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 .
Status of the Claims
Claims 1-20 are currently pending.
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 a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Step 1
Claims 1-20 are within the four statutory categories. Claims 1-12 are drawn to methods for evaluating and optimizing treatment plans, which is within the four statutory categories (i.e. process). Claims 13-20 are drawn to a system for evaluating and optimizing treatment plans, which is within the four statutory categories (i.e. machine).
Prong 1 of Step 2A
Claim 9, which is representative of the inventive concept, recites: A computer-implemented method for optimizing healthcare outcomes using adversarial imitation deep learning, comprising:
receiving patient data from one or more medical sensors monitoring a patient;
processing the patient data to remove irrelevant data based on correlation to a healthcare key performance indicator (KPI);
generating, using a policy generator network with a transformer-based architecture, an optimal sequence of treatment actions based on the patient data;
employing a discriminator network to differentiate between the optimal sequence of treatment actions and real-world high-performance treatment sequences, wherein the discriminator network generates a first reward signal based on divergence from the real-world high-performance treatment sequences;
estimating final healthcare KPI results based on the optimal sequence of treatment actions using a performance prediction network, wherein the performance prediction network generates a second reward signal based on estimated KPI improvement at each time step;
generating a composite loss function by combining the first reward signal and the second reward signal, and updating the policy generator network based on the composite loss function; and
applying the optimal sequence of treatment actions to a patient's care plan to optimize healthcare KPI in real-time, wherein the optimal sequence of treatment actions is generated by the policy generator network trained using the composite loss function.
The underlined limitations as shown above, given the broadest reasonable interpretation, cover the abstract idea of a mathematical concept and/or a certain method of organizing human activity because they recite mathematical relationships, formulas, equations, and/or mathematical calculations (in this case, the step of generating of a composite loss function by combining the first and second reward signals recites at least a mathematical relationship and/or calculation), and/or managing personal behavior or relationships or interactions between people (i.e. social activities, teaching, and following rules or instructions – in this case, the steps of receiving patient data, processing the patient data to remove irrelevant data, generating an optimal sequence of treatment actions based on the patient data, differentiating between the optimal sequence of treatment actions and real-world high-performance treatment sequences, estimating final healthcare KPI results based on the optimal sequence of treatment actions, and applying the optimal sequence of treatment actions to a patient’s care plan recite following rules or instructions to optimize behavior and/or actions in the form of the determined patient care plan), e.g. see MPEP 2106.04(a)(2). Any limitations not identified above as part of the abstract idea are deemed “additional elements,” and will be discussed in further detail below.
Furthermore, the abstract idea for Claims 1 and 13 is identical as the abstract idea for Claim 9, because the only difference between Claims 1, 9, and 13 is that Claim 9 recites a method, whereas Claim 1 recites a method but does not recite that the sensors are medical sensors, that the sensor data is patient data, and/or that the sequence of actions are treatment actions, and Claim 13 recites a system that executes the same functions as those claimed in Claim 9.
Dependent Claims 2-8, 10-12, and 14-20 include other limitations, for example Claims 2, 12, and 14 recite a mechanism to capture temporal dependencies in the sensor data, Claims 3 and 15 recite minimizing discrepancies between action sequences and real-world high-performance sequences, Claims 4 and 16 recite a mechanism to estimate the final KPI results, Claims 5-6, 11, and 17-18 recite performing simulations on the data to determine future states of the process and consequences of potential actions, Claims 7, 10, and 19 recite utilizing historical data, Claim 8 recites that the sequence of actions comprise patient treatment actions, and Claim 20 recites types of medical interventions for the optimal sequence, but these only serve to further narrow the abstract idea, and a claim may not preempt abstract ideas, even if the judicial exception is narrow, e.g. see MPEP 2106.04, and/or do not further narrow the abstract idea and instead only recite additional elements, which will be further addressed below. Hence dependent Claims 2-8, 10-12, and 14-20 are nonetheless directed towards fundamentally the same abstract idea as independent Claims 1, 9, and 13.
Hence Claims 1-20 are directed towards the aforementioned abstract idea.
Prong 2 of Step 2A
Claims 1, 9, and 13 are not integrated into a practical application because the additional elements (i.e. the non-underlined limitations above – in this case, the medical sensors, the policy generator network with a transformer-based architecture, the discriminator network, the performance prediction network, the hardware processor and memory, the steps of generating the first reward signal and the second reward signal, and the fact that the optimal sequence of treatment actions is generated by the policy generator network trained using the composite loss function) amount to no more than limitations which:
amount to mere instructions to apply an exception – for example, the recitation of the medical sensors, the hardware processor, the memory, and the various types of machine-learning architecture (i.e. the policy generator network with a transformer-based architecture, the discriminator network, the performance prediction network, the steps of generating the first reward signal and the second reward signal, and the fact that the optimal sequence of treatment actions is generated by the policy generator network trained using the composite loss function), which amounts to merely invoking a computer as a tool to perform the abstract idea, e.g. see [0033]-[0035], [0041]-[0044], [0048]-[0051], [0054], and [0057] of the as-filed Specification, and see MPEP 2106.05(f); and/or
generally link the abstract idea to a particular technological environment or field of use – for example, the claim language of the data being patient data, the sensors being medical sensors, and the sequence of actions being patient treatment actions, which amounts to limiting the abstract idea to the field of healthcare, e.g. see MPEP 2106.05(h).
Additionally, dependent Claims 2-8, 10-12, and 14-20 include other limitations, but these limitations also amount to no more than mere instructions to apply an exception (e.g. the various machine learning limitations recited of dependent Claims 2-4, 7, 10, 12, 14-16, and 19), and/or do not include any additional elements beyond those already recited in independent Claims 1, 9, and 13, and hence also do not integrate the aforementioned abstract idea into a practical application.
Hence Claims 1-20 do not include additional elements that integrate the judicial exception into a practical application.
Step 2B
Claims 1, 9, and 13 do not include additional elements that are sufficient to amount to “significantly more” than the judicial exception because the additional elements (i.e. the non-underlined limitations above – in this case, the medical sensors, the policy generator network with a transformer-based architecture, the discriminator network, the performance prediction network, the hardware processor and memory, the steps of generating the first reward signal and the second reward signal, and the fact that the optimal sequence of treatment actions is generated by the policy generator network trained using the composite loss function), as stated above, are directed towards no more than limitations that amount to mere instructions to apply the exception, generally link the abstract idea to a particular technological environment or field of use, and/or add insignificant extra-solution activity to the abstract idea, wherein the additional elements comprise limitations which:
amount to elements that have been recognized as well-understood, routine, and conventional activity in particular fields, as demonstrated by:
The present Specification expressly disclosing that the structural additional elements are well-understood, routine, and conventional in nature:
[0033]-[0035], [0041]-[0044], [0048]-[0051], [0054], and [0057] of the as-filed Specification discloses that the additional elements (i.e. the medical sensors, the policy generator network with a transformer-based architecture, the discriminator network, the performance prediction network, the hardware processor and memory, and the fact that the optimal sequence of treatment actions is generated by the policy generator network trained using the composite loss function) comprise a plurality of different types of generic computing systems;
Relevant court decisions: The functional limitations interpreted as additional elements are analogized to the following examples of court decisions demonstrating well-understood, routine and conventional activities, e.g. see MPEP 2106.05(d)(II):
Storing and retrieving information in memory, e.g. see Versata Dev. Group, Inc. v. SAP Am., Inc. – similarly, the additional elements recite receiving sensor data and real-world high performance sequence data, and utilizing (i.e. retrieving) the received data in order to ultimately estimate the final KPI results and apply the optimal sequence to a plan to optimize the KPI; and/or
Performing repetitive calculations, e.g. see Parker v. Flook, and/or Bancorp Services v. Sun Life – similarly, the additional elements recite performing basic calculations (i.e. the calculations of the first and second reward signals are recited at a high level, are used to determine the composite loss function, wherein the composite loss function is used as a basis for the updating of the policy generator network) and does not impose meaningful limits on the scope of the claims;
Dependent Claims 2-8, 10-12, and 14-20 include other limitations, but none of these limitations are deemed significantly more than the abstract idea because the additional elements recited in the aforementioned dependent claims similarly amount to mere instructions to apply an exception (e.g. the various machine learning limitations recited independent Claims 2-4, 7, 10, 12, 14-16, and 19), and/or do not include any additional elements beyond those already recited in independent Claims 1, 9, and 13, and hence do not amount to “significantly more” than the abstract idea.
Hence, Claims 1-20 do not include any additional elements that amount to “significantly more” than the judicial exception.
Thus, taken alone, the additional elements do not amount to significantly more than the abstract idea identified above. Furthermore, looking at the limitations as an ordered combination adds nothing that is not already present when looking at the elements taken individually, and there is no indication that the combination of elements improves the functioning of a computer or improves any other technology, and their collective functions merely provide conventional computer implementation.
Therefore, whether taken individually or as an ordered combination, Claims 1-20 are nonetheless rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter.
Subject Matter Free From Prior Art
Claims 1-20 are not presently rejected under 35 U.S.C. 102 or 103, and hence would be in condition for allowance if amended to overcome the rejections presented under 35 U.S.C. 101. The following represents Examiner’s characterization of the most relevant prior art references and the differences between the present claim language and the prior art references in view of 35 U.S.C. 102 and/or 103:
With regards to 35 U.S.C. 102 and/or 103, the following represents the closest prior art to the claimed invention, as well as the differences between the prior art and the limitations of the presently claimed invention.
White (US 2020/0342969) teaches a system that calculates a compliance score based on a comparison between an ideal treatment plan and an actual treatment plan, wherein the compliance score reflects how similar the treatments of the actual treatment plan are to the ideal treatment plan. Additionally, White teaches comparing the compliance score to a threshold, and when the threshold is not satisfied, determines and performs remedial actions. However, White does not teach receiving data from sensors, and removing sensor data determined to be irrelevant, and further does not teach utilizing a transformer network in any of the compliance score calculations. Additionally, White does not teach generating a first reward signal based on the divergence between the ideal treatment plan, generating a second reward signal based on estimated KPI improvement at each time step, generating a composite loss function based on the first and second reward signals, updating the policy generator network based on the composite loss function, and further does not teach that the optimal sequence of actions is generated by the policy generator network trained using the composite loss function.
Cheng (US 2018/0264258) teaches removing a noise signal (i.e. irrelevant data) in order to evaluate the efficacy of the treatment without the possible corruption that the noise signal may cause. However, Cheng does not teach determining any type of compliance or KPI score to determine an optimal sequence of actions, and does not specifically teach making this determination based on a comparison between the optimal sequence and a real-world sequence. Furthermore, Cheng does not teach generating a first reward signal based on the divergence between the ideal treatment plan, generating a second reward signal based on estimated KPI improvement at each time step, generating a composite loss function based on the first and second reward signals, updating the policy generator network based on the composite loss function, and further does not teach that the optimal sequence of actions is generated by the policy generator network trained using the composite loss function.
Inam (US 2021/0338387) teaches determining review metrics that correspond to analysis results of the necessity or justifiability of certain clinical treatment procedures, wherein the calculation of the review metrics is performed through machine learning implementations including neural networks such as a transformer network. Additionally, although, Inam teaches utilizing the neural network to provide optimal results for patients, for example by optimizing measurement speed of patient data, Inam does not specifically teach determining an optimal sequence of actions. Additionally, Inam does not teach generating a first reward signal based on the divergence between the ideal treatment plan, generating a second reward signal based on estimated KPI improvement at each time step, generating a composite loss function based on the first and second reward signals, updating the policy generator network based on the composite loss function, and further does not teach that the optimal sequence of actions is generated by the policy generator network trained using the composite loss function.
Kuusela (US 2018/0161596) teaches determining a treatment plan for a patient based on a model that is trained using a set of treatment plans previously devised for past patients that are of high quality, and further teaches improving a dosing model for a patient. However, Kuusela does not specifically teach determining an optimal sequence of actions, and further does not teach generating a first reward signal based on the divergence between the ideal treatment plan, generating a second reward signal based on estimated KPI improvement at each time step, generating a composite loss function based on the first and second reward signals, updating the policy generator network based on the composite loss function, and further does not teach that the optimal sequence of actions is generated by the policy generator network trained using the composite loss function.
Parker (US 2022/0361779) teaches computing similarity scores between a patient’s glucose values and a sequence of historic glucose values, wherein the similarity score enables the facilitating of diabetes coaching, linking of events, and providing recommendations. However, Parker does not teach the specifics of the diabetes coaching, linking of events, and recommendations. That is, Parker does not teach determining an optimal sequence of actions, generating a first reward signal based on the divergence between the ideal treatment plan, generating a second reward signal based on estimated KPI improvement at each time step, generating a composite loss function based on the first and second reward signals, updating the policy generator network based on the composite loss function, and further does not teach that the optimal sequence of actions is generated by the policy generator network trained using the composite loss function.
The aforementioned references are understood to be the closest prior art. Various aspects of the claimed invention are known individually, but for the reasons disclosed above, the particular manner in which the elements of the present invention are claimed, when considered as an ordered combination, distinguishes from the aforementioned references and hence the invention recited in Claim 1-20 is not considered to be disclosed by and/or obvious in view of the inventions of the closest prior art references.
Response to Arguments
Applicant’s arguments, see Remarks, filed April 2, 2026, with respect to the rejections of Claims 1-20 under 35 U.S.C. 101 have been fully considered but are not persuasive.
Applicants first allege that the claimed invention is patent eligible because it integrates any abstract idea into a practical application, by virtue of reciting an improvement in the technological field of industrial operations, e.g. see pgs. 8-11 of Remarks – Examiner disagrees.
The Specification discloses that the claimed invention addresses the problems of executing actions in an optimal sequence to achieve the best KPI, and clarity regarding the effects/rewards of actions on the final KPI, e.g. see [0004]-[0005] of the as-filed Specification. The aforementioned problems are not specifically technological problems because they have existed since long before the advent of any computer technology, and moreover exist independent of any technology involved in the actions. That is, the problem of “optimizing workflow” exists across every process, and thus is not a problem arising specifically from any technology involved in the process. Hence, even assuming, arguendo, that the claimed invention achieves the allege improvements (i.e. optimized sequence of actions, high value KPIs), these nonetheless do not address technological problems and further represent improvements to the abstract idea rather than technological improvements, and an improvement in the abstract idea itself (in this case, a certain method of organizing human activities) is not an improvement in technology, e.g. see MPEP 2106.05(a)(II).
For the aforementioned reasons, Claims 1-20 are rejected under 35 U.S.C. 101.
Applicant’s arguments, see Remarks, filed April 2, 2026, with respect to the rejections of Claims 1-20 under 35 U.S.C. 103 have been fully considered and, in combination with the claim amendments, are persuasive. The rejections of Claims 1-20 under 35 U.S.C. 103 have been withdrawn for the reasons disclosed above.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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 JOHN P GO whose telephone number is (703)756-1965. The examiner can normally be reached Monday-Friday 9am-6pm Pacific.
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, PETER H CHOI can be reached at (469)295-9171. 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.
/JOHN P GO/Primary Examiner, Art Unit 3681
Prosecution Insights