DETAILED ACTION
This action is in response to the initial filing of application no. 18/433,014 on 02/05/2024.
Claims 1- 20 are still pending in this application, with claims 1, 15 and 20 being independent.
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 Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 17 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 17 recites the limitation "the one or more training instances”. There is insufficient antecedent basis for this limitation in the claim.
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 – 6, 11 – 13 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sacheti et al. (US 2024/0354309) (“Sacheti”) in view of HE (US 2012/0254158) and further in view of Paulino et al. (US 2024/0346233) (“Paulino”).
For claims 1 and 20, Sacheti discloses a system comprising one or more processors and memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the operations (Abstract) of: receiving a user query that seeks information about an entity (product) class (lawn tractors, tennis rackets, camping tents, vacuum cleaners) (Fig.2, 202 and Fig.8, 802; [0007] [0032] [0053] [0065]); searching, based at least on the user query, an entity database (supplemental content database comprising supplemental content, Fig.1, 116; [0035]) to acquire supplemental content associated with one or more entities of the entity class (Fig.2, 204, 210, 212, 214 and Fig.8, 804; [0024] [0028] [0033 – 0035] [0065]); generating a prompt based on the user query and the supplemental content (Fig.2, 206, 214 and Fig.8, 806; [0033] [0035] [0036] [0065]); processing the prompt as input, using one or more of the generative models (GLM, Fig.2, 112; [0002] [0024]) (Fig.8, 808; [0036] [0065]); generating a model output from which a response to the user query is derived ([0036] [0065]); and causing the response to the user query to be rendered at an output device (Fig.2, 216; [0037]). Yet, Sacheti fails to teach the following: the entity database is an external entity-summarization database; and the supplemental content comprises summarized description of one or more of the features of one or more entities of the entity class, wherein the external entity-summarization database includes one or more entities each storing a summarized description of one or more features of a respective entity of the entity class, and wherein the summarized description of the respective entity is generated based on summarizing one or more public descriptions that each describe multiple features of the respective entity, using one or more generative models.
However, HE discloses a system and method for generating a product catalog (Abstract), comprising the following: an entity (product) database (product catalog, Fig.1, 130) is an external entity database (The information included in the product catalog is provided by external entities such as merchants, information aggregators and/or websites, Fig.2, 205; [0019 – 0021]); and the entity database further comprises entity information corresponding to one or more entities (products), including one or more public descriptions that each describe multiple features of the respective entity (Fig.2, 210, 215, 220 and 230; [0022 – 0027] [0029 – 0032]).
Additionally, Paulino discloses a system and method for efficiently generating a review summary (Abstract), comprising the following: a summarized description (review summary) for an entity (product) is generated using a generative model (LLM) (An automated review summary request initiates the generation of a review summary., Fig.2, 252, 220, 222, 224 and 226; [0045 – 0051] [0056] [0057] [0059 – 0064] [0067] [0077] [0078] [0080 -0082]); and the summarized descriptions are provided to a data store in association with the entity ([0083 - 0085]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve Sacheti’s invention in the same way that He’s invention has been improved to achieve the following, predictable results for the purpose of storing and providing sufficient information to enable a user to compare and select a variety of products to purchase (Sacheti, [0006] [0007]) (HE, [0003] [0004]): the entity database is further an external entity database which comprises one or more public descriptions (reviews) corresponding to a respective entity.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Sacheti and HE in the same way that Paulino’s invention has been improved to achieve the following, predictable results for the purpose storing and providing sufficient information to enable a user to compare and select a variety of products to purchase, without unnecessarily consuming computing resources (Sacheti,[0006] [0007]) (Paulino, [0001]): the external entity database further comprises summarized descriptions, wherein the summarized descriptions are generated based on the one or more public descriptions using a generative model; and the supplemental content used to generate the prompt further comprises the summarized descriptions.
For claim 2, Sacheti further discloses, wherein the user query seeks a recommendation of one or more entities of the entity class (Sacheti, A user requests a to list and/or compare products, such as biking apparel, vacuums, cars tennis rackets. camping tents, lawn tractors, wherein the a entity corresponds to a specific instance of the product, and the a entity class corresponds to the group of products., Fig.2, 202 and Fig.8, 802; [0004] [0007] [0030] [0032] [0046] [0053] [0065]), and wherein the response to the user query includes a recommendation of one or more particular entities of the entity class (Sachet, Fig.2, 214 and 216, Fig.3, 312, Fig.4, 402 and Fig.8, 808 and 810; [0035 – 0037] [0046] [0052 - 0055] [0065]).
For claim 3, Sacheti, HE and Paulino further disclose, wherein: the external entity-summarization database identifies one or more data sources for the
one or more public descriptions, respectively (Sacheti, The supplemental content database comprises source information (store, vendor, manufacturer) for respective products., [0024] [0035] [0046] [0053] [0060]) (HE, [0011] [0019] [0020]); and the one or more data sources are associated with the summarized description for the respective entity (HE, [0011] [0019] [0020]) (Paulino, [0083 - 0085]).
For claim 4, Sacheti and HE further disclose, wherein the rendered response to the user query presents a particular data source for the recommendation (Sachet, The data source (store, vendor, manufacturer) is displayed in the table., Fig.2, Fig.3, 312, Source; [0035 – 0037] [0043] [0046]) (HE, [0011] [0019] [0020]).
For claim 5, Sacheti and HE further disclose, wherein the rendered response to the user query includes a selectable element (Sacheti, Buy button, Fig.4, 404; [0052 – 0054]) that, when selected, causes the particular data source to be rendered (Sacheti, Fig.5, 502; [0058]) (HE, [0011] [0019] [0020]).
For claim 6, HE and Paulino further disclose wherein the one or more public descriptions include a first public review that reviews the respective entity from a public data source, and a second public review that reviews the respective entity from an additional public data source that is different from the public data source (HE, Fig.2, 210, 215, 220 and 230; [0022 – 0027] [0029 – 0032]) (Paulino, [0045 – 0051] [0056] [0057] [0059 – 0064] [0067] [0077] [0078] [0080 -0082]).
For claim 11, Sacheti and Paulino further disclose wherein the generative model (Sacheti, [0006] [0024] [0042]) (Paulino, [0080]) is trained to summarize reviews from one or more public data sources (Paulino, The LLM model used to generate the review summary has been trained to perform text summarization by being trained to receive a prompt comprising text, e.g. review data, and generate a summary, e.g. review summary, [0019 – 0021] [0080]).
For claim 12, Sacheti, HE and Paulino further disclose, wherein the external entity-summarization database (Sacheti, [0035]) (HE, [0019 – 0027] [0029 – 0032]) (Paulino, ([0080 - 0085]) is updated in real-time based on detection of an update to the reviews published at the one or more public data sources (Paulino, “A website or application service, such as review service 118, associated with item presentation may automatically initiate generation of review summaries, for instance, based on a lapse of a time period, a reception of a review or set of reviews (e.g., upon obtaining a predetermined number of reviews), [0047]).
For claim 13, Sacheti further discloses, wherein searching, based at least on the user query, the external entity-summarization database is performed in response to the user query being classified as a query requesting recommendation of one or more entities for an entity class (Sacheti, [0030] [0032]), or in response to determining that the user query includes one or more keywords.
Claim(s) 7, 8 and 10 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sacheti et al. (US 2024/0354309) (“Sacheti”) in view of HE (US 2012/0254158), and further in view of Paulino et al. (US 2024/0346233) (“Paulino”) and further in view of Chalkley et al. (US 2025/0200115) (“Chalkley”).
For claim 7, the combination of Sacheti, HE and Paulino fails to teach the following, wherein searching, based at least on the user query, the external
entity-summarization database to acquire the summarized description of the entity comprises: generating a first numeric embedding based on the user query and/or context data associated with the user query; and searching the external entity-summarization database using the first numerical embedding, to identify the summarized description of the entity.
However, Chalkley teaches a system and method for providing product recommendation services (Abstract), comprising the following: generating a first numeric embedding (search vector) based on a user query (search query) and context data associated with the user query (set off prior user interactions associated with the search query) (A search vector is generated by applying one or more generative artificial intelligence models to a combination of the search query itself and one or mere sets of prior user interactions stored in the user interaction histories, Fig.1, 112, Fig.2, 202, 202A – 202D and Fig.4, 402; [0043 – 0047] [0058] [0066 – 0069 [0071] [0077]) and searching a database (The database stores item/product listings, item listing, vectors, user interaction histories, etc., Fig.6, 624; [0082]), using to first numeric embedding, to identify entity information (item listing data)(Fig.4, 406; [0031] [0050] [0051] [0078] [0079]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Sacheti, HE and Paulino in the same way that Chalkley’s invention has been improved to achieve the following, predictable results for the purpose of providing e-commerce methods and system (Sacheti, Abstract; [0032]) which accurately identify particular item listings associated with a user's true intent and further provide an enhanced conceptualized search based on enriched categorizations of item listings that are instrumental in assisting the user (Chalkley, [0025]): wherein searching, based at least on the user query, an item listing database, e.g. the external entity-summarization database, to acquire item listing information, e.g. summarized description, of the entity comprises generating a first numeric embedding based on the user query and context data associated with the user query; and searching the external entity-summarization database using the first numerical embedding, to identify the summarized description of the entity.
For claim 8, Sacheti, HE, Paulino and Chalkley further disclose, wherein searching the external entity-summarization database (Sacheti, [0024] [0035]) (HE, [0019 – 0027] [0029 – 0032]) (Paulino, [0023] [0083 - 0085]) (Chalkley, [0031] [0050] [0051] [0078] [0079]) using the first numerical embedding (Sacheti, [0035) (Chalkley, [0031] [0050] [0051] [0078] [0079]) comprises: for each respective summarized description in the external entity-summarization database (Sacheti, [0024] [0035]) (HE, [0019 – 0027] [0029 – 0032]) (Paulino, [0023] [0083 - 0085]) (Chalkley, [0031] [0050] [0051] [0078] [0079]), calculate a distance between the first numeric embedding and a numeric embedding for the respective summarized description (Paulino, [0023] [0083 - 0085]) (Chalkley, [0006] [0064]); and selecting one summarized description from the external entity-summarization database that satisfies a distance threshold as the summarized description of the entity (Paulino, [0023] [0083 - 0085]) (Chalkley, [0006] [0064]).
For claim 10, the combination of Sacheti, HE, and Paulino fails to teach wherein the external entity-summarization database includes a respective numeric embedding for the summarized description for the respective entity, stored in association with the summarized description for the respective entity.
However, Chalkley teaches a system and method for providing product recommendation services (Abstract), comprising the following: generating an item listing vector/embedding of values for an item by applying one or more generative artificial intelligence models to an items description ([0050] [0051] [0056] [0058]); and storing the item listing vector in association with item listings in a database (The enhanced conceptual platform database(s) stores item/product listings, item listing, vectors, user interaction histories, etc., Fig.6, 624; [0082]), using to search vector, to identify item listing data (Fig.4, 406; [0031] [0050] [0051] [0078] [0079]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Sacheti, HE and Paulino in the same way that Chalkley’s invention has been improved to achieve the following, predictable results for the purpose of providing e-commerce methods and system (Sacheti, Abstract; [0032]) which accurately identify particular item listings associated with a user's true intent and further provide an enhanced conceptualized search based on enriched categorizations of item listings that are instrumental in assisting the user (Chalkley, [0025]): a first numeric embedding (vector) is further generated for the entity (product/item) information stored in the external entity-summarization database, wherein the entity information comprises the summarized description; and storing the first numeric embedding in association with the entity information, which includes the summarized description, in the external entity-summarization database.
Claim(s) 9 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sacheti et al. (US 2024/0354309) (“Sacheti”) in view of HE (US 2012/0254158), and further in view of Paulino et al. (US 2024/0346233) (“Paulino”), and further in view of Chalkley et al. (US 2025/0200115) (“Chalkley”) and further in view of Pathak et al. (US 2024/0394479)(“Pathak”).
For claim 9, the combination Sacheti, HE, Paulino and Chalkley further discloses, wherein the user query is received during an interactive chat (Sacheti, [0006] [0065]). Yet, the combination of Sacheti, HE and Paulino fails to teach that the context data associated with the user query includes at least a portion of a chat history of the interactive chat that precedes the user query.
However, Pathak discloses a system and method for interacting with a machine-trained language model (Abstract), wherein a prompt generated for a dialogue system used in conjunction with a shopping-related application ([0052]) comprises both a query and context information comprising at least a portion of a dialogue history of the interactive dialogue that precedes the user query (A user’s input query with targeted context information is used to generate the prompt. The targeted context information comprises a portion of the dialogue preceding the query, Fig.15, 1504 – 1512; [0042] [0044 - 0046] [0055] [0056] [0062] [0128]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Sacheti, HE, Paulino and Chalkley in the same way that Pathak’s invention has been improved to achieve the following predictable results to achieve the following, predictable results for the purpose of providing e-commerce methods and system (Sacheti, Abstract; [0032]) which accurately identify particular item listings associated with a user's true intent and further provide an enhanced conceptualized search based on enriched categorizations of item listings that are instrumental in assisting the user (Chalkley, [0025]): the context data associated with the user query further includes at least a portion of a chat (dialogue) history of the interactive chat that precedes the user query.
Claim(s) 14 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sacheti et al. (US 2024/0354309) (“Sacheti”) in view of HE (US 2012/0254158), and further in view of Paulino et al. (US 2024/0346233) (“Paulino”) and further in view of MOMO et al. (US 2025/0200089) (“MOMO”).
For claim 14, the combination of Sacheti, HE and Paulino fails to teach, wherein the prompt based on the user query and the summarized description of one or more features of the entity includes an instruction to generate a response to the user query utilizing the external entity-summarization database if no satisfying response is generated using inherent knowledge of the generative models.
However, MOMO discloses a system and method for the purpose of generating a search prompt based on input involving a user of a computing device (Abstract, wherein a prompt (PT2) based on a user query and search results retrieved from a database ([0112]) comprises instructions to generate a response to a user query utilizing the search results (ANS2) ([0112] [0114]) if no satisfying response is generated using inherent knowledge of a generative model ([0090 – 0099] [0106 – 0108] [0115]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Sacheti, HE and Paulino in the same way that MOMO’s invention has been improved to achieve the following predictable results for the purpose of improving the accuracy and efficiency of the e-commerce system (Sacheti, Fig.3 and Fig.4; [0005 – 0007]) while reducing resource consumption (preventing unnecessary database querying) in identifying product listings associated with a user’s true intent: the query classifier (Sacheti, [0032] [0033]), which generates prompts for the generative model (Sacheti, [0032] [0033]), further generates a prompt which instructs the generative model (Sacheti, [0002] [0024]) to generate a response using inherent knowledge; and the query classifier further generates the including instruction to generate a response to the user query utilizing search results (Sacheti, [0029] 0032] [0033]), e.g. search results from the external entity-summarization database, if no satisfying response is generated using inherent knowledge of the generative models.
Claim(s) 15, 18, 18, 9 is/are rejected under 35 U.S.C. 103 as being unpatentable over Poliak et al. (US 2024/0420208) (“Poliak”) in view of Wu (US 2024/0289823) and further in view of Daultani et al. (US 2020/0134011) (“Daultani”).
For claim 15, Poliak discloses a method implemented using one or more processors (Abstract), the method comprising: receiving a user query that seeks information about an entity class (class of products, e.g. hiking boots) (Fig.2A,122 and 140, Fig.2B, “End User Query Received”, Fig.3, 314 and Fig.4, 404; [0064] [0065] [0067] [0070]); generating, based at least on the user query, a query-based embedding that semantically represents at least the user query (Fig.2A, “Get embedding from end user prompt text”, 182, Fig.2B, “Embed Conversation Context”, and Fig.3B, 316 – 318; [0065 – 0067]); searching an external entity (product) database (database comprising product catalog information, including product descriptions, from external customers/businesses, Fig.1, 144, Fig.2A, 144 and Fig.2B, 144; [0050] [0059 - 0061]) using the query-based embedding to identify an entity (product) description embedding that matches the query-based embedding (Fig.2A, “Look up list of products”, Fig.2B, “Product are vector embedded and stored in the PVDB/Requested closest products”, and Fig.3, 320 and 322) ([0064] [0065] [0067]), wherein the external entity database stores one or more description embeddings each generated from a respective one of the one or more descriptions (Fig.3A, 302 – 312; [0059] [0066]); processing the description entity embedding, using a generative model (Fig.2A, 180 and 184, Fig.2B, 184, Fig.3C, 326 and 328), to generate a response to the user query ([0065] [0068] [0069]), wherein the response includes a recommendation to a particular entity that the description embedding corresponds (Fig.3C, 330 and Fig.4, 406 and 408; [0069] [0070]); and causing the response to be rendered (Fig.3C, 330 and Fig.4, 406 and 408; [0069] [0070]).
Yet, Poliak fails to teach the following: the one or more descriptions are summarized descriptions; the one or more description embeddings are summarized description embeddings generated from the summarized descriptions; and the external entity database is an entity summarization database which stores the one or more summarized descriptions each for the distinct entity with the one or more summarized description embeddings.
However, Daultani discloses a summary generating device and method (Abstract), comprising the following: an entity (product) summary description is generated for an entity description (A summary is generated for a product description or a plurality word of mouth reviews for a product., [0058] [0061] [0069] [0076] [0083] [0094] [0116] [0117]); and the entity description and the entity summary description are stored in a product database (Fig.5, Description and Summary; [0045 – 0047]).
Additionally, Wu discloses a system and method for processing product information (Abstract), comprising the following: an external entity (product) database (data store comprising product information received from different merchants) (Fig.4, 416; [0033] [0040] [0041] [0063] [0067] [0068] [0070]) stores entity descriptions ([0079 – 0085]) and entity description embeddings ([0099]).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve Poliak’s invention in the same way that Daultani’s invention has been improved to achieve the following, predictable results for the purpose of reducing time and/or resources used in searching and processing stored entity (product) information: one or more entity summary descriptions are further generated for each of the one or more entity descriptions; and both the one or more entity summarized descriptions and the one or more entity descriptions are stored as product information in the external entity database.
Additionally, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Poliak and Daultani in the same way that Wu’s invention has been improved to achieve the following, predictable results for the purpose of leveraging a machine-learning model to improve a product recommendation system, wherein machine learning model recognizes, summarizes, predicts and generates text based on knowledge gained from massive datasets (Poliak, [0003]) (Wu, [0008]): the product information stored in the external entity database, including the entity descriptions and summarized entity descriptions, are further encoded to generate product information embeddings, including entity description embeddings and summarized entity description embeddings; and the product information stored in the external entity database further comprises natural language data and embeddings of the natural language data, including summarized entity descriptions and summarized entity description embeddings.
For claim 18, Poliak, Daultani and Wu further disclose, wherein the response to the user query includes an identifier of a data source associated with the summarized description embedding that matches the query based embedding (Poliak, REI as identifier of data source., Fig.4, 406; [0067- 0070]) (Daultani, [0058] [0061] [0069] [0076] [0083] [0094] [0116] [0117]) (Wu, [0033] [0040] [0041] [0063] [0067] [0068] [0070] [0079 – 0085] [0099]).
For claim 19, Poliak and Daultani further disclose wherein the one or more summarized descriptions are generated based on processing descriptions from one or more public data sources (Poliak, Fig.3A, 302 – 312; [0059] [0066]) (Daultani, Public reviews are the descriptions from one or more data sources., [0058] [0061] [0069] [0076] [0083] [0094] [0116] [0117]).
Claim(s) 16 and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Poliak et al. (US 2024/0420208) (“Poliak”) in view of Wu (US 2024/0289823) and further in view of Daultani et al. (US 2020/0134011) (“Daultani”) and further in view of Liu et al. (“Conversational Recommender System and Large Language Model Are Made for Each Other in E-commerce Pre-sales Dialogue”) (“Liu”).
For claim 16, the combination of Poliak, Wu and Daultani fails to teach the following, wherein the generative model is fine-tuned using one or more
training instances, and wherein the one or more training instances include a first training
instance that includes: a first prompt that includes a first instruction to use a summarized description from the external entity-summarization database to generate recommendations for a first user query that seeks recommendation of one or more entities for a particular entity class, as first training instance input; and a first ground truth response that includes a recommendation to a particular entity that belongs to the particular entity class, as first training instance output.
However, Liu discloses a system and method for providing a conversational recommender system and large language model to perform e-commerce (Abstract), comprising the following: a large language model (LLM) is fine-tuned using one or more training instances (I. Introduction; 3 Method, 3.1 Overview, 3.2 LLMs for E-Commerce pre-sales dialogue, pg. 9587 - 9589), wherein the one or more training instances includes a first training instance that includes: a first prompt that includes a first instruction to use a description to generate recommendations for a first user query that seeks recommendation of one or more entities (product) for a particular entity (product) class, as first training instance input (Figure 3; F. Examples of Fine-tuning LLMs; Table 7 pg. 9591, 9592, 9604); and a first ground truth response that includes a recommendation to a particular entity that belongs to the particular entity class, as first training instance output (Figure 3; F. Examples of Fine-tuning LLMs; Table 7 pg. 9591, 9592, 9604).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Poliak, Wu and Daultani in the same way that Liu’s invention has been improved to achieve the following, predictable results for the purpose of enhancing accuracy of conversational information retrieval for commerce using LLMs which are trained on a large amount of data which enables people to interact with them using natural language (Poliak, [0002]) (Liu, 1. Introduction, pg. 9587): the generative model is further fine-tuned using one or more training instances, and wherein the one or more training instances include a first training instance that includes a first prompt that includes a first instruction to use a description, e.g. summarized description from the external entity-summarization database, to generate recommendations for a first user query that seeks recommendation of one or more entities for a particular entity class, as first training instance input; and a first ground truth response that includes a recommendation to a particular entity that belongs to the particular entity class, as first training instance output.
For claim 17, the combination of Poliak, Wu and Daultani fails to teach the following, wherein the one or more training instances include a second
training instance that includes: a second prompt that includes a second instruction to generate recommendations for a second user query not seeking recommendation, as second training instance input; and a second ground truth response that includes a response to the second user query, as second training instance output.
However, Liu discloses a system and method for providing a conversational recommender system and large language model to perform e-commerce (Abstract), comprising the following: a large language model (LLM) is fine-tuned using one or more training instances (I. Introduction; 3 Method, 3.1 Overview, 3.2 LLMs for E-Commerce pre-sales dialogue, pg. 9587 - 9589); the one or more training instances include a second prompt that includes a second instruction to generate recommendations for a second user query not seeking recommendation (The second prompt is a prompt to generate a dialogue turn) as second training instance input (Figure 3; F. Examples of Fine-tuning LLMs; Table 8, pg. 9591, 9592, 9605); and a second ground truth response that includes a response to the second user query, as second training instance output (Figure 3; F. Examples of Fine-tuning LLMs; Table 8, pg. 9591, 9592, 9605).
Therefore, it would have been obvious to one of ordinary skill in the art at the time of applicant’s filing to improve the invention disclosed by the combination of Poliak, Wu and Daultani in the same way that Liu’s invention has been improved to achieve the following, predictable results for the purpose of enhancing accuracy of conversational information retrieval for commerce using LLMs which are trained on a large amount of data which enables people to interact with them using natural language (Poliak, [0002]) (Liu, 1. Introduction, pg. 9587): the generative model is further fine-tuned using one or more training instances; and the one or more training instances include a second prompt that includes a second instruction to generate recommendations for a second user query not seeking recommendation, as second training instance input, and a second ground truth response that includes a response to the second user query, as second training instance output.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SONIA L GAY whose telephone number is (571)270-1951. The examiner can normally be reached Monday-Friday 9-5 ET.
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, Daniel Washburn can be reached at 571-272-5551. 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.
/SONIA L GAY/Primary Examiner, Art Unit 2657