DETAILED ACTION
This action is in reply to the Amendments filed on 04/08/2026.
Claims 2-10 are newly added.
Claims 1-10 are rejected.
Claims 1-10 are currently pending and have been examined.
Response to Amendment
Applicant’s amendment, filed 04/08/2026, has been entered. Claim 1 has been amended.
Claim Objections
The claim objections from the prior Office Action have been withdrawn pursuant Applicant’s amendments. Examiner notes that a new claim objection has been added in light of the newly added claims.
Claim Rejections Under 35 USC § 112(b)
The claim rejections under 35 USC § 112(b) from the prior Office Action have been withdrawn pursuant Applicant’s amendments.
Priority
The current Application claims priority from Provisional Application 63/515,024, filed 07/21/2023. Therefore, the instant claims receive the effective filing date of 07/21/2023.
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 .
Claim Objections
Claim 7 is objected to because of the following informalities:
-Claim 7 reads “the fused representation” but should likely read “a fused representation”
Appropriate correction is required.
Claim Rejections - 35 USC § 112(a)
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 6-7 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.
Claim 6 contains the limitation “wherein the process of fusing the relevant content with the dialogue history comprises encoding the dialogue history and the relevant content into a shared representation space.” The specification does not describe, or even recite, a “shared representation space,” and, accordingly, the way in which the dialogue history and the relevant content are encoded into a shared representation space is not described in such a way as to reasonably convey to one of ordinary skill in the art that the inventors had possession of the claimed invention at the time the application was filed. For example, the disclosure does not adequately describe the way in which the dialogue history and the relevant content are encoded into a shared representation space. Applicant’s failure to disclose any meaningful description via explanation, or the like, as to the way in which the dialogue history and the relevant content are encoded into a shared representation space raises questions whether Applicant truly had possession of this feature at the time of filing.
Claim 7 contains the limitation “wherein the final response is generated using a decoder component of the deep learning model conditioned on the fused representation.” The specification does not describe, or even recite, a “fused representation,” and, accordingly, the way in which the deep learning model is conditioned on the fused representation is not described in such a way as to reasonably convey to one of ordinary skill in the art that the inventors had possession of the claimed invention at the time the application was filed. For example, the disclosure does not adequately describe the way in which the deep learning model is conditioned on the fused representation. Applicant’s failure to disclose any meaningful description via explanation, or the like, as to the way in which the deep learning model is conditioned on the fused representation raises questions whether Applicant truly had possession of this feature at the time of filing.
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 1-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., law of nature, a natural phenomenon, or an abstract idea) without significantly more.
Under Step 1 of the Subject Matter Eligibility Test for Products and Processes, the claims must be directed to one of the four statutory categories (see MPEP 2106.03). All the claims are directed to one of the four statutory categories (YES).
Under Step 2A of the Subject Matter Eligibility Test, it is determined whether the claims are directed to a judicially recognized exception (see MPEP 2106.04). Step 2A is a two-prong inquiry.
Under Prong 1, it is determined whether the claim recites a judicial exception (YES). Taking Claim 1 as representative, the claim recites limitations that fall within the certain methods of organizing human activity groupings of abstract ideas, including:
-using, via a deep learning model executing on the one or more computer systems, search engines to look for external and factual knowledge across the internet;
-finding relevant content for at least one aspect of a product, wherein the at least one aspect includes one or more of price, reviews, and features; and
-completing a process of fusing the relevant content with dialogue history in order to provide a final response;
-wherein the using, the finding, and the completing are to connect an Alexa® socialbot with an Amazon@ Store, opening novel functions including one or more of better recommendations, conversational shopping guidance, automatic seeking of new product types, and personalization
The above limitations recite the concept of searching for and providing relevant product information. The above limitations fall within the “Certain Methods of Organizing Human Activity” groupings of abstract ideas, enumerated in MPEP 2106.04(a).
Certain methods of organizing human activity include:
fundamental economic principles or practices (including hedging, insurance, and mitigating risk)
commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; and business relations)
managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions)
The limitations of finding relevant content for at least one aspect of a product, wherein the at least one aspect includes one or more of price, reviews, and features; and completing a process of fusing the relevant content with dialogue history in order to provide a final response are processes that, under their broadest reasonable interpretation, cover a commercial interaction. For example, “finding” and “completing” in the context of this claim encompass advertising, and marketing or sales activities.
Similarly, the limitations of using, via a deep learning model executing on the one or more computer systems, search engines to look for external and factual knowledge across the internet; wherein the using, the finding, and the completing are to connect an Alexa® socialbot with an Amazon@ Store, opening novel functions including one or more of better recommendations, conversational shopping guidance, automatic seeking of new product types, and personalization are processes that, under their broadest reasonable interpretation, cover a commercial interaction. That is, other than reciting that the model is a deep learning model executing on the one or more computer systems, that search engines are used to look for external and factual knowledge, that the looking is across the internet, and that the connecting with the Amazon Store is via the Alexa socialbot, nothing in the claim element precludes the step from practically being performed by people. For example, but for the “deep learning model,” “one or more computer systems,” “search engines,” “the internet,” and “the Alexa socialbot” language, “using” and “connect” in the context of this claim encompasses advertising, and marketing or sales activities.
Under Prong 2, it is determined whether the claim recites additional elements that integrate the exception into a practical application of the exception. This judicial exception is not integrated into a practical application (NO).
-using, via a deep learning model executing on the one or more computer systems, search engines to look for external and factual knowledge across the internet;
-finding relevant content for at least one aspect of a product, wherein the at least one aspect includes one or more of price, reviews, and features; and
-completing a process of fusing the relevant content with dialogue history in order to provide a final response;
-wherein the using, the finding, and the completing are to connect an Alexa® socialbot with an Amazon@ Store, opening novel functions including one or more of better recommendations, conversational shopping guidance, automatic seeking of new product types, and personalization
The additional elements of claim 1 are recited at a high level of generality (i.e. as generic computing hardware) such that they amount to nothing more than mere instructions to implement or apply the abstract idea on a generic computing hardware (or, merely use a computer as a tool to perform an abstract idea) as supported by paragraph [0042] of Applicant’s specification – “The processor 304 can be any custom made or commercially available processor, a central processor unit (CPU), an auxiliary processor among several processors associated with the computer controller 300, a semiconductor based microprocessor (in the form of a microchip or chip set), a macroprocessor, or generally any device for executing software instructions.” Specifically, the additional elements of one or more computer systems, a deep learning model, search engines, the internet, and the Alexa socialbot are recited at a high-level of generality (i.e. as a generic processor performing the generic computer functions of using a model to look for data, finding data, fusing data [i.e. collecting data], and connecting) such that they amount do no more than mere instructions to apply the exception using generic computer components. Accordingly, 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. The claim is directed to an abstract idea. Further, the additional elements do no more than generally link the use of the judicial exception to a particular technological environment or field of use (such as computers or computing networks). Employing well-known computer functions to execute an abstract idea, even when limiting the use of the idea to one particular environment, does not integrate the exception into a practical application.
Additionally, the additional elements are insufficient to integrate the abstract idea into a practical application because the claim fails to i) reflect an improvement in the functioning of a computer or an improvement to another technology or technical field, ii) apply the judicial exception with, or use the judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim, iii) effect a transformation or reduction of a particular article to a different state or thing, or iv) apply or use the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment.
Accordingly, the judicial exception is not integrated into a practical application.
Under Step 2B, it is determined whether the claims recite additional elements that amount to significantly more than the judicial exception. The claims of the present application do not include additional elements that are sufficient to amount to significantly more than the judicial exception (NO).
In the case of claim 1, taken individually or as a whole, the additional elements of claim 9 do not provide an inventive concept. As discussed above under step 2A (prong 2) with respect to the integration of the abstract idea into a practical application, the additional elements used to perform the claimed functions amount to no more than a general link to a technological environment.
Even considered as an ordered combination (as a whole), the additional elements do not add anything significantly more than when considered individually.
Dependent claims 2-10, when analyzed as a whole, are held to be patent ineligible under 35 U.S.C. § 101 because they do not add “significantly more” to the abstract idea. More specifically, dependent claims 2-10 further fall within the “Certain Methods of Organizing Human Activity” grouping of abstract ideas in that they recite commercial interactions. Dependent claims 4-5 and 9-10, do not recite any farther additional elements, and as such are not indicative of integration into a practical application for at least similar reasons discussed above. Dependent claims 2-3 and 6-8 recite the additional elements of the deep learning model, transformer-based architecture that is trained, the search engines, internet-based search services, third-party websites, encoding, a decoder component, and reinforcement learning but similar to the analysis under prong two of Step 2A these additional elements are used as a tool to perform the abstract idea. As such, under prong two of Step 2A, claims 2-10 are not indicative of integration into a practical application for at least similar reasons as discussed above. Thus, dependent claims 2-10 are “directed to” an abstract idea. Next, under Step 2B, similar to the analysis of claim 1, dependent claims 2-10 when analyzed individually and as an ordered combination, merely further define the commonplace business method (i.e. searching for and providing relevant product information) being applied on a general-purpose computer and, therefore, do not amount to significantly more than the abstract idea itself. Accordingly, the Examiner concludes that there are no meaningful limitations in the claims that transform the judicial exception into a patent eligible application such that the claims amount to significantly more than the judicial exception itself. The analysis above applies to all statutory categories of invention.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 3, 5, and 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over Bright et al. (US 2019/0012714 A1), hereinafter Bright, in view of Baligar et al. (US 10,997,963 B1), hereinafter Baligar.
Regarding claim 1, Bright discloses a deep learning method with improved search and dialogue properties, implemented by one or more computer systems and connecting, the method comprising:
-using, via a deep learning model executing on the one or more computer systems, search engines to look for external and factual knowledge across the internet (Bright, see at least: “A networked system 102, in the example forms of a network-based marketplace or payment system, provides server-side functionality via a network 104 (e.g., the Internet or a wide area network (WAN)) [i.e. across the internet] to one or more client devices 110. FIG. 1 illustrates, for example, a web client 112 (e.g., a browser, such as the Internet Explorer® browser [i.e. using search engines] developed by Microsoft® Corporation of Redmond, Wash. State), a client application 114, and a programmatic client 116 executing on the client device 110” [0040] and “Deep-learning models, deep neural networks (DNNs), recurrent neural networks (RNNs), convolutional neural networks (CNNs), and long short-term CNNs, as well as other ML models and IR models, may be used. For example, the search component 218 may use n-gram, entity, and semantic vector-based queries to product matching. Deep-learned semantic vectors give the ability to match products to non-text inputs directly. Multi-leveled relevance filtration may use BM25, predicted query leaf category+product leaf category, semantic vector similarity between query and product, and other models to pick the top candidate products [i.e. via a deep learning model executing on the one or more computer systems] for the final re-ranking algorithm” [0091] and “the networked system 102 is a network-based marketplace that responds to requests for product listings, publishes publications comprising item listings of products available on the network-based marketplace [i.e. to look for external and factual knowledge], and manages payments for marketplace transactions. One or more users 106 may be a person, a machine, or other means of interacting with the client device 110. In embodiments” [0041] and “the third-party application 132, utilizing information retrieved from the networked system 102, supports one or more features or functions on a website hosted by a third party. The third-party website, for example, provides one or more promotional, marketplace [i.e. external knowledge], or payment functions that are supported by the relevant applications of the networked system 102” [0046]);
-finding relevant content for at least one aspect of a product, wherein the at least one aspect includes one or more of price, reviews, and features (Bright, see at least: “The natural-language text input can thus be transformed into a structured query using rich information from additional knowledge to enrich the query even further. This information is then passed on to the dialogue manager 204 through the orchestrator 220 for further actions with the user or with the other components in the overall system. The structured and enriched query is also consumed by the search component 218 for improved matching [i.e. finding relevant content]” [0052] and “The AIF 144 knows about user details 312, such as user preferences, desired price ranges, sizes, affinities, etc. [i.e. for at least one aspect of a product, wherein the at least one aspect includes one or more of price, reviews, and features]” [0061] and “the AIF 144 may process input queries such as: “Hey!Can you help me find a pair of light pink shoes for my girlfriend please? With heels. Up to $200. Thanks;” “I recently searched for a men's leather jacket with a classic James Dean look. Think almost Harrison Ford's in the new Star Wars movie. However, I'm looking for quality in a price range of $200-300 [i.e. for at least one aspect of a product, wherein the at least one aspect includes one or more of price, reviews, and features]. Might not be possible, but I wanted to see!”” [0063]); and
-completing a process of fusing the relevant content with the dialogue history in order to provide a final response (Bright, see at least: “The NLU component 206 determines the object, the aspects associated with the object, how to create the search interface input, and how to generate the response. For example, the AIF 144 may ask questions to the user to clarify what the user is looking for [i.e. completing a process of fusing the relevant content with the dialogue history]. This means that the AIF 144 not only generates results [i.e. in order to provide a final response], but also may create a series of interactive operations to get to the optimal, or close to optimal, results 222” [0054] and “The dialogue manager 204 is the component that analyzes the query of a user to extract meaning, and determines if there is a question that needs to be asked in order to refine the query, before sending the query to the search component 218. The dialogue manager 204 uses the current communication in the context of the previous communication [i.e. with the dialogue history] between the user and the AIF 144. The questions are automatically generated dependent on the combination of the accumulated knowledge (e.g., provided by a knowledge graph) and what the search component 218 can extract out of the inventory [i.e. fusing the relevant content]” [0056] and “the AIF 144 performs proactive data extraction 310 from multiple sources, such as social networks, email, calendar, news, market trends, etc. The AIF 144 knows about user details 312, such as user preferences, desired price ranges, sizes, affinities, etc.” [0061] Examiner notes that while art is applied “in order to provide the final response” is an intended result and therefore holds little patentable weight);
-wherein the using, the finding, and the completing are to connect a socialbot with a Store, opening novel functions including one or more of better recommendations, conversational shopping guidance, automatic seeking of new product types, and personalization (Bright, see at least: “the AIF 144 performs proactive data extraction 310 from multiple sources, such as social networks, email, calendar, news, market trends, etc. The AIF 144 knows about user details 312, such as user preferences, desired price ranges, sizes, affinities, etc. The AIF 144 facilitates a plurality of services within the service network, such as product search, personalization, recommendations, checkout features, etc. Output 308 may include recommendations, results [i.e. wherein the using, the finding, and the completing]” [0061] and “the intelligent personal assistant system of FIG. 2 is shown to include a front-end component 502 (FE) by which the intelligent personal assistant system 142 communicates (e.g., over the network 104) with other systems within the network architecture 100. The front-end component 502 can communicate with the fabric of existing messaging systems. As used herein, the term “messaging fabric” refers to a collection of APIs and services that can power third-party platforms such as Facebook messenger, Microsoft Cortana, and other “bots.” In one example, a messaging fabric can support an online commerce ecosystem that allows users to interact with commercial intent [i.e. to connect a socialbot with a Store]” [0079] and “An application programming interface (API) server 120 and a web server 122 are coupled to, and provide programmatic and web interfaces respectively to, one or more application servers 140. The application server 140 hosts an intelligent personal assistant system 142 [i.e. to connect a socialbot with a Store]” [0044] and “Embodiments present a personal shopping assistant, also referred to as an intelligent assistant, that supports a two-way communication with the shopper to build context and understand the intent of the shopper, enabling delivery of better, personalized shopping results [i.e. opening novel functions including one or more of better recommendations, conversational shopping guidance, automatic seeking of new product types, and personalization]” [0050]).
Bright does not explicitly disclose the socialbot being an Alexa socialbot and the store being an Amazon Store.
Baligar, however, teaches utilizing a voice assistant (i.e. abstract), including the known technique of connecting the Alexa socialbot with the Amazon Store (Baligar, see at least: “Examples of voice assistant systems include Alexa® [i.e. the Alexa socialbot] provided by Amazon.com® [i.e. the Amazon Store] of Seattle, Wash.” Col. 2 Ln. 7-9 and “FIG. 1 is a schematic diagram of an illustrative computing environment 100 that includes a voice assistant service 102 that augments information provided by a content provider 104 [i.e. connect the Alexa socialbot with the Amazon Store]. The environment may include a user device 106 operated by a user 108. The user device 106 may exchange information from the content provider 104. As an example, the user device 104 may execute a browser application that enables the user 108 to interact with content provided by the content provider 104. For example, the content provider 104 may host an electronic marketplace [i.e. the Amazon Store] that enables the user 108 to consume products and/or services, referred to collectively herein as “items”” Col. 3 Ln. 37-48). This known technique is applicable to the method of Bright as they both share characteristics and capabilities, namely, they are directed to utilizing a voice assistant.
It would have been recognized that applying the known technique of connecting the Alexa socialbot with the Amazon Store, as taught by Baligar, to the teachings of Bright would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such references into similar methods. Further, adding the modification of connecting the Alexa socialbot with the Amazon Store, as taught by Baligar, into the method of Bright would have been recognized by those of ordinary skill in the art as resulting in an improved method that would provide accurate, intuitive, and relatively quick interactions in order to instruct the computing devices to perform desired functions (Baligar, Col. 1 Ln. 19-21).
Regarding claim 3, Bright in view of Baligar teaches the method of claim 1. Bright further discloses:
-wherein the search engines include internet-based search services configured to retrieve real-time product information from third-party websites (Bright, see at least: “A networked system 102, in the example forms of a network-based marketplace or payment system, provides server-side functionality via a network 104 (e.g., the Internet or a wide area network (WAN)) to one or more client devices 110. FIG. 1 illustrates, for example, a web client 112 (e.g., a browser, such as the Internet Explorer® browser [i.e. wherein the search engines include internet-based search services configured to retrieve real-time product information] developed by Microsoft® Corporation of Redmond, Wash. State), a client application 114, and a programmatic client 116 executing on the client device 110” [0040] and “the networked system 102 is a network-based marketplace that responds to requests for product listings, publishes publications comprising item listings of products available on the network-based marketplace [i.e. from third-party websites], and manages payments for marketplace transactions. One or more users 106 may be a person, a machine, or other means of interacting with the client device 110. In embodiments” [0041] and “the third-party application 132, utilizing information retrieved from the networked system 102, supports one or more features or functions on a website hosted by a third party. The third-party website, for example, provides one or more promotional, marketplace [i.e. from third-party websites], or payment functions that are supported by the relevant applications of the networked system 102” [0046]).
Regarding claim 5, Bright in view of Baligar teaches the method of claim 1. Bright further discloses:
-wherein the dialogue history comprises a sequence of user utterances and system responses from a multi-turn conversation session (Bright, see at least: “The dialogue manager 204 is the component that analyzes the query of a user to extract meaning, and determines if there is a question that needs to be asked in order to refine the query, before sending the query to the search component 218. The dialogue manager 204 uses the current communication in the context of the previous communication [i.e. wherein the dialogue history comprises a sequence of user utterances and system responses] between the user and the AIF 144.” [0056] and “FIG. 9 is a graphical representation of a service sequence for a chat turn [i.e. from a multi-turn conversation session] with a structured answer, according to some example embodiments. In some example embodiments, the client application performs functions of the NLU component 206 or provides choices to the user regarding filters for browsing. As a result, the client sends structured data ready for consumption by the dialogue manager 204” [0122] and “the BFF 504 sends the “structured answer” received from the client to the orchestrator 220, which then sends it to the dialogue manager 204. The dialogue manager 204 returns actions and parameters for the structured answer, and the orchestrator 220 sends the search request with the parameters to the search component 218. If necessary, narrowing questions may be sent to the user for narrowing the search, by using the dialogue manager 204 to formulate the questions” [0123]).
Regarding claim 9, Bright in view of Baligar teaches the method of claim 1. Bright further discloses:
-wherein the personalization function comprises adapting the final response based on a user profile including prior purchase history and stated preferences (Bright, see at least: “a knowledge graph is utilized to identify the aspects, based on analysis of user behavior while interacting with the system. For example, when users look for messenger bags, what is the click pattern of these users while searching for the messenger bags (e.g., selecting brand or color, or adding results to the search query). The NLU component 206 may provide questions to be asked with reference to the intent and the aspects. For example, the NLU may indicate asking, “I have messenger bags for these four brands, A, B, C, and D; do you have a brand preference?” [i.e. wherein the personalization function comprises adapting the final response based on a user profile including stated preferences]” [0104] and “Another example may provide a personalized default option based on a user's historical choices (e.g., a user always selects Adidas shoes when buying footwear) [i.e. the personalization function comprises adapting the final response based on a user profile including prior purchase history]” [0140]).
Regarding claim 10, Bright in view of Baligar teaches the method of claim 1. Bright further discloses:
-wherein the conversational shopping guidance includes suggesting alternative products based on user constraints such as budget, brand, or product features (Bright, see at least: “Other examples of proactive action shown in FIG. 18 deal with user inaction in the instance of a non-click. Here, the bot may present the next top items, with a prompt. Alternatively, similar items may be presented from a similar but different query, such as a different brand or price range [i.e. wherein the conversational shopping guidance includes suggesting alternative products based on user constraints such as budget, brand, or product features]” [0141] and “The AIF 144 knows about user details 312, such as user preferences, desired price ranges, sizes, affinities, etc. [i.e. based on user constraints such as budget, brand, or product features]” [0061] and “the AIF 144 may process input queries such as: “Hey!Can you help me find a pair of light pink shoes for my girlfriend please? With heels. Up to $200. Thanks;” “I recently searched for a men's leather jacket with a classic James Dean look. Think almost Harrison Ford's in the new Star Wars movie. However, I'm looking for quality in a price range of $200-300. Might not be possible, but I wanted to see!” [i.e. based on user constraints such as budget, brand, or product features]” [0063]).
Claims 2 and 6 are rejected under 35 U.S.C. 103 as being unpatentable over Bright, in view of Baligar, in further view of Jungmeisteris et al. (US 2022/0398635 A1), hereinafter Jungmeisteris.
Regarding claim 2, Bright in view of Baligar teaches the method of claim 1. Bright further discloses:
-wherein the deep learning model is trained on multi-turn conversational data (Bright, see at least: “the AIF 144 is trained using sample queries (e.g., a development set) and tested on a different set of queries (e.g., an evaluation set), both sets to be developed by human curation or from use data. Also, the AIF 144 is to be trained on transaction and interaction flows defined by experienced curation specialists, or human override 524. The flows and the logic encoded within the various components of the AIF 144 define what follow-up utterance or presentation (e.g., question, result set) is made by the intelligent assistant based on an identified user intent [i.e. the deep learning model is trained on multi-turn conversational data]” [0086] and “Such machine-learning algorithms operate by building a model from example inputs in order to make data-driven predictions or decisions expressed as outputs. Machine-learning algorithms may also be used to teach how to implement a process, such as the time-sensitive training of the dialogue manager 204 to improve user interaction and facilitate turns of speech discussed further above” [0090]).
Bright in view of Baligar does not explicitly disclose the deep learning model comprising a transformer-based architecture trained on multi-turn conversational data and external knowledge sources.
Jungmeisteris, however, teaches utilizing a deep learning system (i.e. [0039]), including the known technique of the deep learning model comprising a transformer-based architecture trained on multi-turn conversational data and external knowledge sources (Jungmeisteris, see at least: “sentiment analysis from text is performed by one or more supervised or unsupervised algorithms. In an exemplary embodiment, a transformer-based deep learning technique is used for sentiment analysis and other large scale NLP processing tasks. The transformer may be trained on the dataset described above with regard to step 502. Exemplary transformer models may include XLM-RoBERTa, or other models based on BERT [i.e. the deep learning model comprises a transformer-based architecture]. In some embodiments, sentiment analysis task is modeled as a classification problem, whereby a classifier is fed a text input and returns a category, e.g. positive, negative, or neutral. This may involve feature extraction from freeform text e.g., to generate vectors for words or sentences” [0077] and “one or more machine learning models have been trained on training sets. In step 502, the training set is a set of character string data simulating potential input text typed in by a user. In step 520, the training set is an exemplary or curated set of customer survey data from a variety of different sources 251-257 [i.e. trained on external knowledge sources]. This training data in steps 502 and 520 may encompass text directed to a variety of topics and a variety of languages, formality of speech, and so on, so as to provide a variety of possible input text … The machine learning models extract features from this training data to develop one or more trained models capable of sentiment analysis and topic classification. In an exemplary embodiment, NLP (natural language processing) models may be trained on the training set on a periodic or scheduled basis, for instance daily, weekly, monthly or the like, or in real-time [i.e. trained on multi-turn conversational data], depending on the size of the collection and the frequency of relevant change within that collection, to optimize the weighting applied by the various ML (machine learning) models applied to the systems described herein” [0067] and “The memory 210 may also, in one embodiment, include communication logic 224, including one or more APIs for obtaining information from or communicating information with database 260 (or other external or third party databases) and obtaining survey data 251-257 from web server 140 and/or via network 130 (FIG. 1 [i.e. trained on external knowledge sources])” [0037] and “the ML models described above are not limited to the text input by the user and may additionally or alternately use historical data regarding the user's prior interactions with the system 110. For instance, in a first customer support instance, a tone (or style) of response (e.g., language, formality, linguistic traits) may be detected [i.e. trained on multi-turn conversational data], and an identifier for such a tone of response may be stored in memory 210 in association with user data 231” [0075]). This known technique is applicable to the method of Bright in view of Baligar as they both share characteristics and capabilities, namely, they are directed to utilizing a deep learning system.
It would have been recognized that applying the known technique of the deep learning model comprising a transformer-based architecture trained on multi-turn conversational data and external knowledge sources, as taught by Jungmeisteris, to the teachings of Bright in view of Baligar would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such references into similar methods. Further, adding the modification of the deep learning model comprising a transformer-based architecture trained on multi-turn conversational data and external knowledge sources, as taught by Jungmeisteris, into the method of Bright in view of Baligar would have been recognized by those of ordinary skill in the art as resulting in an improved method that would provide a more dynamic, accurate, and timely customer sentiment analysis (Jungmeisteris, [0005]).
Regarding claim 6, Bright in view of Baligar teaches the method of claim 1.
Bright in view of Baligar does not explicitly disclose the process of fusing the relevant content with the dialogue history comprising encoding the dialogue history and the relevant content into a shared representation space.
Jungmeisteris, however, teaches utilizing a deep learning system (i.e. [0039]), including the known technique of the process of fusing the relevant content with the dialogue history comprising encoding the dialogue history and the relevant content into a shared representation space (Jungmeisteris, see at least: “one or more of sentiment analysis logic 124, feedback aggregation logic 220, or autoencoder 240 or any subset of any of those logics) may be implemented at least in part as one or more machine learning algorithms. For instance, autoencoder 240 may be understood as a type of artificial neural network used to produce encodings (e.g., vectors) representative of features in a set of data [i.e. comprises encoding the dialogue history and the relevant content into a shared representation space] in an unsupervised manner. In general, autoencoder 240 may include one or more machine learning models for dimensionality reduction of text and one or more machine learning models for reconstructing (generating a representation close to the original text from the reduced encoding)” [0039] and “FIGS. 3-4D illustrate exemplary user interfaces through which a variety of input can be entered by the user of device 150. Web server 140 (or a component of system 110) may extract, from the input data, various information about the input text and/or other values including, in an exemplary embodiment, the actual content of the query input string (in some embodiments, in a tokenized format). In addition to freeform text data entered, the data obtained from web server 140 may also variously include information sufficient to identify the user, such as a user ID (if the user is logged in or otherwise authenticated) or a session ID, as well as the particular text input by the user. The input text is transmitted from the web server to customer support system 110 and processed by the autoencoder 240 [i.e. the process of fusing the relevant content with the dialogue history comprises encoding the dialogue history and the relevant content]” [0069] and “text classification or topic extraction from text is performed by one or more supervised or unsupervised algorithms … this may involve feature extraction from freeform text e.g., to generate vectors for words or sentences (step 512). Topic classifiers are defined in advance, and stored in memory 210 as thematic response data 235. As examples, some predefined topics may include: “user account”, “payment”, “booking”, “cancellation”, “confirmation”, and so on. The specific topics can be generally understood to be specific to the purpose and use of the website or application. In some embodiments, the predefined topics may correspond to products, ticket topics or identifiers, or other delimiters created by a backend customer support system. For each of these topics, one or more machine learning models may be applied to detect patterns in the input freeform text that suggest relevance [i.e. encoding the relevant content]” [0072]). This known technique is applicable to the method of Bright in view of Baligar as they both share characteristics and capabilities, namely, they are directed to utilizing a deep learning system.
It would have been recognized that applying the known technique of the process of fusing the relevant content with the dialogue history comprising encoding the dialogue history and the relevant content into a shared representation space, as taught by Jungmeisteris, to the teachings of Bright in view of Baligar would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such references into similar methods. Further, adding the modification of the process of fusing the relevant content with the dialogue history comprising encoding the dialogue history and the relevant content into a shared representation space, as taught by Jungmeisteris, into the method of Bright in view of Baligar would have been recognized by those of ordinary skill in the art as resulting in an improved method that would provide a more dynamic, accurate, and timely customer sentiment analysis (Jungmeisteris, [0005]).
Claims 4 and 7-8 are rejected under 35 U.S.C. 103 as being unpatentable over Bright, in view of Baligar, in further view of Lindgren et al. (US 2023/0385903 A1), hereinafter Lindgren.
Regarding claim 4, Bright in view of Baligar teaches the method of claim 1.
Bright in view of Baligar does not explicitly disclose the relevant content being selected based on a similarity score between the retrieved content and the current dialogue context.
Lindgren, however, teaches intelligent context-based personalized product recommendation (i.e. abstract), including the known technique of the relevant content being selected based on a similarity score between the retrieved content and the current dialogue context (Lindgren, see at least: “The data analysis and recommendation engine 160 may use the input response (e.g., quiz/questionnaire responses) by the user to create a mapping of the requirements 609. The environmental conditions may be extracted for personalized recommendation by using the locational input provided by the user 610. Using all user responses and the environmental factors a requirement vector may be created 611. All available products are represented as vectors and a similarity score of all products may be calculated 612 using the requirement vector [i.e. between the retrieved content and the current dialogue context]. The top product or products with the highest similarity score [i.e. wherein the relevant content is selected based on a similarity score] in different product categories may then be presented (e.g., displayed on a webpage) to the customer (i.e., system user) as a product recommendation 613” [0098]). This known technique is applicable to the method of Bright in view of Baligar as they both share characteristics and capabilities, namely, they are directed to intelligent context-based personalized product recommendation.
It would have been recognized that applying the known technique of the relevant content being selected based on a similarity score between the retrieved content and the current dialogue context, as taught by Lindgren, to the teachings of Bright in view of Baligar would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such references into similar methods. Further, adding the modification of the relevant content being selected based on a similarity score between the retrieved content and the current dialogue context, as taught by Lindgren, into the method of Bright in view of Baligar would have been recognized by those of ordinary skill in the art as resulting in an improved method that would improve recommendations (Lindgren, [0076]).
Regarding claim 7, Bright in view of Baligar teaches the method of claim 1.
Bright in view of Baligar does not explicitly disclose the final response being generated using a decoder component of the deep learning model conditioned on the fused representation.
Lindgren, however, teaches intelligent context-based personalized product recommendation (i.e. abstract), including the known technique of the final response being generated using a decoder component of the deep learning model conditioned on the fused representation (Lindgren, see at least: “the training process begins by creating training datasets using retrieved subsets of user response data and product information 1101. The subsets of data may be pre-processed subsets ready to be input into a neural network. The training datasets may then be fed into a recurrent neural network 1102 comprising one or more hidden recurrent layers which may embed the input training datasets and extract weighted features which define the datasets as it passes through each of the one or more hidden recurrent layers. The hidden recurrent layers constitute an encoder. The next step is to train the encoder to learn the features of the inputted subset of user response data and the subset of product information such that they may be encoded into a user requirement vector and a product vector, respectively, existing within an encoded feature space 1103. As a next step, product vectors are extracted from the feature space and fed into a decoder to determine one or more beauty products to recommend 1104 [i.e. wherein the final response is generated using a decoder component of the deep learning model conditioned on the fused representation]. The decoder may comprise one or more fully connected layers. At this point in the training cycle the recommended beauty products may be checked for usefulness as well as validating model performance using a pre-determined criteria for success 1105” [0103]). This known technique is applicable to the method of Bright in view of Baligar as they both share characteristics and capabilities, namely, they are directed to intelligent context-based personalized product recommendation.
It would have been recognized that applying the known technique of the final response being generated using a decoder component of the deep learning model conditioned on the fused representation, as taught by Lindgren, to the teachings of Bright in view of Baligar would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such references into similar methods. Further, adding the modification of the final response being generated using a decoder component of the deep learning model conditioned on the fused representation, as taught by Lindgren, into the method of Bright in view of Baligar would have been recognized by those of ordinary skill in the art as resulting in an improved method that would improve recommendations (Lindgren, [0076]).
Regarding claim 8, Bright in view of Baligar teaches the method of claim 1.
Bright in view of Baligar does not explicitly disclose the deep learning model being fine-tuned using reinforcement learning based on user satisfaction feedback.
Lindgren, however, teaches intelligent context-based personalized product recommendation (i.e. abstract), including the known technique of the deep learning model being fine-tuned using reinforcement learning based on user satisfaction feedback (Lindgren, see at least: “The consumer can provide feedback to system 1200 both directly and indirectly … Feedback data [i.e. based on user satisfaction feedback] may be used to train an encoder/decoder model (referring to FIG. 10 above) for product recommendations. In some implementations, the encoder/decoder model may be trained using reinforcement learning wherein the encoder/decoder model [i.e. wherein the deep learning model is fine-tuned using reinforcement learning] is rewarded for making product recommendations that maximize positive user experience” [0081]). This known technique is applicable to the method of Bright in view of Baligar as they both share characteristics and capabilities, namely, they are directed to intelligent context-based personalized product recommendation.
It would have been recognized that applying the known technique of the deep learning model being fine-tuned using reinforcement learning based on user satisfaction feedback, as taught by Lindgren, to the teachings of Bright in view of Baligar would have yielded predictable results because the level of ordinary skill in the art demonstrated by the references applied shows the ability to incorporate such references into similar methods. Further, adding the modification of the deep learning model being fine-tuned using reinforcement learning based on user satisfaction feedback, as taught by Lindgren, into the method of Bright in view of Baligar would have been recognized by those of ordinary skill in the art as resulting in an improved method that would improve recommendations (Lindgren, [0076]).
Response to Arguments
Rejections under 35 U.S.C. §101
Applicant argues that, as amended, claim 1 is not directed to an abstract idea, but rather to a specific improved technological process for integrating an intelligent conversational agent with online information retrieval to facilitate enhanced e-commerce interactions. The claim recites a series of concrete, technical steps performed by a deep learning model on a computing system: (i) using search engines to retrieve external, factual information from the Internet; (ii) analyzing and identifying relevant product-related content (price, reviews, features); and (iii) dynamically fusing that externally retrieved content with the ongoing user dialogue history to generate a contextually appropriate final response. These steps are inextricably tied to computer technology (including web search engines and a trained deep neural-network-based conversational model) and cannot be practically performed in the human mind. Far from a mere "method of organizing human activity," the claimed method improves the functioning of computerized conversational assistants and e-commerce systems. These features are disclosed, for example, in Applicant's originally filed specification throughout, are illustrated in FIGs 1-3, and in specification references (Spec. refs.) [l] - [11] cited in corresponding provisional application 63/515,024, the disclosure of which is incorporated in the present application in its entirety, by reference (Remarks, page 5).
Examiner respectfully disagrees. The claims recite the concept of searching for and providing relevant product information which fall within the “Certain Methods of Organizing Human Activity” groupings of abstract ideas, enumerated in MPEP 2106.04(a). Assistance and commerce are not technical fields and merely automating them does not make them technical fields. Additionally, merely utilizing computer technology such as web search engines and a trained deep neural-network-based conversational model to apply the abstract idea fails to reflect an improvement in the functioning of a computer or an improvement to another technology or technical field. For instance, neither the web search engines or the trained deep neural-network-based conversational model themselves are improved. The additional elements do no more than generally link the use of the judicial exception to a particular technological environment or field of use (such as computers or computing networks). Employing well-known computer functions to execute an abstract idea, even when limiting the use of the idea to one particular environment, does not integrate the exception into a practical application. Accordingly, the claims are directed to an abstract idea.
Applicant further argues that, even assuming, arguendo, that claim 1's method is deemed to recite an abstract idea (e.g., providing product information and recommendations, as the Office Action suggests), the claim as a whole integrates any such idea into a practical technical application. The claimed invention is rooted in computer technology and addresses a specific technical challenge in the field of artificial intelligence and e-commerce: namely, how to enable a conversational AI platform (e.g., a voice-based socialbot) to provide accurate, personalized product recommendations and information by leveraging vast external data sources in real time. The Office Action acknowledges that the claim involves use of a "model", "search engines," and a "socialbot" on an e-commerce platform. These are not generic concepts of doing business or fundamental economic practices; they are specific computing tools and techniques operating in concert to achieve a technical outcome - an automated system for advanced retrieval and contextual processing of information to enhance a user's online shopping experience (Remarks, pages 5-6).
Examiner respectfully disagrees. Providing accurate, personalized product recommendations and information by leveraging vast external data sources in real time are not technical improvements as they do not improve the technology itself. Additionally, this solution is to a business problem, not a problem rooted in technology. Merely utilizing machine learning, search engines, and a socialbot to implement the abstract idea amounts to nothing more than mere instructions to implement or apply the abstract idea on a generic computing hardware (or, merely use a computer as a tool to perform an abstract idea) and does no more than generally link the use of the judicial exception to a particular technological environment or field of use (such as computers or computing networks). Furthermore enhancing a user’s shopping experience is not a technical improvement. Accordingly, the claims are not integrated into a practical application.
Applicant further argues that the closest human analogy - a sales representative manually looking up product information and making suggestions - is fundamentally different
from the claimed computerized process. The invention's deep learning model (e.g., a neural
network-based conversational agent) can process vast amounts of data (internet-scale factual
knowledge) and fuse it with context far beyond human mental capability or traditional static
chatbots. This yields a qualitatively improved machine functionality: the ability for a dialog system to provide factually accurate, context-aware product recommendations and answers in a
conversational manner. The Specification highlights that conventional e-commerce chatbots lack
comprehensive knowledge and struggle with complex queries (see Spec. [0003]-[0005]). The
claimed solution employs a specialized deep learning architecture (see, e.g., Spec. [0029]) that
significantly enhances the chatbot's operation by automatically retrieving up-to-date external information and injecting it into the dialog. Such integration of a search module with a neural
conversational model is a technological improvement in conversational Al, as it increases the
factual accuracy and relevance of the responses (Spec. [0027]; Spec. [0049]-[0054]). In other
words, the claim is directed to a specific improved conversational system, not an abstract business or human-organizational practice (Remarks, page 6).
Examiner respectfully disagrees. The recited claims within the “Certain Methods of Organizing Human Activity” groupings of abstract ideas as they encompass advertising, and marketing or sales activities. Examiner has not stated that the recited claims fall within mental processes as Applicant appears to be arguing. Additionally, the machine learning algorithm itself, as claimed, is not improved, the data provided to it improved; improving data is not a technical improvement. Accordingly, the claims are not integrated into a practical application.
Applicant further argues that the Office Action contends (at pp. 5-9) that claim 1 falls within the category of "certain methods of organizing human activity" (e.g. marketing or sales activities) and thus recites an abstract idea. Applicant respectfully disagrees. The focus of claim 1 is on how a computer-based conversational system performs information retrieval and processing (using particular technological means such as search engine interfaces and a trained deep learning model) to generate useful output. The claim does not recite a mere scheme of selling or marketing products; instead, it recites a technological solution (a multi-step data processing method executed by a machine) to improve the capabilities of a conversational agent. The steps of querying search engines for external product knowledge and algorithmically fusing that knowledge with a user's dialogue context are specific computer-implemented operations facilitating a more effective human-computer interaction. This is analogous to improvements in computer-functionality recognized as patent-eligible in precedent (cf. McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299 (Fed. Cir. 2016), patentable rules for lip-synching animation improved computer operation for automated 3D animation; Finjan, Inc. v. Blue Coat Sys., Inc., 879 F.3d 1299 (Fed. Cir. 2018), specialized virus scanning method was patent-eligible) (Remarks, pages 6-7).
Examiner respectfully disagrees. Merely utilizing machine learning, search engines, and a socialbot to implement the abstract idea amounts to nothing more than mere instructions to implement or apply the abstract idea on a generic computing hardware (or, merely use a computer as a tool to perform an abstract idea) and does no more than generally link the use of the judicial exception to a particular technological environment or field of use (such as computers or computing networks). The technology itself is not improved.
Regarding, McRO, the claims “focused on a specific asserted improvement in computer animation, i.e., the automatic use of a particular type” [see McRO, Inc. v. Bandai Namco Games America Inc., 837 F.3d 1299, 120 U.S.P.Q.2d 1091 (Fed. Cir. 2016) page 24]. The claims were found eligible because of a specific improvement in computer animation (i.e. a problem rooted in technology). This is not the case with the claimed invention. Unlike in McRO, the claimed invention fails to reflect an improvement in the functioning of a computer or an improvement to another technology or technical field.
Regarding Finjan, the claims were determined to be an improvement over technology. Finjan describes that “Because security profiles communicate the granular information about potentially suspicious code made available by behavior-based scans, they can be used to protect against previously unknown viruses as well as "obfuscated code" — known viruses that have been cosmetically modified to avoid detection by code-matching virus scans.” This proved a technical solution to a problem rooted in computer technology. This is not the case with the claimed invention. Unlike in Finjan, the claimed invention fails to reflect an improvement in the functioning of a computer or an improvement to another technology or technical field.
Applicant further argues that the Office Action further asserts (p. 8-10) that the additional elements - such as use of search engines, the Internet, and a socialbot - are recited at a high level of generality and merely invoke generic computer functions. However, the Examiner has provided no evidence or citation establishing that the specific combination of features in claim 1 was well-understood, routine, or conventional as of the effective filing date. On the contrary, the modular architecture claimed (involving a deep learning model synergistically operating with internet search and knowledge fusion) represents a non-conventional arrangement that yields unprecedented capabilities, as documented in the scientific literature and the present Specification. For example, Shuster et al., "Language Models that Seek for Knowledge: Modular Search & Generation for Dialogue" (Findings of EMNLP 2022) (Spec. ref. [6]) describe the recent development of dialogue agents that intelligently search the internet and incorporate found knowledge into their responses – an approach which outperformed prior art systems that lacked such modular integration. The claimed invention likewise leverages this innovative paradigm in the context of a retail chatbot, resulting in a system that improves the accuracy and contextual relevance of product recommendations and answers (Spec. [0049]; Spec. [0053]). This is a technical improvement in computer capabilities (a more effective AI agent), not a mere use of computers as a tool for abstract economic concepts (Remarks, page 7).
Examiner respectfully disagrees. The additional elements are insufficient to integrate the abstract idea into a practical application because the additional elements amount to nothing more than mere instructions to implement or apply the abstract idea on a generic computing hardware (or, merely use a computer as a tool to perform an abstract idea) and do no more than generally link the use of the judicial exception to a particular technological environment or field of use (such as computers or computing networks). Improving the data used by the AI agent does not improve the AI technology itself. Additionally, as is described in the MPEP 2106.05(II) (i.e. “Thus, in Step 2B, examiners should: … Re-evaluate any additional element or combination of elements that was considered to be insignificant extra-solution activity per MPEP § 2106.05(g), because if such re-evaluation finds that the element is unconventional or otherwise more than what is well-understood, routine, conventional activity in the field, this finding may indicate that the additional element is no longer considered to be insignificant”), step 2B considers whether additional elements concluded to be insignificant extra-solution activity in Step 2A are more than well-understood, routine, conventional activity in the field. Examiner did not identify any of the additional elements as insignificant extra-solution activity in Step 2A so there weren’t elements to be evaluated in terms of whether they are more than well-understood, routine, conventional activity in the field. Accordingly, the claims are ineligible.
Applicant further argues that when properly considered as a whole, claim 1 does not monopolize an abstract idea but recites a patent-eligible, specific application of artificial intelligence and information retrieval techniques to solve a technological problem in e-commerce platforms. The method is firmly rooted in computer technology (involving a trained deep neural network-based agent, web search engines, and e-commerce servers) and yields a novel, tangible improvement in computer functionality (improved dialog system performance and user satisfaction via enhanced information processing). Therefore, Applicants respectfully request withdrawal of the § 101 rejection (Remarks, pages 7-8).
Examiner respectfully disagrees. The problem being solved is not rooted in computer technology and commerce (even when applied to a computer) is not a field of technology (see MPEP 2106.05(a)).
Additionally, preemption is not the test for eligibility. While preemption is the concern underlying the judicial exceptions, it is not a standalone test for determining eligibility. Rapid Litig. Mgmt. v. CellzDirect, Inc., 827 F.3d 1042, 1052, 119 USPQ2d 1370, 1376 (Fed. Cir. 2016). Instead, questions of preemption are inherent in and resolved by the Subject Matter Eligibility Test. Synopsys, Inc. v. Mentor Graphics Corp., 839 F.3d 1138, 1150, 120 USPQ2d 1473, 1483 (Fed. Cir. 2016); Ariosa Diagnostics, Inc. v. Sequenom, Inc., 788 F.3d 1371, 1379, 115 USPQ2d 1152, 1158 (Fed. Cir. 2015). It is necessary to evaluate eligibility using the Subject Matter Eligibility Test, because while a preemptive claim may be ineligible, the absence of complete preemption does not demonstrate that a claim is eligible. Diamond v. Diehr, 450 U.S. 175, 191-92 n.14, 209 USPQ 1, 10-11 n.14 (1981) ("We rejected in Flook the argument that because all possible uses of the mathematical formula were not pre-empted, the claim should be eligible for patent protection"). See also Return Mail, Inc. v. U.S. Postal Service, -- F.3d --, -- USPQ2d –, slip op. at 34 (Fed. Cir. August 28, 2017); Synopsys v. Mentor Graphics, 839 F.3d at 1150, 120 USPQ2d at 1483; FairWarning IP, LLC v. Iatric Sys., Inc., 839 F.3d 1089, 1098, 120 USPQ2d 1293, 1299 (Fed. Cir. 2016); Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1320-21, 120 USPQ2d 1353, 1362 (Fed. Cir. 2016); Sequenom, 788 F.3d at 1379, 115 USPQ2d at 1158 (See MPEP 2106.04 section I). Accordingly, the claims are ineligible.
Rejections under 35 U.S.C. §103
Applicant argues that neither Bright nor Baligar, alone or in combination, teaches or suggests the full scope of amended claim 1. Bright's disclosure is limited to searching and filtering products within an e-commerce marketplace's own catalog. Bright's "search component" uses semantic queries to find matching products in a marketplace database and may employ a knowledge graph for understanding a user's query (Office Action, p. 13, citing Bright ([0052], [0061], [0056]). Crucially, Bright does not teach using an internet search engine to retrieve external, factual knowledge outside of the product catalog. Instead, Bright's system operates on in-platform data (product listings and associated product attributes already known to the marketplace system - see Bright ([0041], [0061], [0056]). Bright's dialogue manager may ask follow-up questions and use an "AIF 144" component with a knowledge graph and user details (Office Action, p. 13-14), but nowhere does Bright describe leveraging open-web search engines or merging outside factual content into the conversation as required by claim 1. Baligar is not discussed in detail in the Office Action. It is understood to relate generally to e-commerce assistants or voice-based digital assistants. There is no indication that Baligar cures the above-noted deficiencies of Bright (Remarks, pages 8-9).
Examiner respectfully disagrees. Bright discloses that the networked system forms a marketplace that provides functionality via browsers such as Internet Explorer® browser [i.e. using search engines] and is connected to receive data via the internet from third party marketplaces, as well as, that the artificial intelligence framework (AIF) that provides data to the searched databases performs proactive data extraction from multiple sources, using web browsers [i.e. using search engines to look for external and factual knowledge across the internet], to facilitate product searching (see Bright, [0040], [0091], [0060]-[0061], [0041], [0046], and Fig. 1). Accordingly, Bright discloses this feature.
Applicant further argues that Bright fails to disclose the specific "fusion" step of claim 1. While Bright's agent can clarify queries and utilize a knowledge graph to retrieve product data, it does not teach completing a process of fusing retrieved knowledge with the dialogue history to generate a final response in the manner claimed. Bright's system appears to extract information (e.g., using semantic vectors and knowledge graph) to narrow product search results, but it does not detail any machine learning mechanism that integrates retrieved external content into the flow of a conversational response to the user. In contrast, amended claim 1 explicitly requires that the method fuses the externally obtained content with the ongoing dialogue context via a deep learning model to produce a unified final response to the user. This is a distinguishing feature of Applicants' invention that is not taught or suggested in Bright. Baligar is not discussed in detail in the Office Action. It is understood to relate generally to e-commerce assistants or voice-based digital assistants. There is no indication that Baligar cures the above-noted deficiencies of Bright (Remarks, pages 8-9).
Examiner respectfully disagrees. Initially, Examiner points out that the “relevant content” is not recited to include the “external and factual knowledge.” Regardless, Bright discloses this feature, Bright discloses that the artificial intelligence framework (AIF) creates a series of interactive operations to get to the optimal results by asking clarifying questions regarding what the user is looking for and searching for results based on the context of the conversation [i.e. fusing the relevant content with the dialogue history in order to provide a final response] (see Bright, [0054], [0056] and [0061]). The response is based on the ‘fusion’ of the context of the conversation and the content retrieved via search. Additionally, claim 1 does not recite that the ‘fusing’ is done via a deep learning model. Although the claims are interpreted in light of the specification, limitations from the specification are not read into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). Accordingly, Bright discloses this feature.
Applicant further argues that, even considering Bright and Baligar together, the combination would not render the claimed invention obvious. The Office Action has not provided a specific teaching or motivation to modify either reference to arrive at the claimed method. Bright does not recognize any need or benefit of querying open-domain search engines or fusing outside factual knowledge into the conversation; its focus is on the marketplace's internal data. Baligar, while possibly describing a voice interface for shopping, lacks any teaching of using a machine learning model to perform the particular knowledge retrieval and integration steps recited in claim 1. In particular, nothing in Baligar or Bright suggests the technical implementation of a "deep learning" model that retrieves internet-sourced factual content about product aspects and combines it with user dialog context to improve recommendations and Q&A results (Remarks, page 9).
Examiner respectfully disagrees. As detailed in response to the arguments above, Bright discloses the above argued features, Baligar is not cited for the above argued features. Additionally, it would have been obvious to combine the teachings of Bright with Baligar as it would provide accurate, intuitive, and relatively quick interactions in order to instruct the computing devices to perform desired functions (Baligar, Col. 1 Ln. 19-21). Accordingly, the cited references teach the amended claims.
Applicant further argues that the features added to claim 1 through the present amendments further distinguish the claimed invention from the cited art. For example, claim 1 now expressly recites that the external knowledge is retrieved "across the internet" and that the content is fused with the dialogue history to generate the final response. These requirements are nowhere taught in Bright or Baligar. Additionally, claim 1 now makes clear that the goal of these steps is to connect a socialbot (e.g., a voice-controlled conversational agent) with an e-commerce store platform to provide new functions like personalized shopping guidance. Such a holistic integration of a conversational Al's dialog management with real-time web-based information retrieval and an online retail platform is absent from the prior art of record. Indeed, the motivation behind Applicants' invention was to overcome the limitations of prior e-commerce chatbots (Spec. [0003]-[0005]), which, unlike the claimed solution, could not effectively provide personalized, context-aware shopping assistance leveraging the breadth of Internet knowledge (Remarks, pages 9-10).
Examiner respectfully disagrees. As detailed in response to the arguments above Bright discloses that the networked system forms a marketplace that provides functionality via browsers such as Internet Explorer® browser and is connected to receive data via the internet from third party marketplaces, as well as, that the artificial intelligence framework (AIF) that provides data to the searched databases performs proactive data extraction from multiple sources, using web browsers [i.e. across the internet], to facilitate product searching (see Bright, [0040], [0091], [0060]-[0061], [0041], [0046], and Fig. 1). Bright further discloses that the artificial intelligence framework (AIF) creates a series of interactive operations to get to the optimal results by asking clarifying questions regarding what the user is looking for and searching for results based on the context of the conversation [i.e. fusing the relevant content with the dialogue history in order to provide a final response] (see Bright, [0054], [0056] and [0061]). The previous back and forth of the conversation is considered dialog history as it’s previously submitted dialog. Accordingly, the cited references teach the amended claims.
Applicant further argues that for at least the reasons above, neither Bright nor Baligar - whether considered individually or in any combination - teaches or fairly suggests the full scope of claim 1. The Office Action's assertion that the combination would "pick up" all claimed features (see Office Action, pp. 11-15) is not supported by the teachings of the references. Absent hindsight, one of ordinary skill would not have been motivated to modify Bright's marketplace-focused method with the specific internet knowledge integration techniques of the claimed invention. Therefore, assuming arguendo, that one would be motivated to combine these references in the manner suggested by the Office Action, the present invention would not be obvious in view of such a combination. The suggested combination would not result in the presently claimed invention recited above in claim 1. Accordingly, claim 1 is patentable over Bright in view of Baligar. Applicants respectfully request withdrawal of the § 103 rejection. (Remarks, page 10).
Examiner respectfully disagrees. As detailed in response to the arguments above, Bright discloses the above argued features, Baligar is not cited for the above argued features. Additionally, it would have been obvious to combine the teachings of Bright with Baligar as it would provide accurate, intuitive, and relatively quick interactions in order to instruct the computing devices to perform desired functions (Baligar, Col. 1 Ln. 19-21). Accordingly, the cited references teach the amended claims.
Applicant further argues that new claims 2-10 depend from claim 1 and are therefore allowable at least for the reasons claim 1 is allowable, and for the specific features recited therein (Remarks, page 10).
Examiner respectfully disagrees. As detailed in response to the arguments above, claim 1 is not allowable. Accordingly, claims 2-10 are not allowable and are taught by the currently cited references.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
-Vijayaraghavan et al. (US 2020/0302423 A1) teaches searching using speech-to-text algorithms.
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 ARIELLE E WEINER whose telephone number is (571)272-9007. The examiner can normally be reached M-F 8:30-5:00.
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, Maria-Teresa (Marissa) Thein can be reached at 571-272-6764. 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.
/ARIELLE E WEINER/ Primary Examiner, Art Unit 3689
Prosecution Insights