# US 11,443,211 B2 · Extracting important sentences from documents to answer hypothesis that include causes and consequences > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,443,211 B2 - Document ID: US11443211B2 - Status: Granted - Kind: B2 - Application number: 16/737,489 - Filed: 2020-01-08 - Published: 2021-07-08 - Granted: 2022-09-13 - Earliest claimed: 2020-01-08 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Miki Ishikawa; Sachiko Yoshihama; Issei Yoshida; Kohichi Kamijoh - Web page: https://patentgrep.com/patent/US11443211B2 ## CPC - G06F 40/30: Semantic analysis - G06N 5/041: {Abduction} - G06F 16/3329: Natural language query formulation - G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30) - G06F 40/44: Statistical methods, e.g. probability models - G06F 40/49: using very large corpora, e.g. the web - G06F 40/284: Lexical analysis, e.g. tokenisation or collocates - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract A method is provided for validating a hypothesis sentence. The method extracts, from a document database D using a hypothesis sentence that includes a causal part and a consequence part, a set D1 of documents related to the causal part. The method extracts, from set D1, a set S of sentences that include expressions of opinion. The method obtains a word list W of words that have a high co-occurrence in the set D1. The method selects, from the set S, a set S1 of sentences that are positionally close to any of the words in the word list W. The method selects, from the set S, a set S2 of sentences that are related to the words in the consequence part of the hypothesis. The method extracts and displays sentences included in both the set S1 and the set S2 as opinion sentences in relation to the hypothesis sentence. ## Claims ### Claim 1 A computer-implemented method for machine-validating an input hypothesis sentence, comprising: extracting, by a hardware processor from a document database D using the input hypothesis sentence that includes a causal part and a consequence part, a set D1 of documents related to the causal part; extracting, by the hardware processor from the set D1 of documents, a set S of sentences that include one or more expressions of opinion; obtaining, by the hardware processor, a word list W of words that have a high co-occurrence in the set D1 of documents; selecting, by the hardware processor from the set S of sentences, a set S1 of sentences that are positionally close to any of the words in the word list W; selecting, by the hardware processor from the set S of sentences, a set S2 of sentences that are related to the words in the consequence part of the hypothesis; and extracting, by the hardware processor, and displaying, by a display device, sentences included in both the set S1 of sentences and the set S2 of sentences as opinion sentences in relation to the input hypothesis sentence. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, wherein the set D1 of documents is related to the causal part based on word commonality between the causal part and each of the documents in the set D1 of documents. ### Claim 3 (depends on 1) The computer-implemented method of claim 1, wherein a given word in the word list W of words is obtained based on (i) a first ratio of a number of occurrences of the given word in any of the documents in the set D1 that include keywords related to the input hypothesis sentence, and (ii) a second ratio of a number of occurrences of the given word in set D1 over the total number of documents in the set D1. ### Claim 4 (depends on 3) The computer-implemented method of claim 3, wherein the first ratio is divided by the second ratio and evaluated relative to a threshold. ### Claim 5 (depends on 3) The computer-implemented method of claim 3, wherein the keywords related to the input hypothesis sentence are determined based on word commonality and word synonym commonality. ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein the set S2 are selected based on a cosine similarity value for a candidate sentence for the set S2 being greater than a threshold value, wherein the cosine similarity value is determined between (i) words that occur in the consequence part of the input hypothesis sentence and (ii) objective words of an opinion in the candidate sentence for the set S2. ### Claim 7 (depends on 6) The computer-implemented method of claim 6, wherein said step of selecting the set S2 further comprises increasing a similarity weight between (i) the words that occur in the consequence part of the input hypothesis sentence and (ii) the objective words of the opinion in the candidate sentence for the set S2, responsive to a noun match ratio. ### Claim 8 (depends on 6) The computer-implemented method of claim 6, wherein said step of selecting the set S2 further comprises decreasing a similarity weight between (i) the words that occur in the consequence part of the input hypothesis sentence and (ii) the objective words of the opinion in the candidate sentence for the set S2, responsive to opposing word polarities. ### Claim 9 (depends on 1) The computer-implemented method of claim 1, wherein a threshold β is used such that words having a co-occurrence above the threshold β in the documents in the set D1 are added to the word list W. ### Claim 10 (depends on 1) The computer-implemented method of claim 1, wherein the document database D comprises news articles and report documents. ### Claim 11 (depends on 1) The computer-implemented method of claim 1, wherein the sentences in the set S are determined to be positionally close to any of the words in the word list W based on membership in both the word list W and the set S of sentences. ### Claim 12 A computer program product for machine-validating an input hypothesis sentence, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising: extracting, by a hardware processor from a document database D using the input hypothesis sentence that includes a causal part and a consequence part, a set D1 of documents related to the causal part; extracting, by the hardware processor from the set D1 of documents, a set S of sentences that include one or more expressions of opinion; obtaining, by the hardware processor, a word list W of words that have a high co-occurrence in the set D1 of documents; selecting, by the hardware processor from the set S of sentences, a set S1 of sentences that are positionally close to any of the words in the word list W; selecting, by the hardware processor from the set S of sentences, a set S2 of sentences that are related to the words in the consequence part of the hypothesis; and extracting, by the hardware processor, and displaying, by a display device, sentences included in both the set S1 of sentences and the set S2 of sentences as opinion sentences in relation to the input hypothesis sentence. ### Claim 13 (depends on 12) The computer program product of claim 12, wherein the set D1 of documents is related to the causal part based on word commonality between the causal part and each of the documents in the set D1 of documents. ### Claim 14 (depends on 12) The computer program product of claim 12, wherein a given word in the word list W of words is obtained based on (i) a first ratio of a number of occurrences of the given word in any of the documents in the set D1 that include keywords related to the input hypothesis sentence, and (ii) a second ratio of a number of occurrences of the given word in set D1 over the total number of documents in the set D1. ### Claim 15 (depends on 14) The computer program product of claim 14, wherein the first ratio is divided by the second ratio and evaluated relative to a threshold. ### Claim 16 (depends on 14) The computer program product of claim 14, wherein the keywords related to the input hypothesis sentence are determined based on word commonality and word synonym commonality. ### Claim 17 (depends on 12) The computer program product of claim 12, wherein the set S2 are selected based on a cosine similarity value for a candidate sentence for the set S2 being greater than a threshold value, wherein the cosine similarity value is determined between (i) words that occur in the consequence part of the input hypothesis sentence and (ii) objective words of an opinion in the candidate sentence for the set S2. ### Claim 18 (depends on 17) The computer program product of claim 17, wherein said step of selecting the set S2 further comprises increasing a similarity weight between (i) the words that occur in the consequence part of the input hypothesis sentence and (ii) the objective words of the opinion in the candidate sentence for the set S2, responsive to a noun match ratio. ### Claim 19 (depends on 17) The computer program product of claim 17, wherein said step of selecting the set S2 further comprises decreasing a similarity weight between (i) the words that occur in the consequence part of the input hypothesis sentence and (ii) the objective words of the opinion in the candidate sentence for the set S2, responsive to opposing word polarities. ### Claim 20 A computer processing system for machine-validating an input hypothesis sentence, comprising: a memory device including program code stored thereon; a hardware processor, operatively coupled to the memory device, and configured to run the program code stored on the memory device to extract, from a document database D using the input hypothesis sentence that includes a causal part and a consequence part, a set D1 of documents related to the causal part; extract, from the set D1 of documents, a set S of sentences that include one or more expressions of opinion; obtain a word list W of words that have a high co-occurrence in the set D1 of documents; select, from the set S of sentences, a set S1 of sentences that are positionally close to any of the words in the word list W; select, from the set S of sentences, a set S2 of sentences that are related to the words in the consequence part of the hypothesis; and extract and cause to be displayed on a display device sentences included in both the set S1 of sentences and the set S2 of sentences as opinion sentences in relation to the input hypothesis sentence.