# US 12,182,682 B2 · Method for retrieving a recommendation from a knowledge database of a ticketing system > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,182,682 B2 - Document ID: US12182682B2 - Status: Granted - Kind: B2 - Application number: 16/495,243 - Filed: 2018-01-23 - Published: 2020-01-30 - Granted: 2024-12-31 - Earliest claimed: 2017-03-22 - Assignees: Siemens Aktiengesellschaft - Applicants: Siemens Aktiengesellschaft - Inventors: Bernt Andrassy; Pankaj Gupta - Web page: https://patentgrep.com/patent/US12182682B2 ## CPC - G06F 16/00: Information retrieval; Database structures therefor; File system structures therefor - G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks - G06F 16/2462: {Approximate or statistical queries} - G06F 16/252: {between a Database Management System and a front-end application} - G06F 40/30: Semantic analysis - G06N 3/043: based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS] - 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 3/08: Learning methods - G06N 3/09: Supervised learning - G06Q 10/20: Administration of product repair or maintenance - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR ## Abstract Provided is a ticketing system adapted to retrieve a recommendation from a knowledge database in response to a received query, the ticketing system including a processor adapted to perform semantic similarity learning in textual description pairs by calculating similarity scores for similarities between the received query and tickets stored in the knowledge database of the ticketing system, wherein each textual description pair includes a textual description of the received query and a textual description of a ticket of a plurality of tickets stored in the knowledge database of the ticketing system, wherein the ticket having the maximum similarity score is identified and a solution of the identified ticket is output as the retrieved recommendation for the received query by the ticketing system. ## Claims ### Claim 1 A method for automatic monitoring and control of a power plant, a production line, or a turbine, the method comprising: receiving a query regarding the power plant, the production line or the turbine, wherein internal states of the power plant, the production line or the turbine are monitored and queries concerning an abnormal behavior of the power plant, the production line or the turbine are generated automatically depending on the monitored state of the power plant, the production line or the turbine, wherein the received query comprises as a textual description a description text; performing a semantic textual similarity learning in textual description pairs by calculating similarity scores for similarities between the received query and a plurality of tickets stored in a knowledge database, wherein each ticket stored in the knowledge database comprises as a textual description a description text and a solution text, wherein each textual description pair includes a textual description of the received query and a textual description of each of the tickets of the plurality of tickets stored in said knowledge database; and identifying a ticket from the plurality of tickets having the calculated maximum similarity score and reading a solution of the identified ticket as the retrieved recommendation for the received query, wherein the solution of the identified ticket read from the knowledge base as the retrieved recommendation for the received query comprises control data for controlling the power plant, the production line or the turbine; and automatically controlling the power plant, the production line or the turbine to address the abnormal behavior of the power plant, the production line or the turbine using the control data of the identified ticket. ### Claim 2 (depends on 1) The method according to claim 1 wherein the received query comprises as a textual description a subject text. ### Claim 3 (depends on 1) The method according to claim 1 wherein each ticket stored in the knowledge database comprises as a textual description a subject text. ### Claim 4 (depends on 1) The method according to claim 1 wherein each ticket stored in the knowledge database further comprises at least one of a timestamp, metadata, information data or control data. ### Claim 5 (depends on 1) The method according to claim 1 wherein the query is input by a user or generated by a controller of the power plant, the production line or the turbine in response to a monitored state of the power plant, the production line or the turbine. ### Claim 6 (depends on 1) The method according to claim 1 wherein the solution of the identified ticket read from the knowledge database as the retrieved recommendation for the received query comprises a solution text of the identified ticket for a user. ### Claim 7 (depends on 1) The method according to claim 1 wherein the representations of the textual description of a ticket are updated during training of associated neural networks and precalculated for evaluation in response to a received query. ### Claim 8 (depends on 1) The method according to claim 1, wherein the query is generated by a controller of the power plant, the production line or the turbine in response to a monitored state of the power plant, the production line or the turbine and provided via a logical data network. ### Claim 9 (depends on 1) The method according to claim 1, wherein the semantic textual similarity learning includes using a Siamese bi-directional Long Short Term Memory (LSTM) neural network architecture for multi-level and cross-level or asymmetric textual similarity learning. ### Claim 10 (depends on 1) The method according to claim 1, wherein the power plant, the production line or the turbine is connected to a local ticketing system used for a plurality of machines in an industrial plant, wherein the knowledge database is located within the local ticketing system. ### Claim 11 (depends on 1) The method according to claim 1, wherein the power plant, the production line or the turbine is connected to a remote ticketing system used for a plurality of machines in different sites connected by a data network, wherein the knowledge database is located within the local ticketing system. ### Claim 12 (depends on 1) The method according to claim 1, further comprising storing the retrieved recommendation in the knowledge base as an additional ticket for further processing. ### Claim 13 (depends on 1) The method according to claim 1 wherein multi-level symmetric representation differences and/or cross-level asymmetric representation differences between the received query and the tickets stored in the knowledge database are calculated. ### Claim 14 (depends on 13) The method according to claim 13 wherein the similarity score indicating a degree of similarity between the received query and a ticket stored in the knowledge database is calculated on the basis of the multi-level symmetric representation differences and/or on the basis of the cross-level asymmetric representation differences using a predetermined similarity metric. ### Claim 15 (depends on 14) The method according to claim 14 wherein the similarity metric comprises a Manhattan similarity metric. ### Claim 16 (depends on 1) The method according to claim 1 wherein for each textual description of the received query and for each textual description of the ticket at least one associated representation is calculated. ### Claim 17 (depends on 16) The method according to claim 16 wherein the representation of the textual description comprises a word embedding. ### Claim 18 (depends on 16) The method according to claim 16 wherein the representation of the textual description comprises a hidden state of an associated neural network. ### Claim 19 (depends on 18) The method according to claim 18 wherein the representation of the textual description comprises a hidden state of a bidirectional Long Short Term Memory, LSTM, network. ### Claim 20 A ticketing system adapted to retrieve a recommendation from a knowledge database in response to a received query regarding a power plant, a production line or a turbine, wherein the recommendation includes control data for automatically controlling the power plant, the production line or the turbine, said ticketing system comprising: a processor adapted to perform semantic similarity learning in textual description pairs by calculating similarity scores for similarities between the received query and a plurality of tickets stored in the knowledge database of said ticketing system, wherein each ticket of the plurality of tickets stored in the knowledge database comprises as a textual description a description text and a solution text, and wherein the received query comprises as a textual description a description text, wherein each textual description pair includes the textual description of the received query and the textual description of a ticket of the plurality of tickets stored in the knowledge database of said ticketing system, wherein the ticket having the maximum similarity score is identified and the solution text of the identified ticket is output as the retrieved recommendation for the received query by said ticketing system, wherein internal states of the power plant, the production line or the turbine are monitored and queries concerning an abnormal behavior of the power plant, the production line or the turbine are generated automatically depending on the monitored state of the power plant, the production line or the turbine, and wherein the solution of the identified ticket read from the knowledge base as the retrieved recommendation for the received query comprises control data of the identified ticket which is used to automatically control the power plant, the production line or the turbine to address the abnormal behavior of the power plant, the production line or the turbine.