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 .
DETAILED ACTION
This non-final action is responsive to application filed on 12/30/2024. Claims 1-14 are pending, with claims 1 and 14 being independent.
Priority
The present patent application claims priority from Singapore Patent Application Number 10202404092X filed on December 27, 2024.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 12/30/2024 and 01/28/2026 is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claims 1-14 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention. The claims recite “acquiring, from the online platforms, the given in-use message” (emphasis added). The disclosure does not provide any details regarding how one message is acquired via multiple online platforms.
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-14 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor, or for pre-AIA the applicant regards as the invention.
Claim 1 recite “acquiring, from the online platforms, the given in-use message” (emphasis added). It is not clear how one message is acquired via multiple online platforms. Examiner assumes that the given in-use message is acquired from an online platform.
Claims 2-13 depend on claim 1and are rejected for the same reasons as set forth in claim 1.
Claim 14 has the same issue as shown in claim 1, and is ejected for the same reasons as set forth in claim 1.
In addition, claim 3 recites “wherein the generating the respective message vector comprises: replacing values of service fields of the given training message, hyperlinks, and emails with a respective predetermined value” (emphasis added). It is not clear how replacing values of service fields of hyperlinks, and emails with a respective predetermined value are part of the generating the respective message vector. It appears that the claim should read “wherein the generating the respective message vector comprises: replacing values of service fields of the given training message
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 of this title, 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, 2, 9, 13 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Tristan et al. (US 2019/0095805, published Mar. 28, 2019) and Prakash (US 2016/0014151, published Jan. 14, 2016).
As per claim 1, Tristan discloses a computer-implemented method for identifying target messages (Tristan Fig. 1 and par. 64, SPAM email classification), a target message including a malicious ad (from claim 2, target message includes spam message; Tristan par. 53, a SPAM filtering system may receive a decision from a document classifier system, which classifies an incoming email as SPAM, the method comprising:
during a first phase (Tristan Fig. 7, process of training an ensembled decision system):
acquiring a plurality of training messages (Tristan par. 64, in a SPAM email classification example, the training data may be divided by the sender's email address, so that certain decision models may be trained to specialize on emails from particular senders. [This indicates that the training data comprising email messages]);
generating, for a given training message of the plurality of training messages, a respective message vector (Tristan par. 64, in a SPAM email classification example, the training data may be divided by the sender's email address, so that certain decision models may be trained to specialize on emails from particular senders; Tristan Fig. 7, Perform the hashing technique on the input record to generate a feature vector at 708);
generating a training set of data including a plurality of training digital objects, a given one of which includes: (i) the respective message vector of the given training message; and (ii) a respective label representative of the given training message being one selected from the group consisting of: a target message; and a non-target message (Tristan par. 46, at least some of the models may employ different hashing techniques, so that they are trained to make their decisions using different feature vectors; Tristan par. 65, In some cases of supervised learning, the training may involve repeatedly executing the decision models on the training data sets, comparing the decision results against predetermined truth labels);
feeding, to a given prediction model of a plurality prediction models, the given training digital object, thereby training the given prediction model (Tristan par. 46, at least some of the models may employ different hashing techniques, so that they are trained to make their decisions using different feature vectors; Tristan par. 65, In some cases of supervised learning, the training may involve repeatedly executing the decision models on the training data sets, comparing the decision results against predetermined truth labels) to generate a respective prediction of whether a given in-use message is a target one or not (Tristan Fig. 1 and par. 64, SPAM email classification; Tristan par. 18, the decision system 100 may be a regression system, in which case the decision may span a continuous range (e.g., a probability range));
generating, based on respective predictions of the plurality prediction models, a respective training consolidated probability vector for the given training message of the plurality of training messages (Tristan par. 81, after the training of the individual models in the ensemble is completed, a further training may be performed on the combining model to achieve the desired accuracy for the overall model. This training may be performed by using the trained models in the ensemble to generate test results using additional training data, feeding those results to the combining model, and then adjusting the parameters of the combining model based the accuracy of its decisions);
using respective training consolidated probability vectors associated with the plurality of training messages, training a decision tree model (Tristan par. 81, after the training of the individual models in the ensemble is completed, a further training may be performed on the combining model to achieve the desired accuracy for the overall model. This training may be performed by using the trained models in the ensemble to generate test results using additional training data, feeding those results to the combining model, and then adjusting the parameters of the combining model based the accuracy of its decisions) to determine whether the given in-use message is a target one or not (Tristan par. 64, SPAM email classification;
during a second phase, following the first phase (Tristan Fig. 3, process of decision making performed by an ensembled decision system):
acquiring, from the online platforms, the given in-use message (Tristan Fig. 3, Receive input data to an ensembled decision model including multiple decision models at 302);
generating, for the given in-use message, a respective in-use message vector (Tristan Fig. 3, Perform a respective hashing technique for each decision model to reduce features of the input data to a feature vector at 304);
feeding, the respective in-use message vector, to each prediction model of the plurality trained models, thereby causing each one of the plurality prediction models to generate a respective probability value of the given in-use message being a target message (Tristan Fig. 3, Generate respective decision results from each decision model based on the respective feature vector at 306; Tristan par. 18, the decision system 100 may be a regression system, in which case the decision may span a continuous range (e.g., a probability range));
generating, based on the respective probability values of the plurality prediction models, an in-use consolidated probability vector (Tristan par. 46, the contributing models may each produce a confidence metric that indicates a confidence level associated with its output result. Accordingly, the confidence metrics are provided to the combining function or model, which may be trained to use these metrics to make its decision);
feeding the in-use consolidated probability vector to the decision tree model, thereby causing the decision tree model to generate a final probability value of the given in-use message being a target message (Tristan par. 46, the contributing models may each produce a confidence metric that indicates a confidence level associated with its output result. Accordingly, the confidence metrics are provided to the combining function or model, which may be trained to use these metrics to make its decision; see Tristan Fig. 2C, decision errors);
in response to the final probability value being representative of the given in-use message being a target message, causing execution of a remedial action (Tristan par. 53, a SPAM filtering system may receive a decision from a document classifier system, which classifies an incoming email as SPAM. The SPAM filtering system may then immediately place the incoming email in a SPAM folder of the receiving user).
Note, Tristan teaches acquiring a plurality of training messages, but does not explicitly disclose acquiring from online platforms (i.e., acquiring, from online platforms, a plurality of training messages).
Prakash teaches:
acquiring, from online platforms, a plurality of training messages (Prakash par. 136, Messages written by John Smith would then be retrieved from the online social networks and the classification engine would be trained on these retrieved messages; Prakash par. 58, Examples of online social networks include but are not limited to Facebook, Linkedin, Google+, Twitter, alunmi associations, Instagram, Reddit, Pinterest, Vine, Tumblr, Nexopia, Badoo, Myspace, Xanga and Friendster).
It would have been obvious to one skilled in the art before the effective filing date of the claimed invention to modify the method of Tristan with the teaching of Prakash for acquiring, from online platforms, a plurality of training messages. One of ordinary skilled in the art would have been motivated because it offers the advantage of providing various sources for acquiring the training messages.
As per claim 2, Tristan-Prakash discloses the method of claim 1, wherein the target message one selected from the group consisting of:
a message about a sale or a purchase of illegal goods and services;
a message advertising selling access to a private network;
a message with proposals of illegal jobs;
a message with proposals to participate in illegal actions;
a message aimed at committing a crime; and
a spam message (Tristan par. 53, a SPAM filtering system may receive a decision from a document classifier system, which classifies an incoming email as SPAM).
As per claim 9, Tristan-Prakash discloses the method of claim 1, wherein each prediction model of the plurality of prediction models has a different architecture (Tristan par. 64, the classifiers may be implemented using different machine learning models, such as, for example, decision trees, support vector machines, neural networks, Bayesian networks, and the like).
As per claim 13, Tristan-Prakash discloses the method of claim 1, wherein the remedial action comprises at least one selected from the group consisting of:
submitting a complaint of an author the given in-use message to a respective customer support service;
generating a warning notification about a cybersecurity incident;
storing information of the given in-use message in a target message database (Tristan par. 53, a SPAM filtering system may receive a decision from a document classifier system, which classifies an incoming email as SPAM. The SPAM filtering system may then immediately place the incoming email in a SPAM folder of the receiving user); and
generating a notification for displaying to an operator.
Claim 14 does not teach or further define over the limitations in claim 1. As such, claim 14 is rejected for the same reasons as set forth in claim 1.
Claims 3-6 are rejected under 35 U.S.C. 103 as being unpatentable over Tristan et al. (US 2019/0095805, published Mar. 28, 2019), Prakash (US 2016/0014151, published Jan. 14, 2016), Danis et al. (US 2011/0225200, published Sep. 15, 2011) and Parrish et al. (US 2022/0382977, published Dec. 1, 2022).
As per claim 3, Tristan-Prakash discloses the method of claim 1, but does not explicitly disclose wherein the generating the respective message vector comprises:
replacing values of service fields of the given training message, hyperlinks, and emails with a respective predetermined value;
tokenizing, comprising bringing all words in the message text to their initial form;
generating a statistical metric representative of a frequency of occurrence of each word.
Danis teaches:
replacing values of service fields of message with a respective predetermined value (Danis par. 15, if the policy provides for anonymity of the senders and receivers, then personally identifying information, such as a name, is replaced with a character string or text such as a pseudonym).
It would have been obvious to one skilled in the art before the effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Danis for replacing values of service fields of the given training message, hyperlinks, and emails with a respective predetermined value. One of ordinary skilled in the art would have been motivated because it offers the advantage of anonymizing personal identifying information for preserving privacy of individuals.
Tristan-Prakash-Danis does not explicitly disclose:
tokenizing, comprising bringing all words in the message text to their initial form;
generating a statistical metric representative of a frequency of occurrence of each word.
Parrish teaches:
tokenizing, comprising bringing all words in the message text to their initial form (Parrish par. 38, pre-processing of the requirements text (which can involve; (i) cleaning, lemmatization, and other standard preprocessing procedures);
generating a statistical metric representative of a frequency of occurrence of each word (Parrish par. 38, vectorising the requirements documents using, for example, TF-IDF and/or BERT).
It would have been obvious to one skilled in the art before the effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Parrish for pre-processing and vectorizing text (e.g., tokenizing, comprising bringing all words in the message text to their initial form; generating a statistical metric representative of a frequency of occurrence of each word). One of ordinary skilled in the art would have been motivated because it offers the advantage of allowing allows machines and AI models to understand language, context, and nuance.
As per claim 4, Tristan-Prakash-Danis-Parrish discloses the method of claim 3, wherein the service fields comprise at least one selected from the group consisting of:
a user identifier of an author (Danis par. 15, if the policy provides for anonymity of the senders and receivers, then personally identifying information, such as a name, is replaced with a character string or text such as a pseudonym) of the given training message (Tristan par. 64, in a SPAM email classification example, the training data may be divided by the sender's email address, so that certain decision models may be trained to specialize on emails from particular senders);
a username of the author of the given training message; and
a password of the author of the given training message.
As per claim 5, Tristan-Prakash-Danis-Parrish discloses the method of claim 3, wherein the generating the statistical metric comprises executing a Term Frequency Inverse Document Frequency (TF/IDF) algorithm (Parrish par. 38, vectorising the requirements documents using, for example, TF-IDF and/or BERT). The same rationale as in claim 3 applies.
As per claim 6, Tristan-Prakash-Danis-Parrish discloses the method of claim 3, wherein the generating the statistical metric comprises executing a Bidirectional Encoder Representations from Transformers (BERT) algorithm (Parrish par. 38, vectorising the requirements documents using, for example, TF-IDF and/or BERT). The same rationale as in claim 3 applies.
Claims 7 and 8 are rejected under 35 U.S.C. 103 as being unpatentable over Tristan et al. (US 2019/0095805, published Mar. 28, 2019), Prakash (US 2016/0014151, published Jan. 14, 2016), Lauwers (US 2024/0248925, published Jul. 25, 2024) and Li (US 9,946,789, published Apr. 17, 2018).
As per claim 7, Tristan-Prakash discloses the method of claim 1, but does not explicitly disclose the method of claim 1, wherein after the generating the respective message vector, the method further comprises:
clustering respective message vectors;
in response to a given cluster including at least one training message that has been assigned the respective label being indicative of the at least one training message being a target one, determining all training messages of the given cluster as being target messages.
Lauwers teaches:
clustering respective vectors (Lauwers par. 29, the vector representations are clustered).
It would have been obvious to one skilled in the art before the effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Lauwers for clustering respective message vectors. One of ordinary skilled in the art would have been motivated because it offers the advantage of accelerating analysis and reducing computational costs.
Tristan-Prakash-Lauwers does not explicitly disclose:
in response to a given cluster including at least one training message that has been assigned the respective label being indicative of the at least one training message being a target one, determining all training messages of the given cluster as being target messages.
Li teaches:
in response to a given cluster including at least one training message that has been assigned the respective label being indicative of the at least one training message being a target one, determining all training messages of the given cluster as being target messages (Li 9:24-27, the classification system 106 may assign a single label to all messages within the same cluster, overriding the existing labels of some messages. This is because a classification model trained or constructed based on messages that are in the same cluster; Li 10:44-46, the message label having the greatest weight within a cluster is assigned to all messages within that cluster).
It would have been obvious to one skilled in the art before the effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Li for in response to a given cluster including at least one training message that has been assigned the respective label being indicative of the at least one training message being a target one, determining all training messages of the given cluster as being target messages. One of ordinary skilled in the art would have been motivated because it offers the advantage of improving classification accuracy.
As per claim 8, Tristan-Prakash-Lauwers-Li discloses the method of claim 7, wherein the clustering the respective message vectors comprises executing a Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm (Lauwers par. 29, A non-limiting example of a suitable clustering is HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise)). The same rationale as in claim 7 applies.
Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Tristan et al. (US 2019/0095805, published Mar. 28, 2019), Prakash (US 2016/0014151, published Jan. 14, 2016) and Bruss et al. (US 2019/0349400, published Nov. 14, 2019).
As per claim 10, Tristan-Prakash discloses the method of claim 9, wherein the plurality models includes:
a logistic regression model (Tristan par. 21, in a decision system that uses a logistic regression for document classification);
a neural network (Tristan par. 64, the classifiers may be implemented using different machine learning models, such as, for example, decision trees, support vector machines, neural networks, Bayesian networks, and the like).
Tristan-Prakash does not explicitly disclose:
a random forest model;
a gradient boosting model.
Bruss teaches:
a random forest model (Bruss par. 49, random-forest classifier);
a gradient boosting model (Bruss par. 49, gradient boosting machine).
It would have been obvious to one skilled in the art at the time of effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Bruss because a simple substitution of one known element (random-forest model and gradient boosting model of Bruss) for another (models of Tristan) would yield the predictable results of analyzing messages using random-forest model and gradient boosting model.
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Tristan et al. (US 2019/0095805, published Mar. 28, 2019), Prakash (US 2016/0014151, published Jan. 14, 2016) and Singh et al. (US 2019/0158070, published May 23, 2019).
As per claim 11, Tristan-Prakash discloses the method of claim 1, wherein the training consolidated probability vector, along with the respective predictions of the plurality of prediction models, further comprises:
an arithmetic mean of the respective predictions (Tristan par. 50, the combining function may implement an averaging of the results of the different models).
Tristan-Prakash does not explicitly disclose:
values of a pairwise summation of the respective predictions;
values of a triple summation of the respective predictions.
Singh teaches:
values of a pairwise summation (Singh pg. 5, the first expression is a double summation performed on M);
values of a triple summation of the respective predictions (Singh pg. 5, the first expression is a triple summation performed on M).
It would have been obvious to one skilled in the art at the time of effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Singh to combine the pairwise and triple summation features for values of a pairwise summation of the respective predictions; values of a triple summation of the respective predictions because in combination, each element merely performs the same function as it does separately and the combination would yield the predictable results of allowing the classification to further performing pairwise summation and triple summation.
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over Tristan et al. (US 2019/0095805, published Mar. 28, 2019), Prakash (US 2016/0014151, published Jan. 14, 2016) and Brabec et al. (US 2024/0356969, filed Jul. 10, 2023).
As per claim 12, Tristan-Prakash discloses the method of claim 1, but does not explicitly disclose the wherein the causing the execution of the remedial action is executed in response to the final probability value exceeding a pre-determined threshold value.
Brabec teaches:
the causing the execution of the remedial action is executed in response to the final probability value exceeding a pre-determined threshold value (Brabec par. 51, a final probability score exceeding the predetermined threshold value may be indicative of a fraudulent email classification; in other words, exceeding the threshold may not longer classify the email as being simply suspicious requiring a blocking or other more restrictive protective action to be warranted).
It would have been obvious to one skilled in the art at the time of effective filing date of the claimed invention to further modify the method of Tristan with the teaching of Brabec for causing the execution of the remedial action is executed in response to the final probability value exceeding a pre-determined threshold value. One of ordinary skilled in the art would have been motivated because it offers the advantage of finetuning the sensitivity of the SPAM email classification.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20250047639 A1; System And Method For Generating A Signature Of A Spam Message Based On Clustering
A method for generating a signature of a spam message includes determining one or more classification attributes and one or more clustering attributes contained in successively intercepted first and second electronic messages. The first electronic message is classified using a trained classification model for classifying electronic messages based on the one or more classification attributes. The first electronic message is classified as spam if a degree of similarity of the first electronic message to one or more spam messages is greater than a predetermined value. A determination is made whether the first electronic message and the second electronic message belong to a single cluster based on the determined one or more clustering attributes. A signature of a spam message is generated based on the the identified single cluster of electronic messages.
US 20170251006 A1; System For Detecting Fraudulent Electronic Communications Impersonation, Insider Threats And Attacks
A system for detecting fraudulent emails from entities impersonating legitimate senders that are intended to cause the recipients to unknowingly conduct unauthorized transactions, for example, transferring funds or divulging sensitive information. The system monitors emails being sent from and received at the protected domain to detect suspected fraudulent emails. The emails are monitored for, among other aspects, linguistic variations, changes in normal patterns of email communications, new or unfamiliar source domains. Suspicious emails can be held and flagged for later review, discarded or passed through with an alert raised indicating a review is needed.
US 12627708 B2; Systems, Methods, And Apparatuses For Detection Of Data Misappropriation Attempts Across Electronic Communication Platforms
Systems, computer program products, and methods are described herein for detection of data misappropriation attempts across electronic communication platforms. The present invention is configured to identify a recipient user account, wherein the recipient user account has received a current communication; parse the current communication to identify at least one order for the recipient user account; identify at least one potential outcome based on the at least one order for the recipient user account; determine the potential outcome comprises a misappropriation; apply a misappropriation attempt engine to the current communication; and generate, by the misappropriation attempt engine, a misappropriation attempt rating for the current communication.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KHANG DO whose telephone number is (571)270-7837. The examiner can normally be reached Monday-Friday 8:00 - 5:00 EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, RUPAL DHARIA can be reached at (571) 272-3880. 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.
/KHANG DO/Primary Examiner, Art Unit 2492
Prosecution Insights