# US 11,030,226 B2 · Facilitating answering questions involving reasoning over quantitative information > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,030,226 B2 - Document ID: US11030226B2 - Status: Granted - Kind: B2 - Application number: 15/875,051 - Filed: 2018-01-19 - Published: 2019-07-25 - Granted: 2021-06-08 - Earliest claimed: 2018-01-19 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Elizabeth Daly; Charles Arthur Jochim; Akihiro Kishimoto; Vanessa Lopez Garcia; Radu Marinescu - Web page: https://patentgrep.com/patent/US11030226B2 ## CPC - G06F 16/3329: Natural language query formulation - G06F 16/3334: {Selection or weighting of terms from queries, including natural language queries} - G06F 16/3344: {using natural language analysis} - G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] - G06N 3/0455: Auto-encoder networks; Encoder-decoder networks - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound - G06N 5/041: {Abduction} - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 20/00: Machine learning - G06N 7/02: using fuzzy logic (computing arrangements based on biological models G06N3/00; computing arrangements using knowledge-based models G06N5/00) - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Systems, computer-implemented methods and/or computer program products that facilitate answering questions that involve mathematical computations are provided. In one embodiment, a computer-implemented method comprises: transforming, by a system operatively coupled to a processor, a natural language query into a first logical representation and extrinsic knowledge into a second logical representation relevant to the natural language query; merging, by the system, the first logical representation and the second logical representation into a third logical representation; and generating, by the system, answers for the natural language query based on processing of the third logical representation. ## Claims ### Claim 1 A system, comprising: a memory that stores computer executable components; a processor, operably coupled to the memory, and that executes computer executable components stored in the memory, wherein the computer executable components comprise: an analysis component that: transforms a natural language query into a first logical representation, wherein the natural language query involves one or more mathematical computations, determines a context of the natural language query based upon data collected related to an activity of a user associated with the natural language query, selects a domain from a plurality of domains based upon the context, and translates extrinsic knowledge associated with the domain into a second logical representation relevant to the natural language query; a merging component that merges the first logical representation and the second logical representation into a third logical representation; and an answering component that generates answers for the natural language query based on processing of the third logical representation, wherein the answers comprise respective confidence values for the answers, and a confidence value for an answer is based on a quantity of sources in the domain that have a same set of equations for determining the answer. ### Claim 2 (depends on 1) The system of claim 1, further comprising an explanation component that generates explanations for the answers. ### Claim 3 (depends on 2) The system of claim 2, further comprising a probabilistic component that assigns the respective confidence values to the the answers. ### Claim 4 (depends on 3) The system of claim 3, wherein the probabilistic component also ranks the answers and corresponding explanations that are most probable based on the respective confidence values for the answers. ### Claim 5 (depends on 4) The system of claim 4, further comprising a machine learning component that generates and trains models with the answers and the explanations based on user feedback to provide answers with higher respective confidence values. ### Claim 6 (depends on 1) The system of claim 1, wherein the answers are questions. ### Claim 7 (depends on 1) The system of claim 1, wherein the extrinsic knowledge is knowledge compiled from or more sources of the domain relevant to the natural language query. ### Claim 8 A computer-implemented method, comprising: transforming, by a system operatively coupled to a processor, a natural language query into a first logical representation, wherein the natural language query involves one or more mathematical computations; determining, by the system, a context of the natural language query based upon data collected related to an activity of a user associated with the natural language query; selecting, by the system, a domain from a plurality of domains based upon the context; and transforming, by the system, extrinsic knowledge associated with the domain into a second logical representation relevant to the natural language query; merging, by the system, the first logical representation and the second logical representation into a third logical representation; and generating, by the system, answers for the natural language query based on processing of the third logical representation, wherein the answers comprise respective confidence values for the answers, and a confidence value for an answer is based on a quantity of sources in the domain that have a same set of equations for determining the answer. ### Claim 9 (depends on 8) The computer-implemented method of claim 8, further comprising generating explanations for the answers. ### Claim 10 (depends on 9) The computer-implemented method of claim 9, further comprising assigning the respective confidence values to the first logical representation, the second logical representation and the answers. ### Claim 11 (depends on 10) The computer-implemented method of claim 10, further comprising ranking the answers and corresponding explanations that are most probable based on the respective confidence values for the answers. ### Claim 12 (depends on 11) The computer-implemented method of claim 11, further comprising generating and training models with the answers and the explanations based on user feedback to provide answers with higher respective confidence values. ### Claim 13 (depends on 8) The computer-implemented method of claim 8, wherein the answers are questions. ### Claim 14 (depends on 8) The computer-implemented method of claim 8, wherein the extrinsic knowledge is knowledge compiled from or more sources of the domain relevant to the natural language query. ### Claim 15 A computer program product for facilitating answering questions that involve mathematical computations, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to: transform a natural language query into a first logical representation, wherein the natural language query involves one or more mathematical computations; determine a context of the natural language query based upon data collected related to an activity of a user associated with the natural language query; select a domain from a plurality of domains based upon the context; and transform extrinsic knowledge associated with the domain into a second logical representation relevant to the natural language query; merge the first logical representation and the second logical representation into a third logical representation; and generate answers for the natural language query based on processing of the third logical representation, wherein the answers comprise respective confidence values for the answers, and a confidence value for an answer is based on a quantity of sources in the domain that have a same set of equations for determining the answer. ### Claim 16 (depends on 15) The computer program product of claim 15, wherein the program instructions are further executable to cause the processor to: generate explanations for the answers. ### Claim 17 (depends on 16) The computer program product of claim 16, wherein the program instructions are further executable to cause the processor to: assign the respective confidence values to the first logical representation, the second logical representation and the answers. ### Claim 18 (depends on 17) The computer program product of claim 17, wherein the program instructions are further executable to cause the processor to: rank the answers and corresponding explanations that are most probable based on the respective confidence values for the answers. ### Claim 19 (depends on 18) The computer program product of claim 18, wherein the program instructions are further executable to cause the processor to: generate and train models with the answers and the explanations based on user feedback to provide answers with higher respective confidence values. ### Claim 20 (depends on 15) The computer program product of claim 15, wherein the answers are questions.