DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant’s arguments with respect to claim(s) are rejected under 103, have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Applicant argued in the remark that “identify a set of frequent user interaction sub-sequences by: mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences; and selecting, based on the vector representation, the corresponding user interaction sub-sequences the set of frequent user interaction sub-sequences,"
Examiner respectfully disagrees. Ludwig US 2012/0280927 discloses identify a set of frequent user interaction sub-sequences by (0048 a sequence of interactions and the context selection determines, i.e. identify, a context mapping and 0167 a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping ): mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences(0167] However, in general many subtasks are performed in sequence, and at least some sort of context switch is involved between each sequence step. For example, FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations. And see also 0260 ); and selecting, based on the vector representation, the corresponding user interaction sub-sequences the set of frequent user interaction sub-sequences,"( [0126] FIG. 2b depicts use of a pointing device (such as a mouse, trackball, or conventional touchpad), first used for selecting focus or context, and secondly directing pointing device output to GUI elements either in the drawing area or GUI "task selection" area. In each area, the traditional GUI provides such a "graphical phrase dictionary" for the application language control and command instructions, sometimes directing in a drawing/viewing area, other times in a "task selection area." And 0167 FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations.)
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 1-4,6-8,10-17 and 19-23 is rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
As per clams 1,11, and 20, the claimed invention is directed to an abstract idea without significantly more. The claim(s) recite(s) identify a plurality of user interaction sequences indicating user interactions representing actions of a plurality of users during use of an application, wherein each user interaction sequence of the plurality of user interaction sequences corresponds to a session of a different user of the plurality of users;
tokenize the plurality of user interaction sequences to generate a vector representation that indicates a quantity of occurrences of each of a plurality of sub-sequences within the plurality of user interaction sequences, wherein the plurality of sub-sequences includes sub-sequences having different lengths; and
identify a set of frequent user interaction sub-sequences by mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences; and selecting, based on the vector representation, the corresponding user interaction sub-sequences as the set of frequent user interaction sub-sequences.
The above limitations are analyzing data to identify patterns — specifically, collecting user interaction data, converting it to a numerical representation, and identifying frequently occurring patterns within that data. This is potentially drawn to noting more than data analysis and pattern recognition concept under mental steps. This claim is likely ineligible under § 101 because it appears directed to an abstract idea of data analysis / pattern identification in user interaction sequences, implemented on generic computer components without a sufficiently particularized technological improvement or other inventive concept. The above judicial exception is not integrated into a practical application because the above group of limitations can be a mental process. The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because the above groups of abstract ideas can be perform mentally. This judicial exception is not integrated into a practical application. In particular, the claim only recites one additional element. The above steps are recited at a high-level of generality (i.e., as a generic processor) such that it amounts no more than mere instructions to apply the exception using a generic computer component. Accordingly, there is not any 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. The claim is directed to an abstract idea.
All dependents’ claims are rejected based on the same rational set forth in the independent claims respectively.
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 1-4, 6-8, 10-17, and 19-23 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.
As per claims 1,11 and 20,
identify a plurality of user interaction sequences indicating user interactions representing actions of a plurality of users during use of an application, wherein each user interaction sequence of the plurality of user interaction sequences corresponds to a session of a different user of the plurality of users;
tokenize the plurality of user interaction sequences to generate a vector representation that indicates a quantity of occurrences of each of a plurality of sub-sequences within the plurality of user interaction sequences, wherein the plurality of sub-sequences includes sub-sequences having different lengths; and
identify a set of frequent user interaction sub-sequences by mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences; and selecting, based on the vector representation, the corresponding user interaction sub-sequences as the set of frequent user interaction sub-sequences.
The amendment underlines appear to add new matter, if not previously claimed in the original claims. The specification does not clearly state that each user interaction sequence corresponds to a session. This appears new mater added to the claims. The same appears to be the case for “each sequence corresponds to a different user”.
The spec uses phrases like “separated on a per user basis” and “from a plurality of different users,” but does not disclose session-based sequencing tied one-to-one with different users.
Moreover, Other limitations: The spec does repeats vector representation, but the description is high level and does not fully disclose the detail as to how the vector representation is performed or results achieved. MPEP 2161 /written-description concern, especially for a computer-implemented claim that is drafted at a fairly high level of functional desired. Do the specification does more than describe a desired outcome for identify sequences? tokenize them, find frequent subsequences, or whether it actually describes how the inventors possessed the claimed invention in sufficient detail. MPEP 2161 require a description of the invention itself, not merely a statement of a result to be achieved. The example in the specification of Z = AB” / “Y = AAB” looks like an example, not a teaching of how the inventor intend to achieve this. the inventors contemplated one illustrative tokenization outcome, but it does not necessarily teach, how all sequences are tokenized, how the representation is constructed, how frequent subsequences are selected, how mapping is performed, or how the claimed multi-user/session limitation is implemented. This example (fig. 1B, shows one illustrative outcome of token substitution, but it does not provide an adequate description of the claimed genus of tokenizing a plurality of user interaction sequences to generate a vector representation for sub-sequences of different lengths. The Federal Circuit has explained that a specification cannot always support expansive claim language and satisfy the requirements of 35 U.S.C. 112 "merely by clearly describing one embodiment of the thing claimed." LizardTech v. Earth Resource Mapping, Inc., 424 F.3d 1336, 1346, 76 USPQ2d 1731, 1733 (Fed. Cir. 2005). It is not enough that one skilled in the art could write a program to achieve the claimed function because the specification must explain how the inventor intends to achieve the claimed function to satisfy the written description requirement. See, e.g., Vasudevan Software, Inc. v. MicroStrategy, Inc., 782 F.3d 671, 681-683, 114 USPQ2d 1349, 1356, 1357 (Fed. Cir. 2015).
The specification does not reasonably convey to one of ordinary skill in the art that the inventor was in possession of the claimed invention as of the filing date. Rather, the disclosure primarily describes desired results and a high-level functional objective—i.e., detecting patterns in event data, tokenizing user interaction sequences, and identifying frequent sub-sequences—without adequately describing the claimed subject matter at the level of detail required to demonstrate possession of the full scope of the claim. The disclosure primarily states desired results and high-level functional objectives, with only a limited illustrative example, and does not adequately describe the claimed genus, including the specific multi-user/session-based identification, tokenization, mapping, and selection limitations.
The claimed invention recites, in pertinent part, a system configured to: (i) identify a plurality of user interaction sequences indicating user interactions representing actions of a plurality of users during use of an application, wherein each user interaction sequence corresponds to a session of a different user; (ii) tokenize the plurality of user interaction sequences to generate a vector representation indicating a quantity of occurrences of each of a plurality of sub-sequences having different lengths; and (iii) identify, based at least in part on the vector representation, a set of frequent user interaction sub-sequences by mapping the plurality of sub-sequences to corresponding user interaction sub-sequences and selecting the corresponding user interaction sub-sequences as the set of frequent user interaction sub-sequences.
The specification, as filed, does not reasonably convey to those skilled in the art that the inventor had possession of the claimed subject matter, particularly with respect to the limitations recited in claim 1. See MPEP 2161. The disclosure is framed largely in terms of the desired result of detecting patterns in event data and identifying frequent subsequences, rather than a sufficient description of the claimed invention itself.
For example, the specification states that “some implementations described herein include an event sequence system that detects patterns in event data in a computationally efficient manner,” and that the system may “identify patterns in the data” and “surface patterns that may have otherwise gone undetected.” These statements describe the intended result or objective of the disclosed system, but do not, by themselves, provide an adequate written description of the claimed subject matter.
With respect to the tokenization limitation, the specification states that the event sequence system may “tokenize the set of user interaction sequences to generate a vector representation” and that the sub-sequences may include “sub-sequences of different lengths.” The disclosure further indicates that tokenization may use byte pair encoding and may involve a maximum token length threshold. However, these passages are stated at a high level and do not adequately describe the claimed tokenization of a plurality of user interaction sequences into a vector representation indicating a quantity of occurrences of each of a plurality of sub-sequences having different lengths, across the breadth of the claimed genus. No algorithm is described. The specification’s discussion in this regard is primarily functional and result-oriented.
The specification likewise states that the event sequence system may “identify, based at least in part on the vector representation, a set of frequent token sub-sequences” and may identify the “N most frequent user interaction sub-sequences.” However, the disclosure does not sufficiently describe the claimed operation of mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences and selecting, based on the vector representation, the corresponding user interaction sub-sequences as the set of frequent user interaction sub-sequences. Rather, the specification generally describes an end result—i.e., identification of frequent subsequences—without adequately describing the claimed algorithm for achieving that result.
In addition, the claim recites that each user interaction sequence corresponds to a session of a different user of the plurality of users. The specification does not clearly describe this limitation. Introduces New Matter into the claim. The closest disclosure states that the clickstream data may be “separated on a per user basis, combined together for multiple users, or both,” and that the set of user interaction sequences may comprise user interactions “from a plurality of different users.” This disclosure is insufficient to demonstrate possession of the claimed limitation that each sequence corresponds to a session of a different user.
Accordingly, the specification fails to reasonably convey to one of ordinary skill in the art that the inventor was in possession of the claimed subject matter at the time of filing, because the disclosure primarily recites desired results and high-level functional objectives, with only limited illustrative examples, and does not adequately support the full scope of the claimed genus.
All dependents’ claims are rejected based on the rational set forth in the all-independent claims respectively.
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.
Claim(s) 1-4,6-8,10-17, and 19-23 are rejected under 35 U.S.C. 103 as being unpatentable over Hassan et al US 2021/0166104 in view of Sevrens et al US 2018/0300608 and Ludwig US 2012/0280927.
As per claim 1, Hassan discloses A system for processing event data, the system comprising: one or more memories (fig.1, servers 104), coupled to the one or more memories ); and one or more processors, coupled to the one or more memories ((fig.1, servers 104), coupled to the one or more memories ), configured to:
identify a set of user interaction sequences indicating user interactions of a plurality of users with an application (0004 actions include collecting data for interactions performed by users, i.e. plurality of users, through a portal page, the data comprising a plurality of sequences of interactions performed by a user on representations of products displayed in the portal page; 0029 server systems accept requests for application services and provides such services to any number of client devices (e.g., the client device 102 over the network 106 and, 0044 The received interactions by the first user comprise a sequence of actions, i.e. identify a set of user, performed in relation to one or more products displayed in the portal page, i.e. with an application and 0058 data is collected for interactions performed by users through the web portal 440, i.e. with application);
tokenize the set of user interaction sequences to generate a vector representation that indicates a quantity of occurrences of each of a plurality of sub-sequences, having fixed lengths, within the set of user interaction sequences( 0006 the plurality of sequences of interactions are defined to have a fixed equal length for a number of interactions within a sequence, wherein the number corresponds to a sequence length value defined by the neural network. The interactions within a sequence from the plurality of sequences may correspond to a single user from the users and 0007 the plurality of sequences by the neural network comprises initializing an embedded layer to map product identifiers of the products from the portal page to product vector data, wherein the product vector data defines a number of products, i.e. tokenize provided through the portal page, a dimension of the product vector data, and a length value of a sequence to be input to the neural network for training and generating the model for product identification; and generating the model for product identification, wherein the model comprises learning units as hidden layers for iteratively processing input comprising sequences from the plurality of sequences),
and
identify, based at least in part on the vector representation, a set of frequent user interaction sub-sequences that are included in the set of user interaction sequences (0004 actions include collecting data for interactions, i.e. identify, performed by users through a portal page, the data comprising a plurality of sequences of interactions performed by a user on representations of products displayed in the portal page and [0038] At 230, the plurality of sequences of interactions are processed by the neural network through a plurality of learning layers. A model for product recommendation is generated at the neural network. The model is for identifying a product to be recommended and displayed at the portal page).
Hassan does not disclose tokenize the set of user interaction sequences to generate a vector representation that indicates a quantity of occurrences of each of a plurality of sub-sequences, having different lengths, within the set of user interaction sequences(emphasis added); identify a set of frequent user interaction sub-sequences by: mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences; and selecting, based on the vector representation, the corresponding user interaction sub-sequences the set of frequent user interaction sub-sequences,".
Sevrens discloses tokenize the set of user interaction sequences to generate a vector representation that indicates a quantity of occurrences of each of a plurality of sub-sequences, having different lengths, within the set of user interaction sequences (par 0018/0030 the information extraction system 102 receives transaction data 104 from a transaction server 106, A transaction can be an instance, i.e. application, of interaction between a first user and a second user, i.e. set of users (e.g., between two humans and , 0023/0030 the information extraction system 102 builds the language models 108 using an iterative approach of having a trained classifier label the training data. The information extraction system 102 labels the very first set of training data using an unsupervised classifier. 0024/0032 Categorization can include identifying entities, i.e. the token, from background information and labeling, i.e. tokenize, the identified entities. An entity can include one or more tokens. An entity type can be a label of an entity. [0040] The BI-LSTM models include a first-level model. The first-level model of the RNN processing module 206 can be a BI-LSTM model, or another model, that encodes tokens in a transaction record at a character-level. A token in a transaction record can be a character sequence having an arbitrary length. A token may not be clearly delimited from another token. The first-level model of the RNN processing module 206, by encoding tokens in a transaction record at a character level, can recognize the tokens even if the tokens are not delimited or contain delimiters, e.g., spaces, within, or if the tokens have never been exposed to the RNN processing module 206. A token in a transaction record can be a character sequence having an arbitrary length, i.e. having different length).
Hassan and Sevrens are both considered to be analogous to the claimed invention because they are in the same field of token.
Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Hassan to incorporate the teachings of Sevrens and provide arbitrary length of the token.
Doing so would provide an adaptive-length tokenization system that can reduce tokens in a sequence while maintaining accuracy, thereby improving the protection of products from side-channel attacks by masking the network-level signals that attackers use to infer the model's output.
However, Ludwig US 2012/0280927 discloses identify a set of frequent user interaction sub-sequences by (0048 a sequence of interactions and the context selection determines, i.e. identify, a context mapping and 0167 a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping ): mapping the plurality of sub-sequences to corresponding user interaction sub-sequences in the plurality of user interaction sequences(0167] However, in general many subtasks are performed in sequence, and at least some sort of context switch is involved between each sequence step. For example, FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations. And see also 0260 ); and selecting, based on the vector representation, the corresponding user interaction sub-sequences the set of frequent user interaction sub-sequences,"( [0126] FIG. 2b depicts use of a pointing device (such as a mouse, trackball, or conventional touchpad), first used for selecting focus or context, and secondly directing pointing device output to GUI elements either in the drawing area or GUI "task selection" area. In each area, the traditional GUI provides such a "graphical phrase dictionary" for the application language control and command instructions, sometimes directing in a drawing/viewing area, other times in a "task selection area." And 0167 FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations.)
Hassan and Sevrens and Ludwig are both considered to be analogous to the claimed invention because they are in the same field of token.
Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Hassan to incorporate the teachings of Sevrens, including the teaching of Ludwig and provide context mapping for derive the token. Doing so would provide a security fingerprint for the token, thereby improving secure communication in the network.
As per claim 2, Hassan and Sevrens and Ludwig discloses the system of claim 1, Hassan discloses wherein the one or more processors are further configured to: obtain click stream data representing the actions of the plurality of users with the application, wherein the plurality of user interaction sequences is identified from the click stream data(0030 The user interface, i.e. application, may include different tools and interaction controls to navigate between different product categories and select products of certain type and of certain properties. 0058 data is collected for interactions performed by users through the web portal 440, i.e. with application,, The input 455 is data collected, i.e. obtain, from interactions performed by users and may be organized in forms of multiple sequences of interactions and a sequence of interactions performed by user A may be five consecutive clicking operations on different products presented on the user interface of the web portal 440. The interactions within a sequence may correspond also to click operations that define purchase transactions, or viewing request, or request for including a product in a purchase bag. The interactions collected for different user may represent customer behavior in purchasing products through a web portal)).
As per claim 3, Hassan and Sevrens and Ludwig discloses The system of the claim 2, Hassan discloses wherein the one or more processors, to identify the set of user interaction sequences, are configured to: identify a subset of the user interactions with the application from the click stream data as the plurality of user interaction sequences ( 0054 the selection of products may be customized for the particular user 330 and his profile and behavior characteristics. When a different user logs in the e-commerce web portal 310, and start to interact with the web content, based on review of his interaction and/or his user profile characteristics (if such can be identified), a different set of products may be displayed at the product panel 315. And 0058 data is collected for interactions performed by users through the web portal 440, and such data is stored as input 455 to be provided for training at a neural network. The input 455 is data collected from interactions performed by users and may be organized in forms of multiple sequences of interactions, where one sequence correspond to one user from the users. The sequences are defined with fixed length and correspond to different types of interaction events performed by a user. For example, a sequence of interactions performed by user A may be five consecutive clicking operations on different products presented on the user interface of the web portal 440. The interactions within a sequence may correspond also to click operations that define purchase transactions, or viewing request, or request for including a product in a purchase bag. The interactions collected for different user may represent customer behavior in purchasing products through a web portal).
As per claim 4. Hassan and Sevrens and Ludwig discloses The system of the claim 1, Hassan discloses wherein each user interaction sequence of the plurality of user interaction sequences comprises categorical univariate time series data representing one or more users of the actions during the session (0052 The CNN processing module can determine the classification of the transaction record based on a final layer in the series by feeding an output of the transforming to a fully connected feed forward network. The final layer before feeding into the fully connected feed forward network can be a pooling layer).
As per claim 6. Hassan and Sevrens and Ludwig discloses The system of the claim 1, Sevrens discloses wherein the one or more processors, to tokenize the plurality of user interaction sequences, are configured to: tokenize the plurality of user interaction sequences to compress the plurality of user interaction sequences (0040 The first-level model of the RNN processing module 206 can be a BI-LSTM model, or another model, that encodes tokens in a transaction record at a character-level. A token in a transaction record can be a character sequence having an arbitrary length. A token may not be clearly delimited from another token. The first-level model of the RNN processing module 206, by encoding tokens in a transaction record at a character level ).
As per claim 7. Hassan and Sevrens and Ludwig discloses The system of claim 6, Hassan discloses The system of wherein the one or more processors, to tokenize the plurality of user interaction sequences, are configured to: compress the plurality of user interaction sequences based at least in part on a maximum token length threshold (0003 collecting data for interactions on products displayed at a user interface of a portal page and inputting the collected data for training a neural network on temporal dependencies between interactions within a sequence of certain length ).
As per clam 8. Hassan and Sevrens and Ludwig discloses The system of claim 1, Hassan discloses wherein the one or more processors, to tokenize the plurality of user interaction sequences, are configured to: tokenize the plurality of user interaction sequences using byte pair encoding (0024/0032 Categorization can include identifying entities, i.e. the token, from background information and labeling, i.e. tokenize, the identified entities. An entity can include one or more tokens. An entity type can be a label of an entity. [0040] The BI-LSTM models include a first-level model. The first-level model of the RNN processing module 206 can be a BI-LSTM model, or another model, that encodes tokens in a transaction record at a character-level. A token in a transaction record can be a character sequence having an arbitrary length. A token may not be clearly delimited from another token. The first-level model of the RNN processing module 206, by encoding tokens in a transaction record at a character level, can recognize the tokens even if the tokens are not delimited or contain delimiters, e.g., spaces, within, or if the tokens have never been exposed to the RNN processing module 206. A token in a transaction record can be a character sequence having an arbitrary length, i.e. having different length and 0033 The CNN processing module 202 includes a one-dimensional convolutional neural network where textual data are encoded via a one-of-m encoding for each character ).
As per claim 10. Hassan and Sevrens and Ludwig discloses The system of claim 1, Hassan discloses wherein the set of frequent user interaction sub-sequences comprises N most frequent user interaction sub-sequences based on the vector representation(0060 the input 455 provided to the product recommendation system 410 is in form of a data matrix of size (N×M), where M is the size of a number of interactions collected for a single sequence, and N is the number of multiple sequences that are input to the product recommendation system for the neural network training ).
As per claim 11, this method claim is rejected based on the same rational set forth in the claim 1.
As per claim 12. Hassan and Sevrens and Ludwig discloses the method of claim 11, Hassan discloses further comprising: obtaining event stream data representing the actions of the plurality of users with the application, wherein the plurality of user interaction sequences are identified from the event stream data( 0030 The user interface, i.e. application, may include different tools and interaction controls to navigate between different product categories and select products of certain type and of certain properties. 0058 data is collected for interactions performed by users through the web portal 440, i.e. with application,, The input 455 is data collected, i.e. obtain, from interactions performed by users and may be organized in forms of multiple sequences of interactions and a sequence of interactions performed by user A may be five consecutive clicking operations on different products presented on the user interface of the web portal 440. The interactions within a sequence may correspond also to click operations that define purchase transactions, or viewing request, or request for including a product in a purchase bag. The interactions collected for different user may represent customer behavior in purchasing products through a web portal).
As per claim 13. Hassan and Sevrens and Ludwig discloses the method of claim 11, Sevrens disclose wherein identifying the plurality of user interaction sequences comprises: identifying a subset of the event stream data as the plurality of user interaction sequences (0052 The CNN processing module can determine the classification of the transaction record based on a final layer in the series by feeding an output of the transforming to a fully connected feed forward network. The final layer before feeding into the fully connected feed forward network can be a pooling layer ).
As per claim 14. Hassan and Sevrens and Ludwig discloses the method of claim 11, Hassan discloses wherein the plurality of user interaction sequences comprises user interactions with a web page of the application( 0010 in response to receiving the interaction by the first user through the portal page, displaying the first product as part of content dynamically presented in the portal page).
As per claim 15. Hassan and Sevrens and Ludwig discloses the method of claim 11, Sevrens discloses wherein tokenizing the plurality of user interaction sequences comprises: tokenizing the plurality of user interaction sequence sequences to compress the plurality of user interaction sequences(0040 The first-level model of the RNN processing module 206 can be a BI-LSTM model, or another model, that encodes tokens in a transaction record at a character-level. A token in a transaction record can be a character sequence having an arbitrary length. A token may not be clearly delimited from another token. The first-level model of the RNN processing module 206, by encoding tokens in a transaction record at a character level ).
As per claim 16. Hassan and Sevrens and Ludwig discloses the method of claim 15 Hassan discloses wherein tokenizing the plurality of user interaction sequences comprises: compressing the plurality of user interaction sequence sequences until token sub- sequences satisfy a maximum token length threshold(0003 collecting data for interactions on products displayed at a user interface of a portal page and inputting the collected data for training a neural network on temporal dependencies between interactions within a sequence of certain length. ).
As per claim 17. Hassan and Sevrens and Ludwig discloses the method of claim 11 Hassan discloses wherein tokenizing the plurality of user interaction sequences comprises: tokenizing the plurality of user interaction sequences using byte pair encoding ( 0024/0032 Categorization can include identifying entities, i.e. the token, from background information and labeling, i.e. tokenize, the identified entities. An entity can include one or more tokens. An entity type can be a label of an entity. [0040] The BI-LSTM models include a first-level model. The first-level model of the RNN processing module 206 can be a BI-LSTM model, or another model, that encodes tokens in a transaction record at a character-level. A token in a transaction record can be a character sequence having an arbitrary length. A token may not be clearly delimited from another token. The first-level model of the RNN processing module 206, by encoding tokens in a transaction record at a character level, can recognize the tokens even if the tokens are not delimited or contain delimiters, e.g., spaces, within, or if the tokens have never been exposed to the RNN processing module 206. A token in a transaction record can be a character sequence having an arbitrary length, i.e. having different length and 0033 The CNN processing module 202 includes a one-dimensional convolutional neural network where textual data are encoded via a one-of-m encoding for each character).
As per claim 19. Hassan and Sevrens and Ludwig discloses The method of claim 11 Hassan discloses wherein the set of frequent user interaction sub-sequences comprises N most frequent user interaction sub-sequences based on the vector representation(0060 the input 455 provided to the product recommendation system 410 is in form of a data matrix of size (N×M), where M is the size of a number of interactions collected for a single sequence, and N is the number of multiple sequences that are input to the product recommendation system for the neural network training).
As per claim 20, this claim is rejected based on the same rational set forth in the claim 1.
As per claim 21. Hassan and Sevrens and Ludwig discloses The non-transitory computer-readable medium of claim 20, Ludwig discloses wherein the one or more processors cause the event data processing system to: obtain click stream data representing the actions of the plurality of users with the application, wherein the plurality of user interaction sequences is identified from the click stream data(0167] However, in general many subtasks are performed in sequence, and at least some sort of context switch is involved between each sequence step. For example, FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations. And see also 0260).
As per claim 22. Hassan and Sevrens and Ludwig discloses The non-transitory computer-readable medium of claim 20, Ludwig discloses wherein each of the plurality of user interaction sequences comprises categorical univariate time series data representing one or more user actions during the session(0167] However, in general many subtasks are performed in sequence, and at least some sort of context switch is involved between each sequence step. For example, FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations. And see also 0260).
As per claim 23. Hassan and Sevrens and Ludwig discloses The non-transitory computer-readable medium of clam 20, Ludwig discloses wherein the one or more processors, to tokenize the plurality of user interaction sequences, cause the event data processing system to: tokenize the plurality of user interaction sequences using byte pair encoding (0167] However, in general many subtasks are performed in sequence, and at least some sort of context switch is involved between each sequence step. For example, FIG. 5 depicts a sequence of interactions such as those depicted in FIG. 4a, wherein a context switch is required before each task, and the context selection determines a context mapping. Further, however, the context mapping step can internally comprise its own collection of additional internal steps, for example as depicted in FIG. 6, calling out examples of further detail involved in the context mapping as including dialog and/or interactive adjustment, enter operations, cancel operations, and undo operations. And see also 0260).
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 ABU S SHOLEMAN whose telephone number is (571)270-7314. The examiner can normally be reached EST: 9am-5pm.
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, JORGE ORTIZ CRIADO can be reached at 571-272-7624. 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.
/ABU S SHOLEMAN/ Primary Examiner, Art Unit 2496
Prosecution Insights