# US 2013/0159272 A1 · Understanding and addressing complex information needs > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2013/0159272 A1 - Document ID: US20130159272A1 - Status: Published application - Kind: A1 - Application number: 13/327,734 - Filed: 2011-12-15 - Published: 2013-06-20 - Earliest claimed: 2011-12-15 - Assignees: Yahoo! Inc. - Inventors: Meryem Pinar Donmez; Ana-Maria Popescu - Web page: https://patentgrep.com/patent/US20130159272A1 ## CPC - G06F 16/338: Presentation of query results - G06F 16/332: Query formulation - G06F 16/3338: {Query expansion} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Techniques described herein assist users in satisfying complex information needs represented as long, detailed questions. A generalized search assistance framework for complex information needs is disclosed. Given a detailed question, the techniques enrich the original question with a set of related concepts. The types of questions handled are detailed, complex questions similar to the ones posted in Q&A portals. A generalized search assistance framework enriches complex detailed questions with topically related concepts. A basic pipeline represents an instantiation of such the search assistance framework. Given a detailed question, the pipeline relies on semantic and syntactic relationships in the detailed question in order to construct a set of related queries. The queries are issued to a commercial search engine and the retrieved results are processed by state-of-the-art document understanding techniques in order to retrieve important concepts. A final concept set for enriching the original question is then assembled. ## Claims ### Claim 1 A computer-implemented method comprising steps of: receiving an original question from a user; generating two or more queries based on the original question; submitting each of the two or more queries to a search engine; for each particular query of the two or more queries, receiving, from the search engine, a separate document set returned as results for that particular query, thereby receiving a plurality of document sets; for each particular document set in the plurality of document sets, identifying concepts to which terms in the particular document set relate, and adding said concepts to a concept set; wherein the steps are performed by one or more computing devices. ### Claim 2 (depends on 1) The method of claim 1, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences. ### Claim 3 (depends on 1) The method of claim 1, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence. ### Claim 4 (depends on 1) The method of claim 1, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises: identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence; and determining one or more dependencies between (a) various noun phrases of the one or more noun phrases and either (b) various other noun phrases of the one or more noun phrases or (c) various verbs of the one or more verbs. ### Claim 5 (depends on 1) The method of claim 1, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises: identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence; determining one or more dependencies between (a) various noun phrases of the one or more noun phrases and either (b) various other noun phrases of the one or more noun phrases or (c) various verbs of the one or more verbs; and merging together only (1) pairs of noun phrases that have a dependency that is contained in a specified set of dependencies and (2) verbs and noun phrases that have a dependency that is contained in the specified set of dependencies; and wherein generating the two or more queries comprises using, as at least one query of the two or more queries, at least one phrase produced as a result of said merging. ### Claim 6 (depends on 1) The method of claim 1, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises: identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence; determining one or more dependencies between (a) various noun phrases of the one or more noun phrases and either (b) various other noun phrases of the one or more noun phrases or (c) various verbs of the one or more verbs; and merging together only (1) pairs of noun phrases that have a dependency that is contained in a specified set of dependencies and (2) verbs and noun phrases that have a dependency that is contained in the specified set of dependencies; wherein generating the two or more queries comprises using, as at least one query of the two or more queries, at least one phrase produced as a result of said merging; and wherein the specified set of dependencies consists of: agent, appositional, conjunct, clausal subject, clausal passive subject, direct object, indirect object, negation modifier, noun compound modifier, noun phrase as adverbial modifier, nominal subject, passive nominal subject, purpose clause modifier, relative clause modifier, and prepositional clause modifier. ### Claim 7 (depends on 1) The method of claim 1, wherein the step of generating the two or more queries comprises: parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences, thereby generating a plurality of queries; and generating, as at least one of the two or more queries, a super query that contains at least parts of two queries of the plurality of queries. ### Claim 8 (depends on 1) The method of claim 1, wherein the step of identifying concepts to which terms in the particular document set relate comprises ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document. ### Claim 9 (depends on 1) The method of claim 1, wherein the step of identifying concepts to which terms in the particular document set relate comprises ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document; and wherein ranking the particular terms comprises ranking the particular terms relative to each other based at least in part on positions at which the particular terms occur within the particular document. ### Claim 10 (depends on 1) The method of claim 1, wherein the step of identifying concepts to which terms in the particular document set relate comprises ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document; and wherein ranking the particular terms comprises ranking the particular terms relative to each other based at least in part on font sizes of the particular terms within the particular document. ### Claim 11 (depends on 1) The method of claim 1, wherein the step of identifying concepts to which terms in the particular document set relate comprises mapping particular terms in a particular document within the particular document set to Wikipedia entries that correspond to the particular terms; and wherein the step of adding said concepts to said concept set comprises adding, to said concept set, particular concepts that are primary topics of at least some of said Wikipedia entries. ### Claim 12 (depends on 1) The method of claim 1, wherein the step of identifying concepts to which terms in the particular document set relate comprises: ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document; selecting a specified quantity of the particular terms based on said ranking; and wherein the step of adding said concepts to said concept set comprises: for each and only each selected term for which there exists a Wikipedia entry that corresponds to that selected term, adding, to said concept set, a particular concept that is a primary topic of that Wikipedia entry. ### Claim 13 (depends on 1) The method of claim 1, further comprising: for each particular concept identified in the particular document set, (1) determining, for the particular concept, a plurality of rank scores that indicate, for each particular document in the particular document set, a rank of the particular concept within that particular document relative to other concepts within that particular document, and then (2) determining an aggregate score for the particular concept based at least in part on the plurality of rank scores; for each particular query of the two or more queries, selecting a subset of the concepts identified in the document set generated by that particular query based at least in part on the aggregate scores determined for concepts identified in the document set generated by that query, thereby generating a separate topical concept subset for each particular query of the two or more queries; and for each selected concept selected for inclusion in a topical concept subset, determining a final score for that that selected concept based at least in part on a quantity of topical concept subsets in which that selected concept occurs. ### Claim 14 A non-transitory computer-readable storage medium that stores instructions which, when executed by one or more processors, cause the one or more processors to perform steps comprising: receiving an original question from a user; generating two or more queries based on the original question; submitting each of the two or more queries to a search engine; for each particular query of the two or more queries, receiving, from the search engine, a separate document set returned as results for that particular query, thereby receiving a plurality of document sets; for each particular document set in the plurality of document sets, identifying concepts to which terms in the particular document set relate, and adding said concepts to a concept set; wherein the steps are performed by one or more computing devices. ### Claim 15 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences. ### Claim 16 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence. ### Claim 17 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises: identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence; and determining one or more dependencies between (a) various noun phrases of the one or more noun phrases and either (b) various other noun phrases of the one or more noun phrases or (c) various verbs of the one or more verbs. ### Claim 18 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises: identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence; determining one or more dependencies between (a) various noun phrases of the one or more noun phrases and either (b) various other noun phrases of the one or more noun phrases or (c) various verbs of the one or more verbs; and merging together only (1) pairs of noun phrases that have a dependency that is contained in a specified set of dependencies and (2) verbs and noun phrases that have a dependency that is contained in the specified set of dependencies; and wherein generating the two or more queries comprises using, as at least one query of the two or more queries, at least one phrase produced as a result of said merging. ### Claim 19 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of generating the two or more queries comprises parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences; and wherein the step of generating at least one query based on each sentence of the two or more sentences comprises: identifying one or more noun phrases and one or more verbs within a particular sentence of the two or more sentences and extracting the one or more noun phrases and the one or more verbs from the particular sentence; determining one or more dependencies between (a) various noun phrases of the one or more noun phrases and either (b) various other noun phrases of the one or more noun phrases or (c) various verbs of the one or more verbs; and merging together only (1) pairs of noun phrases that have a dependency that is contained in a specified set of dependencies and (2) verbs and noun phrases that have a dependency that is contained in the specified set of dependencies; wherein generating the two or more queries comprises using, as at least one query of the two or more queries, at least one phrase produced as a result of said merging; and wherein the specified set of dependencies consists of: agent, appositional, conjunct, clausal subject, clausal passive subject, direct object, indirect object, negation modifier, noun compound modifier, noun phrase as adverbial modifier, nominal subject, passive nominal subject, purpose clause modifier, relative clause modifier, and prepositional clause modifier. ### Claim 20 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of generating the two or more queries comprises: parsing the original question into two or more sentences and then generating at least one query based on each sentence of the two or more sentences, thereby generating a plurality of queries; and generating, as at least one of the two or more queries, a super query that contains at least parts of two queries of the plurality of queries. ### Claim 21 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying concepts to which terms in the particular document set relate comprises ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document. ### Claim 22 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying concepts to which terms in the particular document set relate comprises ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document; and wherein ranking the particular terms comprises ranking the particular terms relative to each other based at least in part on positions at which the particular terms occur within the particular document. ### Claim 23 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying concepts to which terms in the particular document set relate comprises ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document; and wherein ranking the particular terms comprises ranking the particular terms relative to each other based at least in part on font sizes of the particular terms within the particular document. ### Claim 24 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying concepts to which terms in the particular document set relate comprises mapping particular terms in a particular document within the particular document set to Wikipedia entries that correspond to the particular terms; and wherein the step of adding said concepts to said concept set comprises adding, to said concept set, particular concepts that are primary topics of at least some of said Wikipedia entries. ### Claim 25 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the step of identifying concepts to which terms in the particular document set relate comprises: ranking particular terms in a particular document within the particular document set based at least in part on an extent to which the particular terms represent an overall topic of the particular document; selecting a specified quantity of the particular terms based on said ranking; and wherein the step of adding said concepts to said concept set comprises: for each and only each selected term for which there exists a Wikipedia entry that corresponds to that selected term, adding, to said concept set, a particular concept that is a primary topic of that Wikipedia entry. ### Claim 26 (depends on 14) The non-transitory computer-readable storage medium of claim 14, wherein the steps further comprise: for each particular concept identified in the particular document set, (1) determining, for the particular concept, a plurality of rank scores that indicate, for each particular document in the particular document set, a rank of the particular concept within that particular document relative to other concepts within that particular document, and then (2) determining an aggregate score for the particular concept based at least in part on the plurality of rank scores; for each particular query of the two or more queries, selecting a subset of the concepts identified in the document set generated by that particular query based at least in part on the aggregate scores determined for concepts identified in the document set generated by that query, thereby generating a separate topical concept subset for each particular query of the two or more queries; and for each selected concept selected for inclusion in a topical concept subset, determining a final score for that that selected concept based at least in part on a quantity of topical concept subsets in which that selected concept occurs.