action performed at step t for segment j, st(j) is a patient state at step t for segment j, πθ is a learned policy, and ot(j) is a skill embedding at step t for segment j.
Independent claim 11 is analyzed in the same manner as claim 1 since directed to a system for training a healthcare treatment machine learning model, comprising: a hardware processor; and a memory that stores a computer program which, when executed by the hardware processor, causes the hardware processor to perform the same or similar activities as at claim 1 above.
The dependent claims (claims 2-5, 8-10, 12-15, and 18-20) appear to be encompassed by the abstract idea of the independent claims since they merely indicate using embedding at a layer of the machine learning (ML) model (claims 2 and 12) including a multilayer perceptron and a convolutional layer (claims 3 and 13), measuring a patient state, selecting a treatment action based on the ML model, and notifying a medical professional (claim 4), the skill embedding including weighted prototype vectors based on similarity of the prototype vectors to trajectory (claims 5 and 15), the ML model including a cluster structure regularization term expressed as particular formulas (claims 8 and 18), the treatment actions including a prescription, meal provision, rehabilitation, or discharge destination plan (claims 9 and 19) or instructions to automatically administer or adjust treatment (such as according to albumin readings, as at claim 10) (claims 10 and 20), and/or the prototype layer determining similarity between segments and prototype vectors (claim 14).
The underlined portions of the claims are an indication of elements additional to the abstract idea (to be considered below).
The claim elements may be summarized as the idea of modeling to select a treatment action based on patient state and previous actions; however, the Examiner notes that although this summary of the claims is provided, the analysis regarding subject matter eligibility considers the entirety of the claim elements, both individually and as a whole (or ordered combination). This idea is within the following grouping(s) of subject matter:
Mathematical concepts (e.g., relationships, formulas, equations, and/or calculations) based on the explicit formulas modeling and the embeddings used and as representing the mathematics performed;
Certain methods of organizing human activity (e.g. … commercial or legal interactions such as agreements, contracts, legal obligations, advertising, marketing or sales activities/behaviors, or business relations; and/or managing personal behavior or relationships between people such as social activities, teaching, and following rules or instructions) as based on the claims essentially and/or actually emulating what medical professions do and have done – consider past (i.e., prototype) actions in response to states and treatments so as to recommend a treatment action based on current and past information; and
Mental processes (e.g., concepts performed in the human mind such as observation, evaluation, judgment, and/or opinion) as based on the treatment action being selected based on current and past observations, evaluations, judgments, and/or opinions as related to a current situation and patient.
Therefore, the claims are found to be directed to an abstract idea.
For analysis under revised SME Step 2A, Prong 2, the above judicial exception is not integrated into a practical application because the additional elements do not impose a meaningful limit on the judicial exception when evaluated individually and as a combination. The additional elements are that the method is computer-implemented (at claim 1) and using a system … comprising: a hardware processor; and a memory that stores a computer program which, when executed by the hardware processor, causes the hardware processor to perform the same or similar actions (at claim 11). These additional elements do not reflect an improvement in the functioning of a computer or an improvement to other technology or technical field, effect a particular treatment or prophylaxis for a disease or medical condition (there is no medical disease or condition, much less a treatment or prophylaxis for one), implement the judicial exception with, or by using in conjunction with, a particular machine or manufacture that is integral to the claim, effect a transformation or reduction of a particular article to a different state or thing (there is no transformation/reduction of a physical article), and/or apply or use the judicial exception in some other meaningful way beyond generically linking use of the judicial exception to a particular technological environment.
The claims appear to merely apply the judicial exception, include instructions to implement an abstract idea on a computer, or merely use a computer as a tool to perform the abstract idea. The additional elements appear to merely add insignificant extra-solution activity to the judicial exception and/or generally link the use of the judicial exception to a particular technological environment or field of use.
For analysis under SME Step 2B, the claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the additional elements, as indicated above, are merely “[a]dding the words ‘apply it’ (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, e.g., a limitation indicating that a particular function such as creating and maintaining electronic records is performed by a computer, as discussed in Alice Corp.” that MPEP § 2106.05(I)(A) indicates to be insignificant activity
There is no indication the Examiner can find in the record regarding any specialized computer hardware or other “inventive” components, but rather, the claims merely indicate computer components which appear to be generic components and therefore do not satisfy an inventive concept that would constitute “significantly more” with respect to eligibility. Applicant ¶¶ 0061-0066 appear to be the only description of the computers and/or components used, and indicate that generic or general-purpose computers are envisioned.
The individual elements therefore do not appear to offer any significance beyond the application of the abstract idea itself, and there does not appear to be any additional benefit or significance indicated by the ordered combination, i.e., there does not appear to be any synergy or special import to the claim as a whole other than the application of the idea itself.
The dependent claims, as indicated above, appear encompassed by the abstract idea since they merely limit the idea itself; therefore the dependent claims do not add significantly more than the idea.
Therefore, SME Step 2B=No, any additional elements, whether taken individually or as an ordered whole in combination, do not amount to significantly more than the abstract idea, including analysis of the dependent claims.
Please see the Subject Matter Eligibility (SME) guidance and instruction materials at https://www.uspto.gov/patent/laws-and-regulations/examination-policy/subject-matter-eligibility, which includes the latest guidance, memoranda, and update(s) for further information.
NOTICE
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 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.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-5, 8, 11-15, 18, and 20 are rejected under 35 U.S.C. 102(a)(1) as being unpatentable over Waner et al. (U.S. Patent Application Publication No. 2018/0015282, hereinafter Waner) in view of Xu et al. (U.S. Patent Application Publication No. 2020/0364504, hereinafter Xu).
Claim 1: Waner discloses a computer-implemented method for training a healthcare treatment machine learning model, comprising:
segmenting a patient trajectory, which includes a sequence of patient states and treatment actions (see Waner at least at, e.g., ¶ 0082, “the stimulation component 100 may record the parameters associated with given stimulation pulses over time, and this data, along with physiological data collected from the sensing component 50, may be used by a clinician to titrate the system or to otherwise analyze a sleep state of the patient. The collected data may also be used to generate patient alarms and reports on patient treatment and status. The collected data may also be stored in memory 601 or wirelessly communicated to a clinician. For example, the collected data may be used to identify irregularities in the sleep patterns and if appropriate take action, e.g., send an alert for help. The data collected over time can be useful to identify problems early on, e.g., worsening breathing patterns, worsening sleeping problems”; citation hereafter by number only); and
training a machine learning model based on segments of the patient trajectory, including a prototype layer that learns prototype vectors representing respective classes of trajectory segments and an imitation learning layer that learns a policy to select a treatment action based on an input state and a skill embedding (0082, “The data can then be considered by the treating healthcare professional and/or automatically assessed by the processor. Likewise, by collecting data from an individual patient over time, the system can “learn” patient specific patterns of sleep and patient specific patterns of apnea, e.g., via machine learning algorithms, which can enable the system to predict when an event is likely to occur and enable the system to calibrate and select to what level to activate the device”).
Waner, however, does not appear to explicitly disclose wherein training the machine learning model includes minimizing a loss function that includes a clustering structure regularization term, a prototype segment evidence regularization term, a diversity regularization term, and an imitation learning term that is expressed as:
PNG
media_image1.png
59
339
media_image1.png
Greyscale
where m is a length of a segment, n is a number of segments, πE is an expert policy, at(j) is an action performed at step t for segment j, st(j) is a patient state at step t for segment j, πθ is a learned policy, and ot(j) is a skill embedding at step t for segment j. Xu, however, teaches that a “model may be trained by updating the weighting factors and evaluating the results (e.g., based on a loss function). The process may be iterated until the desired results are achieved. During training the weighting factors may be updated according to a predetermined algorithm (e.g., gradient descent). For accuracy, a cross-entropy loss may be minimized on the training set” (Xu at 0045) and “compute a predicted probability for the labeled sequence dataset using a softmax layer that divides the exponential of each of the one or more prediction values by the sum of the one or more prediction values. A diversity regularization value may also be applied to the one or more prototypes to penalize at least a first of the one or more prototype vectors that is similar to a second of the one or more prototype vectors. A clustering regularization function may also be applied to the one or more labeled datasets and the one or more prototype vectors to ensure a clustering structure in a latent space. An evidence regularization function may also be applied to ensure the one or more prototype vectors are approximately equal to the one or more labeled sequence datasets” (Xu at 0148), where “The disclosed model can present verifiable and understandable prototypes that are very useful in the healthcare domain” (Xu at 0076). Where Xu discloses the use of similar mathematical formulas or algorithms (see Xu at least at 0040-0058), although the formulas being claimed do not appear to be explicitly disclosed, it would appear obvious to use the claimed formulas as expressions of the indicated terms as a form of “Use of Known Technique To Improve Similar Devices (Methods, or Products) in the Same Way” (as per MPEP 2143(I)(C)) since the base device, method, or product would be comparable and applying the claimed formulas would be applied in the same way with the same level of predictability.
Therefore, the Examiner understands and finds that to minimize a loss function including a learning term and clustering, segment, and diversity regularization terms is applying a known technique to a known device, method, or product ready for improvement to yield predictable results so as to present verifiable and understandable information, and the claimed formula expression would appear obvious to use as a form of “Use of Known Technique To Improve Similar Devices (Methods, or Products) in the Same Way” (as per MPEP 2143(I)(C)) since the base device, method, or product would be comparable and applying the claimed formulas would be applied in the same way with the same level of predictability.
The rationale for combining in this manner is that to minimize a loss function including a learning term and clustering, segment, and diversity regularization terms is applying a known technique to a known device, method, or product ready for improvement to yield predictable results so as to present verifiable and understandable information and the claimed formula expression would appear obvious to use as a form of “Use of Known Technique To Improve Similar Devices (Methods, or Products) in the Same Way” (as per MPEP 2143(I)(C)) since the base device, method, or product would be comparable and applying the claimed formulas would be applied in the same way with the same level of predictability as explained above.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine or modify the modeling of Waner with the learning and regularization of Xu and use the formula expressions claimed in order to minimize a loss function including a learning term and clustering, segment, and diversity regularization terms so as to present verifiable and understandable information
Claim 2: Waner in view of Xu discloses the method of claim 1, further comprising embedding the segmented patient trajectory using a segment embedding layer of the machine learning model (Waner at 0115, “Machine learning algorithms described herein can comprise methods of dimensionality reduction, including but not limited to … t-distributed stochastic neighbor embedding, and/or variants thereof and/or combinations thereof”).
Claim 3: Waner in view of Xu discloses the method of claim 2, wherein the segment embedding layer includes a multilayer perceptron and a one-dimensional convolutional layer (Waner at 0118, “The neural networks can comprise feed-forward and/or feed-back connectivity between “neurons” and/or layers thereof.… Neural networks can comprise, without limitation, … feedforward artificial neural network models, multilayer perceptrons, recurrent neural networks, In some instances, restricted Boltzmann machines, self-organizing maps, or self-organizing feature maps, convolutional neural networks, and/or variants thereof and/or combinations thereof”, 0119, “Machine learning algorithms described herein can comprise deep learning methods including but not limited to … convolutional neural networks, convolutional deep belief networks, … and/or variants thereof and/or combinations thereof”).
Claim 4: Waner in view of Xu discloses the method of claim 1, further comprising:
measuring a patient’s state information (Waner at 0082, “record the parameters associated with given stimulation pulses over time, and this data, along with physiological data collected from the sensing component 50, may be used by a clinician to titrate the system or to otherwise analyze a sleep state of the patient. The collected data may also be used to generate patient alarms and reports on patient treatment and status”);
selecting a treatment action based on a skill predicted by the trained model, based on the measured state information (Waner at 0082, “the system can “learn” patient specific patterns of sleep and patient specific patterns of apnea, e.g., via machine learning algorithms, which can enable the system to predict when an event is likely to occur and enable the system to calibrate and select to what level to activate the device”); and
notifying a medical professional of the treatment action to assist the medical professional in decision-making for patient management (Waner at 0082, “the collected data may be used to identify irregularities in the sleep patterns and if appropriate take action, e.g., send an alert for help. The data collected over time can be useful to identify problems early on, e.g., worsening breathing patterns, worsening sleeping problems, etc. The data can then be considered by the treating healthcare professional and/or automatically assessed by the processor”).
Claim 5: Waner in view of Xu discloses the method of claim 1, wherein the skill embedding includes a weighted combination of the prototype vectors based on how similar the prototype vectors are to the segmented patient trajectory (Waner at 0046, “a plurality of vital signs are collected, the data may be index or a weight index may be generated, and that index or weighted index may be used by the computer systems, described further herein, to determine whether a sleep apnea event has occurred or is likely to occur (e.g. by comparing a computer or derived index or weighted index to a pre-determined threshold index or a pre-determined threshold weighted index, each specific to the particular patient being treated)”).
Claim 8: Waner discloses the method of claim 1, but does not appear to explicitly disclose
wherein the clustering structure regularization term is expressed as:
PNG
media_image2.png
63
221
media_image2.png
Greyscale
the prototype segment evidence regularization term is expressed as:
PNG
media_image3.png
65
232
media_image3.png
Greyscale
and the diversity regularization term is expressed as:
PNG
media_image4.png
65
328
media_image4.png
Greyscale
where n is a number of segments, k is a number of prototype vectors, pi is an ith prototype vector, zt(j) is a segment embedding at step t for segment j, and dmin is a proximity threshold.
However, where Waner in view of Xu discloses the use of similar mathematical formulas or algorithms (see Xu at least at 0040-0058), although the formulas being claimed do not appear to be explicitly disclosed, it would appear obvious to use the claimed formulas as expressions of the indicated terms as a form of “Use of Known Technique To Improve Similar Devices (Methods, or Products) in the Same Way” (as per MPEP 2143(I)(C)) since the base device, method, or product would be comparable and applying the claimed formulas would be applied in the same way with the same level of predictability.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine or modify the formulas in Zu as combined at Waner in view of Xu with the formulas indicated in order to use the claimed formulas as expressions of the indicated terms.
The rationale for combining in this manner is that to use the claimed formulas as expressions of the indicated terms is a form of using a known technique to improve similar devices, methods, or products in the same way since the base device, method, or product would be comparable and applying the claimed formulas would be applied in the same way with the same level of predictability as explained above.
Claims 11-13, 15, and 18 are rejected on the same basis as claims 1-3 and 5, above since Waner discloses a system for training a healthcare treatment machine learning model, comprising: a hardware processor; and a memory that stores a computer program which, when executed by the hardware processor, causes the hardware processor to perform the same or similar activities as at claims 1-3 and 5 above (see Waner at least at 0009-0010).
Claim 14: Waner in view of Xu discloses the system of claim 11, wherein the prototype layer determines a similarity between the segments and the prototype vectors (Waner at 0109-0113, indicating various forms or types of clustering, including BIRCH, hierarchical, k-means, EM, and/or DBSCAN)
Claim 20: Waner in view of Xu discloses the method of claim 1, but does not appear to explicitly disclose wherein the treatment action includes an instruction to a treatment system to automatically administer treatment to a patient (Waner at 0005, "one aspect of the present disclosure is a sleep apnea treatment system comprising means for detecting an apnea event and means for stimulating a hypoglossal nerve or geniohyoid muscle in response to the detected apnea event").
Claims 9 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Waner in view of Xu and in further view of Bostic et al. (U.S. Patent Application Publication No. 2020/0303047, hereinafter Bostic) .
Claims 9 and 19: Waner in view of Xu discloses the method and system of claims 1 and 11, but does not appear to explicitly disclose wherein the treatment action includes at least one of a prescription plan, a meal provision plan, a rehabilitation plan, and a discharge destination plan. Waner discloses a treatment plan (see Waner at 0005, “one aspect of the present disclosure is a sleep apnea treatment system comprising means for detecting an apnea event and means for stimulating a hypoglossal nerve or geniohyoid muscle in response to the detected apnea event”), Bostic teaches using machine learning modules to simulate health states of a patient “according to potentially prescribing heart disease medication to the patient and advising that the patient take a small dose of aspirin regularly. For example, the digital twin module 1302 may simulate a health state of the patient according to a regimented exercise and diet plan and the patient continuing a current exercise and diet plan thereof. For example, the digital twin module 1302 may simulate a future health state of a patient having a tumor according to whether the tumor becomes cancerous and whether the tumor remains benign. The previous examples are intended to be non-limiting and illustrate a portion of a potential scope of simulations that the one or more digital twin module 1302 modules may perform. In some embodiments, the digital twin module 1302 may simulate a future health state of the patient based on a plurality of variables, such as by simulating a health state of a patient according to a combination of time frames, treatment plans, prescription drug schedules, and lifestyle changes” (Bostic at 0182, see also Bostic at 0183), “for detecting and addressing situations involving improper prescription of medication, improper utilization of prescribed medications, and diversion of prescribed medications to unprescribed uses” (Bostic at 0006), including controlled medication overuse or underuse (Bostic at 0008-0009). Therefore, the Examiner understands and finds that a treatment plan for prescriptions, meal provision, rehabilitation, and/or discharge destination is applying a known technique to a known device, method, or product ready for improvement to yield predictable results so as to detect and address medication overuse and/or underuse.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine or modify the treatments of Waner in view of Xu with the treatment plans of Bostic in order to provide a treatment plan for prescriptions, meal provision, rehabilitation, and/or discharge destination so as to detect and address medication overuse and/or underuse.
The rationale for combining in this manner is that a treatment plan for prescriptions, meal provision, rehabilitation, and/or discharge destination is applying a known technique to a known device, method, or product ready for improvement to yield predictable results so as to detect and address medication overuse and/or underuse as explained above.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Waner in view of Xu and in further view of Flieg et al. (U.S. Patent Application Publication No. 2015/0273127, hereinafter Flieg).
Claim 10: Waner in view of Xu discloses the method of claim 1, but does not appear to explicitly disclose wherein the treatment action includes an instruction to a treatment system to automatically administer an updated dialysis treatment to a patient by updating operational parameters of a dialysis machine in response to an undesirable albumin reading for the patient. Flieg, however, teaches that “The half-life of albumin is inversely proportional to the plasma albumin concentration, that is, a decreased albumin content results in increased half-life, whereas increasing albumin concentrations cause the metabolic rate to increase by up to 50% (Boldt, Br. J. Anaesth. (2010) 104 (3): 276-284). Therefore, a substitution of the albumin which may be adsorbed by the adsorbent during the treatment with a liver support system according to the invention may not be necessary. However, substitution of albumin may be indicated especially in cases of spontaneous bacterial peritonitis (SBP), hepatorenal syndrome (HRS), and post-paracentesis syndrome (PPS) due to the fact that the liver is severely compromised. Substitution can be done according to the state of the art, mostly by infusion. Therefore, according to one aspect of the invention, liver support or dialysis treatment according to the invention may be followed by the substitution of albumin which was adsorbed during the treatment in order to maintain a serum albumin level of above 30 g/l” (Flieg at 0071 – indicating that where prior treatments had not maintained the albumin level desired, the substitution accomplished, or would accomplish, the desired result). Therefore, the Examiner understands and finds that a dialysis substitution based on albumin levels is applying a known technique to a known device, method, or product ready for improvement to yield predictable results so as to maintain a desired albumin level.
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine or modify the treatments of Waner in view of Xu with the dialysis parameters of Flieg in order to use a dialysis substitution based on albumin levels so as to maintain a desired albumin level.
The rationale for combining in this manner is that a dialysis substitution based on albumin levels is applying a known technique to a known device, method, or product ready for improvement to yield predictable results so as to maintain a desired albumin level as explained above.
Response to Arguments
Applicant's arguments filed 5 March 2026 have been fully considered but they are not persuasive.
Applicant first argues the 101 rejection (Remarks at 8-11), first alleging that “MPEP 2106.06(a)(1)(vii) states that training a machine learning model such as neural networks are not directed to an abstract idea” (Id. at 9). The Examiner notes that there is no “2106.06(a)(1)” (or subsequent sections) in the MPEP. The Examiner believes Applicant intended to refer to MPEP 2106.04(a)(1)(vii). However, that indication does not recite a precedent, but appears based on Example 39, where the claim is NOT just directed to training a model – it includes collecting facial images and transforming those images by mirroring, smoothing, or contrast reduction to create data for a training set. The 4 August 2025 Memo from Charles Kim (“SUBJECT: Reminders on evaluating subject matter eligibility of claims under 35 U.S.C. 101”) clarifies that “Even though ‘training the neural network’ involves a broad array of techniques and/or activities that may involve or rely upon mathematical concepts, the limitation does not set forth or describe any mathematical relationships, calculations, formulas, or equations using words or mathematical symbols” at Example 39. This is then contrasted with Example 47, claim 2, that “requires specific mathematical calculations by referring to the mathematical calculations by name”, which is indicated as ineligible. Applicant’s claims are far more similar to the Example 47, claim 2, recitations in that Applicant claims the specific factors used and the particular formula for training the machine learning model.
Applicant then argues “The present embodiments improve functionality of neural networks” (Remarks at 9); however, neural networks are not recited at the claims. Therefore, the argument is not commensurate with the scope of the claims.
Applicant then argues that “similar to Desjardins, the present embodiments improve the functionality of neural networks by resolving a technical problem rooted in computer technology” (Id.), allegedly since “[t]he present embodiments resolve the issue with interpretability of machine learning models such as neural networks, by utilizing at least regularizing components including terms relating to the clustering structure of segment embedding, the segment-prototype evidence, and the diversity of prototypes” (Id. at 10). However, neural networks are not claimed, and the rest of the argued aspects of the claims is/are the particular formula and/or parts of the formula being claimed. The specification portion cited by Applicant indicates that “the underlying rationales for those historical actions may not be interpretable”; however, the claim only requires a “minimizing a loss function that includes a clustering structure regularization term, a prototype segment evidence regularization term, a diversity regularization term, and an imitation learning term” as expressed by the claimed formula that is comprised of “a length of a segment”, “a number of segments”, “an expert policy”, “an action performed … for [the or each] segment”, “a patient state … for [the or each] segment”, “a learned policy”, and “a skill embedding … for [the or each] segment”. In other words, as best understood in light of the specification, the “underlying rationale” for each action (apparently why an action was taken or performed) is accounted for (i.e., “interpreted”) by including patient state information in addition to the policy used and actions performed. But that is merely designating the data to be used for training the model that is claimed as having the mathematical formula or expression indicated at the claims. This is not improving the training of a model to solve a machine learning problem that would be similar to Desjardins – this is merely designating the data used for the model, and citing a particular formula for the model.
Applicant then argues claim 10 (Remarks at 10); however, as indicated and explained at the 112 rejections, the written description lacks support for claim 10 as amended, and it is not clear whether the treatment action is for training or some form of expected result. The Examiner further notes, though, that the “treatment action” at claim 10 is merely an instruction to a system or device outside the scope of the claims that may or may not actually receive, perform, or disregard the instructions produced by claim 10. The machine learning model indicated at parent claim 1 is never recited as actually used – it is merely trained.
The Examiner notes that the indefiniteness of claim 10 makes it difficult to understand or address any eligibility related to claim 10. Applicant, though, indicates at the argument that claim 10 is allegedly eligible since claim 10 is “to effect a particular treatment or prophylaxis for a disease or medical condition [that] integrates the exception into a practical application” (Remarks at 11). However, MPEP § 2106.04(d)(2) indicates that “If the limitation does not actually provide a treatment or prophylaxis, e.g., it is merely an intended use of the claimed invention or a field of use limitation, then it cannot integrate a judicial exception under the ‘treatment or prophylaxis’ consideration.” As such, per the argument, the “treatment action” at claim 10 is merely an instruction to a system outside the scope of the instant claims, and not actually providing at treatment or prophylaxis, and therefore, is not considered to be a practical application at Step 2A, Prong 2.
The Examiner notes that the instant claims appear exceedingly similar to Maucorps; although Maucorps was to determine a number of visits to a customer (or potential customer), the claims in Maucorps indicated the factors or data to be used, and included the formula or expression those factors were used so as to arrive at the desired answer. In the same manner, the instant claims indicate the factors or data to be used, and include the formula or expression those factors are to be used in so as to arrive at the desired answer – it is just that the desired outcome here is a treatment action rather than a number of visits.
Applicant then argues the prior art rejections (Remarks at 11-14), first reciting the wherein clause and formula/expression at the second element of claim 1 (Id. at 11), alleging that “Waner and/or Xu, taken singly or in combination, fail to disclose or suggest” that wherein clause and formula/expression (Id.), that “Waner and Xu are both silent on the concept of an ‘imitation learning term’, let alone” the wherein clause and formula/expression (Id.), and therefore, “taken singly or in combination, fail to disclose or suggest at least” the wherein clause and formula/expression (Id.). Applicant does not provide any actual reasoning as to why or how the recitations at the rejection do not disclose “an ‘imitation learning term’, let alone” the wherein clause and formula/expression – the Examiner notes that an imitation learning term is apparently any term (i.e., a word or phrase or data/information) from which the model would learn so as to imitate the training data. The recitations at the rejection reflect this.
Applicant then asserts that the Examiner did not “establish a prima facie case of obviousness” since “the Examiner has failed to establish that the cited references teach or suggest at least ‘an imitation learning term’ and the equation for the imitation learning term”. This is apparently based on the allegation that the “Examiner alleged that the claim formulas are a form of known technique despite failing to provide any documentary evidence that would support this assertion” (Remarks at 13 for all). However, the “known technique” is the technique (including mathematics) as indicated at the reference (Xu), not the claimed formulas – the claimed formulas are the “Similar Devices (Methods, or Products)” being improved per the rejection. This is not, then, any indication of official notice – it is citing a reference for what the reference indicates as known.
Therefore, the Examiner is not persuaded by Applicant’s arguments.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Mokhtarani, Shabnam, Embeddings in Machine Learning: Everything You Need to Know, dated 26 August 2021, downloaded 23 June 2025 from https://www.featureform.com/post/the-definitive-guide-to-embeddings, indicating “Embeddings are dense numerical representations of real-world objects and relationships, expressed as a vector” (at p. 4).
Gazzotti et al., Extending electronic medical records vector models with knowledge graphs to improve hospitalization prediction. J Biomed Semant 13, 6 (2022). https://doi.org/10.1186/s13326-022-00261-9, Published 22 February 2022, downloaded from https://link.springer.com/article/10.1186/s13326-022-00261-9 on 8 December 2025, indicating “Artificial intelligence methods applied to electronic medical records (EMRs) hold the potential to help physicians save time by sharpening their analysis and decisions, thereby improving the health of patients. On the one hand, machine learning algorithms have proven their effectiveness in extracting information and exploiting knowledge extracted from data. On the other hand, knowledge graphs capture human knowledge by relying on conceptual schemas and formalization and supporting reasoning. Leveraging knowledge graphs that are legion in the medical field, it is possible to pre-process and enrich data representation used by machine learning algorithms. Medical data standardization is an opportunity to jointly exploit the richness of knowledge graphs and the capabilities of machine learning algorithms” (at Abstract).
“Regularization.” Merriam-Webster.com Dictionary, Merriam-Webster, https://www.merriam-webster.com/dictionary/regularization. Accessed 2 Jun. 2026, indicating “regularization” is “the act or instance of regularizing”.
“Regularize.” Merriam-Webster.com Dictionary, Merriam-Webster, https://www.merriam-webster.com/dictionary/regularize. Accessed 2 Jun. 2026, indicating that “regularizing” is “to make regular by conformance to law, rules, or custom”.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SCOTT D GARTLAND whose telephone number is (571)270-5501. The examiner can normally be reached M-F 8:30 AM - 5 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Kambiz Abdi can be reached at 571-272-6702. 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.
/SCOTT D GARTLAND/
Primary Examiner, Art Unit 3685
Prosecution Insights