DETAILED ACTION
Remarks
Claims 1-4, 6-15 and 17-22 have been examined and rejected. This Office Action is responsive to the amendment filed on 03/23/2026, which has been entered in the above identified application.
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 .
Claims 1-4, 6-15 and 17-22 are presented for examination.
Response to Amendment
Applicant’s amendment filed on 03/23/2026 has been entered. Claims 1-3, 12-14, 20 and 21 are amended. Claims 1-4, 6-15 and 17-22 are pending in the application.
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 1-4, 6-15 and 17-22 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Independent claims
Step 1
Claim 1 is drawn to a computer-implemented method, claim 12 is drawn to a system and claim 20 is drawn to a non-transitory machine-readable medium storing instructions executed by processors to cause the computer to perform the method of claim 1. Therefore, each of these claim groups falls under one of four categories of statutory subject matter (process/method, machines/product/apparatus, manufactures, and composition of matter).
Step 2A – Prong 1
Claims 1, 12 and 20 are directed to a judicially recognized exception of an abstract idea without significantly more.
Claims 1, 12 and 20 recite a method of training a first neural network with a first machine learning algorithm using training data, the first neural network being a recurrent neural network, the training data including a plurality of reference career trajectories, each reference career trajectory in the plurality of reference career trajectories comprising a sequence of reference career segments, each reference career segment in the sequence of reference career segments comprising reference profile data and reference time data indicating a position of the reference career segment within the sequence of reference career segments, the training data also including a corresponding set of reference skills for each reference career segment generated by a second neural network, wherein the second neural network is different from the first neural network that under its broadest reasonable interpretation enumerates a mental concept. A human can mentally perform, with the physical aid such as pen and paper, to train a neural network with machine learning algorithm using training data. Therefore, the step of training a first neural network with a first machine learning algorithm using training data is nothing more than an abstract mental concept (MPEP 2106.04(a)(2)(III)).
Claims 1, 12 and 20 recite further a method of an encoder network that includes a first set of gated recurrent units (GRUs) generating embedding vectors for career segments and a decoder network that includes a second set of GRUs generating embedding vectors for skills that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform the calculation using words or using mathematical symbols to generate embedding vectors. Therefore, the step of generating embedding vectors is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Claims 1, 12 and 20 recite further a method of obtaining a target career trajectory of a target user of an online service, the target career trajectory comprising a sequence of target career segments, each target career segment in the sequence of target career segments comprising target profile data of the target user and target time data indicating a position of the target career segment within the sequence of target career segments that under its broadest reasonable interpretation enumerates a mental concept. A human can mentally perform, with the physical aid such as pen and paper, to obtain a career trajectory of a target user. Therefore, the step of obtaining a target career trajectory of a target user of an online service is nothing more than an abstract mental concept (MPEP 2106.04(a)(2)(III)).
Claims 1, 12 and 20 recite further a method of for each target skill in a set of target skills, applying the target career trajectory to the trained first neural network to obtain a corresponding score for the respective target skill that under its broadest reasonable interpretation enumerates a mathematical concept. A human can perform the calculation using words or using mathematical symbols to compute a score for a target skill. Therefore, the step of computing a corresponding score is nothing more than a mathematical concept (MPEP 2106.04(a)(2)(I)).
Step 2A – Prong 2
Claims 1, 12 and 20 recite further the additional step of using the corresponding scores for the set of target skills in an application of the online service that fail to integrate the abstract idea into a practical application. The step of using the corresponding scores for target skills is form of insignificant input and output solution activities, where using the corresponding scores for the set of target skills is necessary for all uses of the judicial exception. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Claims 1, 12 and 20 recite further the additional steps of using the first set of GRUs to output a compressed vector representation of the sequence of target career segments of the target career trajectory; and inputting the compressed vector representation into an individual GRU configured to generate a first embedding vector of a set of embedding vectors representing the set of target skills those fail to integrate the abstract idea into a practical application. The steps of outputting and inputting the compressed vector representation is a form of insignificant input and output solution activities, where outputting and inputting the compressed vector representation into an individual GRU is necessary for all uses of the judicial exception. This additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Step 2B
The additional element in step 2A-Prong 2 that is form of insignificant extra-solution activities, do not amount to significantly more than an abstract idea because the court decisions have determined that this additional element of using the corresponding scores for the set of target skills; and outputting and inputting the compressed vector representation into an individual GRU to be well-understood, routine, and conventional when claimed in a merely generic manner (MPEP 2106.05(d)(II)).
As such, claims 1, 12 and 20 are not patent eligible.
Dependent claims
Claims 2-4, 6-11, 13-15, 17-19, 21 and 22 merely narrow the previously recited abstract idea limitations. For the reasons described above with respect to claims 1, 12 and 20, this judicial exception is not meaningfully integrated into a practical application, or significantly more than the abstract idea. The claims disclose similar limitations described for the independent claims above and do not provide anything more than the mental processes that are practically capable
of being performed in the human mind with the assistance of pen and paper and mathematical
concepts that are achievable through mathematical computation. Therefore, claims 2-11, 13-19, 21 and 22 also recite abstract ideas that do not integrate into a practical application or amount to significantly more than the judicial exception, and are rejected under U.S.C. 101
Step 1
Claim 2-11 are drawn to a computer-implemented method, claims 13-19 are drawn to a system, and claims 21 and 22 are drawn to a non-transitory machine-readable medium storing instructions executed by processors that cause the computer to perform the method of claims 2-11. Therefore, each of these claim groups falls under one of four categories of statutory subject matter (process/method, machines/product/apparatus, manufactures, and composition of matter).
Step 2A – Prong 1
Dependent claims 2 and 13 recite the mental processes by training the second neural network, with a second machine learning algorithm, to output a probability distribution of skills based on an input career segment comprising input profile data, wherein the second neural network is not a recurrent neural network; and using the second neural network to select the corresponding set of reference skills for each reference career segment in the training data; wherein the first set of GRUs and the second set of GRUs are configured to generate the embedding vectors by performing linear transformations of corresponding column vectors to a transformation matrix; and wherein the column vectors are sparse and the transformation matrix projects sparse dimensions into a dense vector of reduced dimensionality those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Dependent claims 3, 14 and 21 recite further the mental process by receiving, by the encoder network, embedding vectors representing the sequence of target career segments that is based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Dependent claims 4, 15 and 22 recite further the mental processes by using the second set of GRUs to generate a second embedding vector representing the set of target skills; passing the first embedding vector and the second embedding vector to a softmax layer; and generating, using the softmax layer, a corresponding probability distribution of skills for each target career segment, wherein the corresponding score is generated based on the corresponding probability distribution of skills those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Dependent claims 8 and 17 recite the mental processes by selecting one or more target skills from the set of target skills based on the corresponding scores of the one or more target skills; and displaying, on a display of a computing device of the target user, a corresponding selectable user interface button for each one of the selected one or more target skills on a computing device of the target user, the corresponding selectable user interface button being configured to trigger storing of the corresponding target skill as part of a profile of the target user in response to a selection of the corresponding selectable user interface button, the profile being stored on the online service those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Dependent claims 9 and 18 recite the mathematical and mental processes by receiving a search query submitted by the target user via a computing device of the target user; computing a corresponding score for each job posting in a plurality of job postings based on the search query and the corresponding scores for the set of target skills; selecting one or more job postings from the plurality of job postings based on the corresponding scores for the one or more job postings; and displaying the selected one or more job postings as search results for the search query on the computing device of the target user those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Dependent claims 10 and 19 recite the mathematical and mental processes by computing a corresponding score for each job posting in a plurality of job postings based on the corresponding scores for the set of target skills; selecting one or more job postings from the plurality of job postings based on the corresponding scores for the one or more job postings; and displaying the selected one or more job postings on a computing device of the target user those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Dependent claim 11 recites the mathematical and mental processes by computing a corresponding score for each online course in a plurality of online courses based on the corresponding scores for the set of target skills; selecting one or more online courses from the plurality of online courses based on the corresponding scores for the one or more online courses; and displaying the selected one or more online courses on a computing device of the target user those are based on one or more features of the ML project (MPEP 2106.04(a)(2)(III)).
Step 2A – Prong 2
Dependent claim 6 recites further the insignificant extra solution activities by the reference profile data comprises one or more profile data comprising: a job title, a seniority level, a company, an industry, or a description of the corresponding reference career segment. These additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
Dependent claim 7 recites further the insignificant extra solution activities by the target profile data comprises one or more profile data comprising: a job title, a seniority level, a company, an industry, or a description of the corresponding target career segment. These additional elements do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea (MPEP 2106.05(g)).
As such, dependent claims 2-4, 6-11, 13-15, 17-19, 21 and 22 are not patent eligible.
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 1-4, 6-15 and 17-22 are rejected under 35 U.S.C. 103 as being unpatentable over Nigam et al (US 20210374681 A1) hereafter Nigam, in view of Malur Srinivasan et al (US 20220300736 A1) hereafter Malur, further in view of Lee et al (US 20220319500 A1) hereafter Lee, and further in view of Ahmed et al (US 20160260064 A1) hereafter Ahmed.
With respect to claim 1, Nigam teaches a computer-implemented method performed by a computer system having a memory and at least one hardware processor (the system comprises a CPU or a processor for executing program components, wherein the processor may be disposed in communication with a memory RAM or ROM [par. 0033-0037]), the computer-implemented method comprising:
training a first neural network with a first machine learning algorithm using training data, the first neural network being a recurrent neural network (tree-based and deep neural networks are used to generate a set of job recommendations. Gated Recurrent Unit (GRU) is used to filter jobs and provide a first set of job recommendations. The GRU technique is applied on the sequence of interaction data vector comprising of the latent competency vector and other feature vector to generate a first set of job recommendations for users [par. 0004, 0009, 0032, 0062]), the training data including a plurality of reference career trajectories, each reference career trajectory in the plurality of reference career trajectories comprising a sequence of reference career segments, each reference career segment in the sequence of reference career segments comprising reference profile data and reference time data indicating a position of the reference career segment within the sequence of reference career segments (the training data comprises a set of sample job postings search results selected by users along with the information (time, location, etc. ) about the people who performed the corresponding searches. The recommendation system comprises a communication interface which users can interact with and can see their job recommendations and user profiles through the feature module. By using GRU, the model is based on predefined time stamps, latent skills and preferences hidden in user interactions [par. 0005, 0009, 0010, 0063-0066]);
obtaining a target career trajectory of a target user of an online service, the target career trajectory comprising a sequence of target career segments, each target career segment in the sequence of target career segments comprising target profile data of the target user and target time data indicating a position of the target career segment within the sequence of target career segments (the recommendation system may predict next career move of a candidate by understanding the current career trajectory of the candidate. By elaborating with the user profile, job industry, location, educational requirement and skills, a job filter may be used to provide some relevant jobs. The system may be based on a time-dependent log of past user activity to capture the user interests to predict future user actions [par. 0005, 0009, 0029]).
However, Nigam does not particularly disclose an encoder network that includes a first set of gated recurrent units (GRUs) generating embedding vectors for career segments and a decoder network that includes a second set of GRUs generating embedding vectors for skills; the training data also including a corresponding set of reference skills for each reference career segment generated by a second neural network, wherein the second neural network is different from the first neural network; for each target skill in a set of target skills, applying the target career trajectory to the trained first neural network to obtain a corresponding score for the respective target skill; and using the corresponding scores for the set of target skills in an application of the online service; wherein the method further includes: using the first set of GRUs to output a compressed vector representation of the sequence of target career segments of the target career trajectory; and inputting the compressed vector representation into an individual GRU configured to generate a first embedding vector of a set of embedding vectors representing the set of target skills.
In the same field of endeavor, Malur teaches the training data also including a corresponding set of reference skills for each reference career segment generated by a second neural network, wherein the second neural network is different from the first neural network (a method for an automated empathetic assessment of a candidate for a job is disclosed. First information about a first set of features associated with a first candidate may be extracted from a resume or profile information. Second information about a second set of features may correspond to the first set of features may be extracted from the databases. A third set of features may be determined based on a difference of corresponding features from the first set of features and the second set of features. A neural network model may determine an empathy score for each of a set of candidates. The method may determine a candidate for a job based on a difference of qualifications or skillsets of the candidate. Examples of the neural network model may include deep neural network, convolutional neural network, artificial neural network, recurrent neural network, long short-term memory network, etc. The neural network model used in here is not necessarily a recurrent neural network. The training data used herein is associated with a candidate. Such data includes candidate information associated with the candidate whose empathy score is to be determined [par. 0020-0022, 0052, 0053, 0059-0061]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated the concept of extracting information of a candidate to be determined based on the information of features extracted from databases as suggested by Malur into the process of recommending jobs based on skills as suggested by Nigam because both of these systems addressing the process of recommending jobs/positions for potential candidates based on their matched skillsets. Doing so would be desirable because the system of Nigam would be more efficient by extracting first information about a first set of features associated with a first candidate from a document or profile information and extracting second information about second set of features those may be corresponding to the first set of features from one or more databases (Malur, [par. 0004-0006]).
However, the combination of Nigam and Malur does not disclose an encoder network that includes a first set of gated recurrent units (GRUs) generating embedding vectors for career segments and a decoder network that includes a second set of GRUs generating embedding vectors for skills; for each target skill in a set of target skills, applying the target career trajectory to the trained first neural network to obtain a corresponding score for the respective target skill; and using the corresponding scores for the set of target skills in an application of the online service; wherein the method further includes: using the first set of GRUs to output a compressed vector representation of the sequence of target career segments of the target career trajectory; and inputting the compressed vector representation into an individual GRU configured to generate a first embedding vector of a set of embedding vectors representing the set of target skills.
In the same field of endeavor, Lee teaches an encoder network that includes a first set of gated recurrent units (GRUs) generating embedding vectors for career segments and a decoder network that includes a second set of GRUs generating embedding vectors for skills (a model is designed to output text data based on a voice signal obtained from the user (ASR model). The ASR model includes an encoder and a decoder, wherein the encoder may convert input data into an embedding vector using a learning model which has been trained. Embedding vectors are then outputted after the learning model is updated. The decoder may be either a DNN implemented with a fully-connected layer or a RNN implemented with a gated recurrent unit [par. 0073-0076]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated the concept of controlling an automatic speech recognition (ASR) model and a language model by an electronic device as suggested by Lee into the combination of Nigam and Malur because all of these systems addressing recommending jobs/positions for potential users/candidates based on their matched skillsets and other matched profile information. Doing so would be desirable because the combination of Nigam and Malur would be more efficient by applying a language model which comprises an input embedding layer, a positional encoding layer, one sub-network layer, a linearization layer and a soft max layer, such that the method may enter input data to generate embedding vectors in the ASR model which includes an encoder and a decoder (Lee, [par. 0008, 0009, 0073]).
However, the combination of Nigam, Malur and Lee does not disclose for each target skill in a set of target skills, applying the target career trajectory to the trained first neural network to obtain a corresponding score for the respective target skill; and using the corresponding scores for the set of target skills in an application of the online service; wherein the method further includes: using the first set of GRUs to output a compressed vector representation of the sequence of target career segments of the target career trajectory; and inputting the compressed vector representation into an individual GRU configured to generate a first embedding vector of a set of embedding vectors representing the set of target skills.
In the same field of endeavor, Ahmed teaches for each target skill in a set of target skills, applying the target career trajectory to the trained first neural network to obtain a corresponding score for the respective target skill (A score may be calculated and assigned to a corresponding skill in a list of skills, whereas a weighted sum may be determined for each skill for each geographic location. A flowchart is also provided to display a method of ranking jobs based on the course history of a user. A weighted score for a course that corresponding to a preferred skill may be calculated which may be used to rank jobs by suitability [par. 0041, 0107-0109 and FIG. 1B]); and
using the corresponding scores for the set of target skills in an application of the online service (The scores calculated from the corresponding jobs and the corresponding courses may be used to display on the talent profile of an online service [par. 0028, 0074]);
wherein the method further includes: using the first set of GRUs to output a compressed vector representation of the sequence of target career segments of the target career trajectory (a first set of job recommendations is computed using GRU model, such that GRU is applied on a set of filtered job. The domain vector is constructed using a plurality of domain skills and the user vector [par. 0009, 0010]); and
inputting the compressed vector representation into an individual GRU configured to generate a first embedding vector of a set of embedding vectors representing the set of target skills (the domain skills and the user skills are obtained through the feature module, such that the job recommendations are obtained using GRU model. User skills or job requirements are represented by a vector where each dimension represents a latent competency group [par. 0009, 0010, 0058-0061]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the invention to have incorporated the concept of calculating corresponding scores for preferred jobs and preferred courses as suggested by Ahmed into the combination of Nigam, Malur and Lee because all of these systems addressing the process of recommending a career path for users based on their profiles, whereas the profiles may show user job interests, location, skills, etc. Doing so would be desirable because the combination of Nigam, Malur and Lee would be more efficient by calculating score for a corresponding skill to determine which skill is trending in the market that can match up with a user who is looking for a job with the respective skill (Ahmed, [par. 0004, 0107-0109]).
With respect to claim 2, the combination of Nigam, Malur, Lee and Ahmed teaches training the second neural network, with a second machine learning algorithm, to output a probability distribution of skills based on an input career segment comprising input profile data, wherein the second neural network is not a recurrent neural network (Malur, the method may determine a candidate for a job based on a difference of qualifications or skillsets of the candidate. Examples of the neural network model may include deep neural network, convolutional neural network, artificial neural network, recurrent neural network, long short-term memory network, etc. The neural network model used in here is not necessarily a recurrent neural network. The training data used herein is associated with a candidate. Such data includes candidate information associated with the candidate whose empathy score is to be determined [par. 0020-0022, 0052, 0053, 0059-0061]); and using the second neural network to select the corresponding set of reference skills for each reference career segment in the training data (Malur, the system may determine a difference between a first set of features of the candidate and corresponding features in a second set of features of a population of candidates belonging to a similar demographic background as the candidate [par. 0020-0022]).
However, the combination of Nigam, Malur, Lee and Ahmed may not teach wherein the first set of GRUs and the second set of GRUs are configured to generate the embedding vectors by performing linear transformations of corresponding column vectors to a transformation matrix; and wherein the column vectors are sparse and the transformation matrix projects sparse dimensions into a dense vector of reduced dimensionality.
With respect to claim 3, the combination of Nigam, Malur, Lee and Ahmed teaches receiving, by the encoder network, embedding vectors representing the sequence of target career segments (Ahmed, an interaction data vector is computed by a plurality of interaction data including interests, selection, preferences, job rejections, of the users, through the communication interface [par. 0009, 0010]).
With respect to claim 4, the combination of Nigam, Malur, Lee and Ahmed teaches using the second set of GRUs to generate a second embedding vector representing the set of target skills (Nigam, the ML model is based on GRU wherein at predefined time stamps, latent skills and/or preferences hidden in user interactions are captured from input in form of sequential data [par. 0006, 0007, 0066]);
passing the first embedding vector and the second embedding vector to a softmax layer (Nigam, A user profile may be analyzed by LSTM model. The job recommendation system may integrate ML models for capturing user’s interests by employing content-based or collaborative filtering techniques to provide a suitable job recommendation list to the user. The position-specific information is used as input to LSTM layer and is linked with embeddings of the company-specific factors and is fed as input to another LSTM layer [par. 0004, 0005, 0031 and FIG. 5]); and
generating, using the softmax layer, a corresponding probability distribution of skills for each target career segment, wherein the corresponding score is generated based on the corresponding probability distribution of skills (Nigam, latent skills and/or preferences are used to predict probability of next job recommendations as well as skills in the sequence [par. 0066]).
With respect to claim 6, the combination of Nigam, Malur, Lee and Ahmed teaches the reference profile data comprises one or more profile data comprising: a job title, a seniority level, a company, an industry, or a description of the corresponding reference career segment (Ahmed, the social profile of the user may comprise job history, certifications, internships, educational history, field of study, degrees, skills, interests, geographic location, etc. [par. 0025]).
With respect to claim 7, the combination of Nigam, Malur, Lee and Ahmed teaches wherein the target profile data comprises one or more profile data comprising: a job title, a seniority level, a company, an industry, or a description of the corresponding target career segment (Ahmed, the talent profile may be extracted that includes most of the fields of the social profile, such as job history, skills, interests, etc. [par. 0025, 0029]).
With respect to claim 8, the combination of Nigam, Malur, Lee and Ahmed teaches wherein the using the corresponding scores for the set of target skills in the application of the online service comprises:
selecting one or more target skills from the set of target skills based on the corresponding scores of the one or more target skills (Ahmed, the method of ranking potential jobs comprises a step of adding a weighted score for a course associated with a preferred skill which can be used to rank jobs based on the user’s skills, and the career server may then select available jobs based on those preferred skills [par. 0041, 0044]); and
displaying on a display of a computing device of the target user, a corresponding selectable user interface button for each one of the selected one or more target skills on a computing device of the target user (Ahmed, the server may display highest-ranking jobs based on the calculated scores. Besides, the trending skills system may also display the most in-demand skills in the job market based on the total score. Each skill may be linked to a course, wherein a link is attached to a button as the trending skills may rearrange [par. 0034, 0100, 0101, 0110]), the corresponding selectable user interface button being configured to trigger storing of the corresponding target skill as part of a profile of the target user in response to a selection of the corresponding selectable user interface button, the profile being stored on the online service (Ahmed, the profiler configured to process information from the career server to create a user profile that displays unlock courses or a global score of the user. The profiler may be a library, API or a set of scripts. The skills those identified based on course information may be added to the profile of user by the profiler. The identified skill may be all kind of skill from computer programming language skill to marketing skill [par. 0113, 0114, 0117]).
With respect to claim 9, the combination of Nigam, Malur, Lee and Ahmed teaches wherein the using the corresponding scores for the set of target skills in the application of the online service comprises:
receiving a search query submitted by the target user via a computing device of the target user (Ahmed, the server may search for jobs and associated skills. The server may search for an instance based on a geographic location. The server may search the first website for a first number instances that the skill is found. The API may perform a blanket search within the first website for the skills [par. 0103]);
computing a corresponding score for each job posting in a plurality of job postings based on the search query and the corresponding scores for the set of target skills (Ahmed, a score is calculated for each job in a plurality of jobs, and one or more jobs may be selected with the highest scores. A weighted score may be added to a preferred skill to rank available jobs [par. 0034, 0041, 0063]);
selecting one or more job postings from the plurality of job postings based on the corresponding scores for the one or more job postings (Ahmed, a career server may add points to a preferred skills score for a job. A list of highest ranking jobs may be generated based on the preferred skills score [par. 0058-0063]); and
displaying the selected one or more job postings as search results for the search query on the computing device of the target user (Ahmed, the selected jobs may be displayed on multiple platforms based on the search query, such as a specific geographic location. For example, the API does a blanket search on country specific job feeds over the internet. The server may display a list of websites based on this country-specific job search on the platforms. An interface provided on a computing device may contain different options to view related jobs [par. 0102-0104, 0125 and FIG. 4]).
With respect to claim 10, the combination of Nigam, Malur, Lee and Ahmed teaches wherein the using the corresponding scores for the set of target skills in the application of the online service comprises:
computing a corresponding score for each job posting in a plurality of job postings based on the corresponding scores for the set of target skills (Ahmed, the server may compute a score for each job in the plurality of jobs. A preferred skills score may be assigned to a job to rank a list of jobs by the career server [par. 0034, 0058-0063]);
selecting one or more job postings from the plurality of job postings based on the corresponding scores for the one or more job postings (Ahmed, a career server may add points to a preferred skills score for a job. A list of highest ranking jobs may be generated based on the preferred skills score [par. 0058-0063]); and
displaying the selected one or more job postings on a computing device of the target user (Ahmed, the selected jobs may be displayed on multiple platforms based on the search query, such as a specific geographic location. For example, the API does a blanket search on country specific job feeds over the internet. The server may display a list of websites based on this country-specific job search on the platforms. An interface provided on a computing device may contain different options to view related jobs [par. 0102-0104, 0125 and FIG. 4]).
With respect to claim 11, the combination of Nigam, Malur, Lee and Ahmed teaches wherein the using the corresponding scores for the set of target skills in the application of the online service comprises:
computing a corresponding score for each online course in a plurality of online courses based on the corresponding scores for the set of target skills (Ahmed, the server may also compute and add a weighted score for each course in a plurality of previous courses that associated with a preferred skill to calculate a preferred skills score, which can then be used to rank jobs based on the user’s skills. The preferred skills score may also be used to determine a course type, such as introductory level course or a higher level course [par. 0041, 0057]);
selecting one or more online courses from the plurality of online courses based on the corresponding scores for the one or more online courses (Ahmed, the career server comprises a database of available courses. The profiler which updates the user profile may dynamically update the global score to unlock more courses [par. 0114, 0115 and FIG. 3B]); and
displaying the selected one or more online courses on a computing device of the target user (Ahmed, the career server may comprise an analogous course finder which may be a library, an API or a set of scripts used to obtain courses and corresponding predetermined courses. An interface may be provided on a computing device to display a list of available courses [par. 0117, 0124 and FIG. 4]).
With respect to claim 12, it is a system claim that corresponding to the computer-implemented method of claim 1. Therefore, it is rejected for the same reason as claimed in claim 1 above.
With respect to claim 13, it is a system claim that corresponding to the computer-implemented method of claim 2. Therefore, it is rejected for the same reason as claimed in claim 2 above.
With respect to claim 14, it is a system claim that corresponding to the computer-implemented method of claim 3. Therefore, it is rejected for the same reason as claimed in claim 3 above.
With respect to claim 15, it is a system claim that corresponding to the computer-implemented method of claim 4. Therefore, it is rejected for the same reason as claimed in claim 4 above.
With respect to claim 17, it is a system claim that corresponding to the computer-implemented method of claim 8. Therefore, it is rejected for the same reason as claimed in claim 8 above.
With respect to claim 18, it is a system claim that corresponding to the computer-implemented method of claim 9. Therefore, it is rejected for the same reason as claimed in claim 9 above.
With respect to claim 19, it is a system claim that corresponding to the computer-implemented method of claim 10. Therefore, it is rejected for the same reason as claimed in claim 10 above.
With respect to claim 20, it is a non-transitory machine-readable medium claim that corresponding to the computer-implemented method of claim 1. Therefore, it is rejected for the same reason as claimed in claim 1 above.
With respect to claim 21, it is a non-transitory machine-readable medium claim that corresponding to the computer-implemented method of claim 3. Therefore, it is rejected for the same reason as claimed in claim 3 above.
With respect to claim 22, it is a non-transitory machine-readable medium claim that corresponding to the computer-implemented method of claim 4. Therefore, it is rejected for the same reason as claimed in claim 4 above.
Response to Arguments
The examiner respectfully acknowledges the applicant’s amendments to claims 1-3, 12-14, 20 and 21.
Applicant’s arguments filed on 03/23/2026 regarding claims 1-4, 6-15 and 17-22 under 35 U.S.C. 101 have been fully considered but are not persuasive.
Applicant argued that “A GRU is not something that a human can implement in their mind. The other steps in the claim are similarly complex enough such that the human mind cannot perform them as a metal process. For example, claim 1 recites "the training data also including a corresponding set of reference skills for each reference career segment generated by a second neural network." The human mind cannot partition itself into two neural networks, where one neural network generates training data to train another a neural network to predict an output.”
Examiner respectfully partially disagrees.
As clearly stated in the Office Action, Examiner indicated that the limitation “training a first neural network … for each reference career segment generated by a second neural network” recites a mental process, and the limitation “an encoder network that includes a first set of gated recurrent unit (GRUs) … embedding vectors for skills” recites a mathematical concept. The first limitation may not recite the mental process. However, the second limitation clearly involves the mathematical concept, as GRUs operates through mathematical operations on vectors and matrices. At high level, a GRU computes gating functions and hidden states using some specific equations and other similar update equations involving matrix multiplication, addition, and nonlinear activation functions. From the perspective of Examiner, the amended claim recites applying a GRU to input data to generate embeddings and scores. These operations amount to mathematical calculations performed on numerical data.
Applicant argued that “Applicant respectfully submits that the pending claims recite an improvement in a computer-related technology. Namely, as set forth in paragraph [0023] of the pending application, training the first neural network using the reference career trajectories improves the effectiveness, efficiency, and scalability of the computer system that is being used to identify skills for users. By using the trained first neural network model instead of a heuristic-based approach, the computer system is able to accommodate the significant variance in the characteristics of the users in a more scalable way. The computer system is also able to adapt to real-world changes more quickly than the heuristic-based approach. The extremely complex encoding of the heuristic-based decision-making process is avoided, and the variance in characteristics of both users and skills is more effectively processed, resulting in a faster and more accurate computer system.
Moreover, the amended features "using the first set of GRUs to output a compressed vector representation of the sequence of target career segments of the target career trajectory," and "inputting the compressed vector representation into an individual GRU configured to generate a first embedding vector of a set of embedding vectors representing the set of target skills" reflect additional improvements to computer related technology. Specifically, the use of a first set of GRUs to output a compressed vector representation serves as a specialized technical tool for dimensionality reduction. This configuration allows the system to capture long-term dependencies within a career trajectory while reducing the memory overhead and computational cycles that typically plague the processing of complex sequences. Rather than a generic computer function, this represents a hardware-integrated logic for filtering noise and retaining relevant historical data.”
Examiner respectfully disagrees.
The newly added limitations merely further define the mathematical model already present in the claim. These limitations constitute additional mathematical processing performed by the recited neural network and do not integrate the judicial exception into a practical application. The additional GRU limitations do not improve computer functionality, network performance, or machine learning training itself. Instead, they merely describe how the abstract analysis is performed. The additional elements merely use generic machine-learning components as tools to perform the abstract idea. The amended claim remains directed to analyzing career-trajectory data and generating skill-related outputs using mathematical models. Encoder GRUs, decoder GRUs, embeddings and compressed vector representation were known machine learning techniques. The claim clearly implements an abstract idea using well-understood, routine, and conventional computer components. In general, the GRU limitations do not provide any concrete technological improvement.
Applicant argued that “As indicated above, conventional neural networks suffer from a technical problem of insufficient training data. The lack of training data results in a neural network generating less accurate predictions. The claimed subject matter solves this technical problem by using a second neural network to generate training data for the first neural network. Specifically, claim 1 recites "the training data also including a corresponding set of reference skills for each reference career segment generated by a second neural network." This additional feature helps to integrate the alleged abstract idea into a practical application by solving the technical issues presented by conventional neural networks.”
Examiner respectfully disagrees.
The amended claim recites training a neural network using career trajectories and associated skills, receiving a user trajectory, generating scores, and generating embedding vectors. The amended claim clearly does not recite using a second neural network to generate training data for a first neural network, such that preventing from insufficient training data and improving prediction accuracy. Accordingly, the alleged improvement is not reflected in the claim language and cannot establish that the claim is directed to a technological improvement.
Therefore, the amended claim recites mathematical concepts and machine-learning operations. The claim does not integrate those concepts into a practical application because it does not recite an improvement to computer technology or machine-learning technology.
Amended claim 1 and its corresponding claims 12 and 20 are not patent eligible for at least the reasons list above. Dependent claims 2-4, 6-11, 13-15, 17-19, 21 and 22, those either directly or indirectly depended on claims 1, 12 and 20, are not patent-eligible for the same reasons.
Applicant’s arguments filed on 03/23/2026 regarding claims 1-4, 6-15 and 17-22 under 35 U.S.C. 103 have been fully considered and moot in view of new ground of rejection (see rejection above).
Conclusion
Applicant’s amendment necessitated the new grounds of rejection presented in this Office Action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP 706.07(a). Applicant is remined of the extension of time policy as set forth in 37 CFR 1.136(a).
Tellex et al (US 20200023514 A1) disclosed a method includes enabling a robot to learn a mapping between English language commands and Linear Temporal Logic (LTL) expressions, wherein neural sequence-to-sequence learning models are employed to infer a LTL sequence corresponding to a given natural language command.
Xie et al (US 20190303798 A1) disclosed profile and/or usage data of a social networking service is leveraged to automatically generate potential career paths for users of the social networking service. Additionally, specific recommendations as to actions the users can take to increase their odds of progressing along particular career paths can be determined, and these recommendations can be shared with users. Both recommendations may be performed in a manner that is scalable for personalized service.
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 filled within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Quoc Phung whose telephone number is (703) 756 1330. The examiner can normally be reached on Monday through Friday from 9am to 5pm PT.
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, Jennifer Welch can be reached on 571-272-7212. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/Q.L.P./Examiner, Art Unit 2143
/BEAU D SPRATT/Primary Examiner, Art Unit 2143
Prosecution Insights