DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 6/30/2026 has been entered.
Response to Amendment
In response to the final action dated 4/29/20226, the applicant has filed a request for continued examination, filed 6/30/2026, amending claims 1 and 11, while arguing to traverse the prior art and 101 rejections. Applicant’s arguments have been fully considered and determined persuasive with respect to the prior art rejections, but not with respect to the 101 rejections for the reasons explained in the response to arguments.
Response to Arguments
In what follows applicant’s arguments will be addressed in the order presented, but arguments directed at similar topics will be grouped together.
Following a broad overview of the latest amendments (page 8 ¶ 1), and a copy of the amended claim 1 (page 8 2nd ¶ and page 9 1st ¶), on page 9 the 2nd ¶ last 5 lines, it is argued that the newly added limitation pertaining to “Chain-of-Searches” “defines a specific, structure computational workflow that proceeds through five distinct stages, namely split, selection, sorting, search, and supplementing, each of which operates on electronic documents and produces outputs that feed into the next stage. This is a precise computational mechanism, not a mental process”. On page 9 the last ¶ lines 2+ on the same topic it is asserted: “Cos” “requires a server to coordinate five computationally distinct stages” “A human who divides a compound question into sub-questions and answers them sequentially does not implement a Cos technique through five computationally coordinated stages, does not maintain a sorted sequence of derived queries based on document counts, and does not perform iterative sequential appending of results until a final answer coverages. These are computational steps that have no mental analogy”. Page 10 the 2nd ¶ last 5 lines it is asserted: “The CoS technique, as a whole, requires the server to coordinate the split, selection, sorting, search, and supplementing stages, where intermediate results move between stages and the appending of each intermediate result to the next derived query continues sequentially through the entire sorted sequence until the final search result is obtained”.
Respectfully a human can easily count a number of documents that are retrieved per each sub-query, and based on that can sort the sub-queries before appending their results by way of giving precedence to higher ranked results or arrange the results in the appending according to their rank in the sorting. Furthermore, in the appending the human responsible for the appending will have to tailor combining them in a way that addresses the query from which all the sub-queries were derived by providing appropriate connections between responses to each sub-query, otherwise it will result in disconnected statements. That connection basically mimics what is defined above as “intermediate result[s]”. Furthermore if one treats “Cos” as a method responsible for all the steps above (“split” “selection” “sorting” “search” “supplementing stages”), the limitations as drafted do not provide any meaningful limits on how “Cos” addresses each specific operation to make it patent eligible, as they are recited at a very high level of generality.
On page 9 the third ¶ lines 3-5 it is recited: “A human” “does not assign numerical weights to the outputs of a syntactic parsing operation and a separate semantic reasoning operation, combine those weighted outputs through a learned algorithm, and then produce derived queries whose relevance scores are determined by those numerical weights”; Page 11 the first ¶ last 5 lines: “The claims as amended now expressly ties the weighted combination to the combination of the syntactic parsing output and the semantic reasoning output, confirming that the operation functions as an algorithmic combiner of two computationally distinct analyses”; page 11 last 3 lines: “The weights govern how the syntactic and semantic analyses are combined to produce derived queries, and they also determine the relevance scores assigned to each derived query”; Page 12 ¶ 2 lines 8-11: “This three-part process, in which two independent analyses each produce outputs are then combined through learned numerical weighing, cannot be performed in the human mind”; Page 14 ¶ 1 lines 1+ : “it learns optimal numerical weights from prior query-response cycles, applies those weights to combine syntactic and semantic analysis, and uses the resulting weighted scores to govern query derivation”; Page 14 ¶ 3 lines 5+: “Instead of applying fixed decomposition rules that may poorly handle semantically ambiguous or syntactically complex queries, the server uses learned weights to balance syntactic structure against semantic intent in a way that improves the quality and independence of the derived queries”; page 15 last ¶ last 4 lines: “The amended claim now requires that the split stage produce derived queries by combining the outputs of two specific computational analyses through a learned numerical weighting mechanism. This is not a description of an abstract search concept implemented on a generic server”; page 16 the ¶ before last lines 4-6: “The weights are not fixed; they are self-learned, meaning the system updates them based on prior query-response performance”.
All these arguments appear to somehow imply the “weighting” employed to be patent eligible because somehow humans would not be able to assign specific numerical weights to different sub-queries of a larger query with implications on each sub-query’s responses, and/or humans would not be able to dynamically on the fly adjust weights. Regardless of whether or not a weight is obtained dynamically (being “self-learned”) and/or specific numerical values assigned to them (which certainly a human is capable and always intuitively does take that action), the action of “weighting” and anything associated with it in post processing is considered “characterizing data gathering steps as insignificant extra-solution activity” (“Bilski, 561 US 593”). This was also pointed out in the previous action and in response the applicant has asserted on page 17 lines 7-9: “Bilski applies to cases where the additional computational step is tacked on to an otherwise abstract method without contributing to the claimed result”. Respectfully a weighting operation which does not “contribute” to the “claimed result”, will become a redundant operation which will result in a 112(b) rejection.
Page 10 ¶ 3: “A human performing a mental analogy of dividing a question and answering sub-questions does not apply the result of each answer as a formal computation constraint that governs and narrows the retrieval operation for the next sub-question”; Page 10 ¶ 4 lines 3+: “each intermediate result generated after appending narrows the retrieval operation for the next derived query in the sorted sequence, producing measurable technical effects of reduced computational time and improved result relevance on the operation of the server” “A human who uses the answer to one sub-question to inform thinking about the next does not implement a formal computational constraint that algorithmically narrows the retrieval space”; page 15 ¶ 1 lines 2-4: “intermediate results” “as part of the CoS technique, acts as a constraint for granular responses, which results in a focused and faster retrieval of the final result”; page 15 ¶ 2, lines 7-8: “This corresponds to reduced computational time and cost, as expressly disclosed in the specification at paragraph [0032]”.
Respectfully the “computational time” as briefly mentioned in Sp. ¶ 0032 2nd column lines 12 and 19+ (i.e., with reference to “intermediate query-result pair” “each derived query is resolved separately and sequentially, which not only reduces computational time and cost in resolving complex queries, but also improves relevance of the end results in a search”), is merely expression of a goal or a claim to performance of the model; i.e., it is merely based on an assumption that this procedure will lead to more accurate result. It was not supported by any specific measurement of e.g., server CPU usage and/or memory usage conducting a search retrieval with and without applying the techniques here. Reduction of “computational time” should be derived from claim limitations as drafted. It is unclear how breaking up an original query to plurality of “sub-queries” by “semantic” and “syntactic” analysis coupled with “weight” calculations and assignment and further based on adding further “intermediate” steps of generating “query-answer pairs” could obtain results faster than avoiding all those steps and directly finding a response to the original query. The claim limitations as drafted do not point out on any steps which could imply faster search times, and/or smaller memory requirements.
As regards to a human not being able to narrow scope of one sub-query depending on a preceding sub-query, that is completely inaccurate. When the two sub-queries are part of one larger query, if the human does not connect the responses to each sub-query in drafting the overall response, the result will be two separate unrelated responses which will not address the main query. The draft of a final response by making connections between the sub-query responses will require correlations between them which inherently imposes constraints between them with resulting of narrowing of each sub-query response. For example in the example provided in the rejection, i.e., 1) when are you going on a vacation?, and 2) which theme parks will you take your children??, let us assume that on the days associated with the vacation, a certain theme park will be closed; then a response to the second sub-query will be narrowed by excluding that theme park among all the available theme parks from consideration.
Pages 18-23 discuss the previous prior art rejections.
Due to the latest amendments, the said rejections are overcome.
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, 3-11, 13-20 stand rejected:
Claims 1 and 11 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The claims recite a search retrieval method (claim 1) and system (claim11) in which a “search query” is broken into smaller “derived queries”, each with a smaller “length” than the “search query”, but are consistent in “intent” with it. The “two more derived queries are derived by syntactic parsing and semantic reasoning based on a self-learned weighted combination operation” which “includes” “combining an output of the syntactic parsing of the user search query with an output of the semantic reasoning of the user search query” “and assigning numerical weights to each derived query based on relevance of each derived query in relation to the user search query”.
Then for each “derived query” a separate “search” is conducted and plurality of “electronic documents” are determined based on “entities” associated with the said “derived query”, which are used to determine a “sorted” “sequence” of “derived queries” (and/or associated “electronic documents” or “results”)), and in so doing the said “derived queries” are “append[ed]” with each respective “retrieved result” (e.g. like a query result pair) to “obtain” “a final search result”.
These steps are carried out by a “server” (i.e., as the only additional element in both claims) and “implement[ed]” using “a chain of searches (Cos)” and they involve “electronic documents” (another additional element requiring search to involve “electronic” and not just a regular result). Therefore other than the word “server”, there is nothing in these claims that cannot be done mentally; e.g., suppose I ask you when you are going on vacation, which theme parks will you take your children. This can easily be broken into: 1) when are you going on a vacation?, and 2) which theme parks will you take your children??. This division is based on realization that each of the sub-queries forms a complete and independent sentence (abide by a syntactic parsing) and they each independently provide a meaningful sentence (abide by a semantic reasoning). Your answer could be June or July (two answers for the first) and (magic mountain, or Disney land or sea world (three answers for the second)), and write them in a sheet of paper starting with the answer to the second question or sorting it according to the second one as it involved more answers and present them with their respective derived queries. These answers in combination (i.e., in a sequential appending analysis) could be further narrowed; e.g., if for example in any one of those vacation dates (e.g., June and/or July), one or more of the theme parks is closed, then that will become unavailable and has to weighted down to zero.
Therefore, these limitations, as drafted, under their broadest reasonable interpretation, cover performance of the said limitations in the mind but for the recitation of a generic computer. If a limitation (or limitations), under their broadest reasonable interpretation, cover performance of the said limitations in the mind but for the recitation of generic computer components, then they fall within the “Mental Processes” grouping of abstract ideas. Accordingly, each of the said claims recite an abstract idea.
The judicial exception is not integrated into a practical application. In particular, the claims recite one additional element – using a server to perform each of the “extracting” “mapping” “sorting” “searching and “appending” steps. The server in all these steps is recited at a high level of generality (i.e., as a generic server or computer) performing generic computer functions of e.g., “sorting” “based on a number of relevant electronic documents related to each query”, such that it amounts no more than mere instructions to apply the exception using a generic computer component. Accordingly, this additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claims are thus directed to an abstract idea.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using a server (using a “chain of searches” (“CoS”)) to perform the “extracting” “mapping” “sorting” “searching and “appending” steps amount to no more than mere instructions to apply the exception using a generic computer component; i.e., although the words “electronic” and “server” are used, but no technique in any limitations requires a “server” to perform it, and/or nothing in the claim limitations is tailored specifically to “electronic” “documents”. No specific technique is taught to be used to “generate” the “derived queries” from the “search query” to assess it to result in helping to enhance the “server” to run more efficiently and/or do something no machine has done before. Similar thing can be said for the other claim limitations. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. The claims are not patent eligible.
Furthermore, in the example above the process of dividing the original query into 1) when are you going on a vacation?, and 2) which theme parks will you take your children??, results in two grammatically correct separate queries (i.e. abides by a syntactic parsing), and both also abide by knowledge of vocabulary (semantic reasoning) and grammatical rules (syntactic reasoning), where the human uses a weighted combination of his knowledge of vocabulary and grammar in a language (a self-learned weighted combination operation).
Regarding claim 3 (13), in the example above the process of dividing the original query into 1) when are you going on a vacation?, and 2) which theme parks will you take your children??, results in two queries which have the following common word: “you”.
Regarding claim 4 (14), It would be quite reasonable to record (e.g. write a log (ontology) ontology) of query answer responses tagging them with e.g. date and time and/or based on name or other identifications pertaining to the queries) (e.g. alike a global ontology) in order to use if such queries are encountered again in order to save time in providing responses and/or guessing answers.
Regarding claim 5 (15), any person with basic literacy can recognize names in a query prior to further processing it.
Regarding claim 6 (16), a person who constructs the log (ontology) would know how to access it to recapture information, in particular if it is constructed based on e.g., name (keyword).
Regarding claim 7 (17), for the person who receives the answers to divided or derived queries above, he could analyze the query that received less answers first simply because there is higher chance that they are incorrect and would be useful to evaluate sooner.
Regarding claim 8 (18) “generating” “database search queries” (i.e., “SQL” ) on derived queries would be simply extra solution activity as providing responses to those derived queries could be made without that extra step. Furthermore, the claimed limitations do not cause “SQL” if implemented to experience enhanced performance, and this application had not invented the “SQL”.
Regarding claim 9 (19) any person with basic knowledge of grammar and vocabulary could handle various query data structures and still divide a query into “derived” or sub-queries according to context. It is very common that a sentence is received which lacks a part of speech component and it is still understandable, and/or there are sentences uttered that do not strictly abide by grammatical rules (i.e., not obey an explicit fixed schema) but their meanings are still inferred and still they could be e.g. divided into smaller portions for more concise analysis.
Regarding claim 10 (20), any person can associate an entity in a query sentence with a certain data type (e.g., medical, business, recreation, etc.) in order to analyze the query which includes that entity and to do search for it.
Allowable Subject Matter
Claims 1, 3-11, and 13-20 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
The following is an examiner’s statement of reasons for allowance: The claims recite a search retrieval method (claim 1) and system (claim11) in which a “search query” is broken into smaller “derived queries”, each with a smaller “length” than the “search query”, but are consistent in “intent” with it. The “two more derived queries are derived by syntactic parsing and semantic reasoning based on a self-learned weighted combination operation” which “includes” “combining an output of the syntactic parsing of the user search query with an output of the semantic reasoning of the user search query” “and assigning numerical weights to each derived query based on relevance of each derived query in relation to the user search query”.
Then for each “derived query” a separate “search” is conducted and plurality of “electronic documents” are determined based on “entities” associated with the said “derived query”, which are used to determine a “sorted” “sequence” of “derived queries” (and/or associated “electronic documents” or “results”)), and in so doing the said “derived queries” are “append[ed]” “sequentially” with each respective “retrieved result” (e.g. like a query result pair) to “obtain” “a final search result”.
The prior art of record THOTA et al. (US 2015/0254357) teach: ¶ 0013 sentence 1: “in response to receiving a query from a user” (from a user query) “where the query specifies a location and an itinerary, a plurality of sub-queries” (two or more derived queries by splitting the user query (split stage)) “is generated” (are generated) “for the query”; these are submitted to “web-based search engines” (a server which is responsible for all the search and retrieval that follow (¶ 0003 sentence 1))); ¶ 0054 lines 1-9: “For example, a submitted may be "kids day out" in "New York, N.Y."”(a user query is used to generate) “In response to receiving this query, a plurality of sub-queries” (two derived) “may be generated, based on predefined information associating itinerary query terms (e.g., "kids day out")” (based on the user intent) “with specific activity query terms (e.g., "kid-friendly restaurant" "kid-friendly entertainment", etc.). In this example, a first sub-query may be "kid-friendly restaurant" in "New York, N.Y."” “and a second sub-query may be "kid-friendly entertainment" in "New York, N.Y."” (self-complete independent queries)); ¶ 0013 sentence 3: “Then, for each sub-query” (for each derived query) “a list of query results is generated where each query result in the list of query results indicates a business entity” (an entity is extracted) “located in the vicinity of the specified location that is relevant to the particular activity”; e.g., ¶ 0054 example “New York” “N.Y.” (one or more entities in each derived query)); ¶ 0013 sentence 3: “Then, for each sub-query” (for each derived query) “a list of query results” (a plurality of electronic documents) “is generated” (are identified) “where each query result in the list of query results indicates a business entity” (each mapped to the entity and associated with a “business” (a data source of financial type)) “located in the vicinity of the specified location that is relevant” (and relevant to the search) “to the particular activity”; other data sources the queries could be directed to is search for “park” (a recreation and outdoor activity data source (¶ 0032 last sentence)) and/or search for “daycare” (child care data source (¶ 0024 sentence 2)); ¶ 0013 sentence 4: “The list” (based on a number of) “of sub-query results” (the relevant electronic documents) “for each sub-query” (for each derived query) “may be ranked” (the derived queries are sorted) “based on both the spatial proximity/relevance and the activity relevance of the sub-query results” (to obtain a sorted sequence of derived queries)); ¶ 0013 last sentence: “Then, all the lists of sub-query results” (following the “ranking” (sorting by analyzing while searching), each “query result” (electronic document) associated with each “sub-query” (derived query)) “may be combined” (to obtain a “top-ranked” “result” (one result (¶ 0054)))) “and provided to the user as query results for the specified itinerary (lists of places matching the activity criteria) at the specified location”); ¶ 0054 lines 13+: “For example, a list of itineraries may include a first itinerary that includes” (appending) “the top-ranked query result” (a final search result retrieved for a derived query) “for” "kid-friendly entertainment"(and its associated or consequent derived query and it is the “top” or first in the “ranked” (sorted) list) “and the top-ranked query result for "kid-friendly restaurant").
THOTA et al. do not specifically disclose wherein each of the two or more derived queries has a length less than a first length of the user search query received originally from a client device; wherein in the split stage, the two or more derived queries are derived by syntactic parsing and semantic reasoning based on a self- learned weighted combination operation by combining an output of the syntactic parsing of the user search query with an output of the semantic reasoning of the user search query, which involves assigning numerical weights to each derived query based on relevance of each derived query in relation to the user search query .
Tumuluri (US 2022/0114463) do teach:¶0038, page 5 lines 3+: “The soft-query” (a user search query is used) “may be divided” (to generate by splitting (split stage)) “into multiple micro-queries” (into two or more derived queries of shorter length as they are obtained by the division of the “soft-query” (user search query)) “based on context, intent”); ¶ 0048 S1: “The semantic and syntactic” (using semantic and syntactic reasoning) “analyzer 608 may be configured to extract entity information”, wherein the “entity” is used to according to ¶ 0038 page 5 lines 3+: “The soft query” (a query) “may be divided” (splits (in a split stage)) “into multiple micro-queries” (into two or more derived queries) “based on” “entities” (using “entities” which were obtained by semantic and syntactic reasoning) “context, intent, objects and previous responses” (a self-learned weighted combination); ¶ 0038 lines 7+: “the results may be sorted” “for example, based on the relevance scores” (a self-learned weighted parameter used which has the same impact as weighting their associated queries according to their “relevance”) “of the micro-queries results” (to validate division of a “soft-query” (query) into the plurality of derived queries) “with respect to soft-query). The micro-query score” “is a relevance score”).
Although THOTA et al. teaches ¶ 0013 last sentence: “Then, all the lists of sub-query results” (the results of queries) “may be combined” (appended) to obtain a “top-ranked” “result” (one result (¶ 0054)))) “and provided to the user”, but this operation is not a “sequential[]” “combin[ing]” action; i.e., these “result[s]” are obtained independently and presented as such without any impact on each other. According to Sp. ¶ 0032 last S: “In the supplementing stage, the appending or supplementing of successive derived queries with previous intermediate results by the system acts as a constraint for granular responses, which results in a focused and faster retrieval of the final result to the original user query”; i.e., the “sequential[]” “appending” imposes “constraints” (limitations and/or narrowing) on individual sub-query results which in principle should enhance their overall validity.
Further search did not produce any reference teaching this phenomenon and therefore these claims became allowable. Claims 3-10 (dependent on claim 1), and 13-20 (dependent on claim 11), further limit the scope of their respective parent claims 1 and 11 and are thus allowable under similar rationale.
.
Any comments considered necessary by applicant must be submitted no later than the payment of the issue fee and, to avoid processing delays, should preferably accompany the issue fee. Such submissions should be clearly labeled “Comments on Statement of Reasons for Allowance.”
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Ross et al. (US Patent 9,235,607) Col. 8 lines 65+: “using the logical relationships to divide” (parse to generate) “the query terms” (a user search query) “into one or more portions” “e.g., sub-queries” (two or more smaller derived queries); Col. 9 lines 25-26: “sort” “search results” but the “sort” is not based on the number of “search results” associated with each “sub-query”.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to FARZAD KAZEMINEZHAD whose telephone number is (571)270-5860. The examiner can normally be reached 10:30 am to 11:30 pm.
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, Paras D. Shah can be reached at (571) 270-1650. 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.
/Farzad Kazeminezhad/
Art Unit 2653
July 25th 2026.
Prosecution Insights