# US 12,015,812 B2 · Method and system for self-correcting match states > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,015,812 B2 - Document ID: US12015812B2 - Status: Granted - Kind: B2 - Application number: 17/930,159 - Filed: 2022-09-07 - Published: 2023-01-05 - Granted: 2024-06-18 - Earliest claimed: 2020-12-02 - Assignees: Simplebet, Inc. - Applicants: SimpleBet, Inc. - Inventors: Simon Gershey; Kyle Engel; Dave Lucia; Zachary Schumacher - Web page: https://patentgrep.com/patent/US12015812B2 ## CPC - H04N 21/251: {Learning process for intelligent management, e.g. learning user preferences for recommending movies (details of learning user preferences for the retrieval of video data in a video database G06F16/739; computer systems using learning methods G06N3/08)} - G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound - G06N 20/00: Machine learning - G06V 10/98: Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns - H04N 21/2187: Live feed - H04N 21/25891: {being end-user preferences (retrieval of video data in a video database based on user preferences G06F16/739; arrangements for recognizing users' preferences H04H60/46; user profiles in network data switching protocols H04L67/306; processing of user preferences or user profiles in wireless networks H04W8/18)} - H04N 21/26291: {for providing content or additional data updates, e.g. updating software modules, stored at the client (deployment, distribution, installation, update of software G06F8/65; error detection or correction during software upgrading G06F11/1433; arrangements for updating broadcast information or broadcast-related information H04H60/25)} - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A system and method arranged to predict a match state event by a machine learning platform. The system and method including monitoring, by the machine learning platform, multimedia content on a live data feed, receiving, by the machine learning platform, current match state data corresponding to the multimedia content, and predicting, by the machine learning platform, an anomaly in the multimedia content based on an analysis of received multimedia content and current match state data. The system and method can include generating, by the machine learning platform, a confidence score regarding a likelihood of the anomaly in the multimedia content, and resolving, by the machine learning platform, the anomaly based on the confidence score. Resolving the anomaly includes extracting a portion of the multimedia content containing the anomaly, ignoring the anomaly, or correcting the anomaly. ## Claims ### Claim 1 A computer-implemented method for predicting a match state event by a machine learning platform on a server, the method comprising: receiving at the server, by the machine learning platform, a live data feed from an application program interface (API); analyzing at the server, by the machine learning platform, multimedia content on the live data feed; receiving at the server, by the machine learning platform, current match state data corresponding to the multimedia content, wherein the current match state data is based on a current state of a live match; predicting at the server, by the machine learning platform, an anomaly in the multimedia content based on the analysis of received multimedia content and an analysis of the current match state data; resolving at the server, by the machine learning platform, the predicted anomaly by (i) encoding a portion of the multimedia content containing the predicted anomaly, (ii) storing the serialized anomaly and a corresponding serialized function that specifies how to resolve the anomaly, and (iii) executing the serialized function based on the serialized anomaly, wherein executing the serialized function causes extracting a portion of the multimedia content containing the anomaly, ignoring the anomaly, or correcting the anomaly; predicting at the server, by the machine learning platform, a future match state based on the current match state data and the multimedia content, including the resolved anomaly; and sending, by a transmitter, the future match state to a communicating device. ### Claim 2 (depends on 1) The computer-implemented method in claim 1, further comprising: forwarding the extracted portion of the multimedia content to an intelligent technology platform to analyze and predict a future match state event based, at least in part, on the multimedia content. ### Claim 3 (depends on 1) The computer-implemented method in claim 1, further comprising: receiving, by the machine learning platform, historical event data comprising past live data feed updates. ### Claim 4 (depends on 1) The computer-implemented method in claim 1, further comprising: receiving, by the machine learning platform, historical event data comprising past match states. ### Claim 5 (depends on 1) The computer-implemented method in claim 1, further comprising: receiving, by the machine learning platform, historical event data comprising past live data feed updates and past match states. ### Claim 6 (depends on 1) The computer-implemented method in claim 1, wherein the predicting, by the machine learning platform, the anomaly in the multimedia content includes analyzing past live data feed updates and past match states. ### Claim 7 (depends on 1) The computer-implemented method in claim 1, further comprising: determining processor computing time or memory capacity required to correct the anomaly. ### Claim 8 (depends on 1) The computer-implemented method in claim 1, wherein resolving the anomaly comprises: predicting, by the machine learning platform, correctability of the anomaly. ### Claim 9 (depends on 8) The computer-implemented method in claim 8, wherein correctability of the anomaly comprises waiting for an update of the multimedia content that resolves the anomaly. ### Claim 10 A system for predicting a match state event by a machine learning platform on a server, the system comprising: a communication unit that includes a transmitter and a receiver, the receiver being arranged to receive a live data feed; a match state collector that includes a memory arranged to store match state data; an anomaly predictor having a machine learning system arranged to monitor multimedia content on the live data feed, receive current match state data that is based on a current state of a live match and that corresponds to the multimedia content, and predict, by a machine learning platform, accuracy of the multimedia content based on the multimedia content and corresponding match state data, utilizing one or more specialized neural network models trained to interpret the multimedia content; and an error corrector having a computing device arranged to resolve an anomaly in the multimedia content based on the predicted accuracy of the multimedia content, wherein resolving the anomaly includes (i) encoding a portion of the multimedia content containing the predicted anomaly, (ii) storing the serialized anomaly and a corresponding serialized function that specifies how to resolve the anomaly, and (iii) executing the serialized function based on the serialized anomaly, wherein executing the serialized function causes extracting a portion of the multimedia content containing the anomaly, ignoring the anomaly, or correcting the anomaly; and a match state predictor having a machine learning event (MLE) model arranged to predict a future match state based on the current match state data and the multimedia content, including the resolved anomaly, wherein the transmitter is configured to send the future match state to a communicating device. ### Claim 11 (depends on 10) The system in claim 10, wherein the match state predictor is arranged to receive the multimedia content and predict a future event based, at least in part, on the multimedia content. ### Claim 12 (depends on 10) The system in claim 10, wherein the anomaly predictor is further arranged to receive historical event data comprising past live data feed updates. ### Claim 13 (depends on 10) The system in claim 10, wherein the anomaly predictor is further arranged to receive historical event data comprising past match states. ### Claim 14 (depends on 10) The system in claim 10, wherein the anomaly predictor is further arranged to receive historical event data comprising past live data feed updates and past match states. ### Claim 15 (depends on 10) The system in claim 10, wherein the machine learning system is trained to: receive historical event data comprising past live data feed updates and past match states; and predict the anomaly in the multimedia content based, at least in part, on the past live data feed updates and the past match states. ### Claim 16 (depends on 10) The system in claim 10, wherein the machine learning system is trained to: determine processor computing time or memory capacity required to correct the anomaly. ### Claim 17 (depends on 10) The system in claim 10, wherein the machine learning system is trained to: predict correctability of the anomaly. ### Claim 18 (depends on 17) The system in claim 17, wherein correctability of the anomaly comprises waiting for an update of the multimedia content that resolves the anomaly. ### Claim 19 A non-transitory computer-readable storage medium containing computer executable instructions that, when executed by a computing device coupled to a communication unit having a transmitter and a receiver and containing a machine learning system, cause the computing device to perform operations comprising: receiving, by the receiver, a live data feed; monitoring, by the machine learning system, multimedia content on the received live data feed; receiving, from a memory coupled to the computing device, current match state data corresponding to the multimedia content, the current match state data being based on a current state of a live match; predicting, by the machine learning system, an anomaly in the multimedia content based on an analysis of received multimedia content and the current match state data, wherein the analysis utilizes one or more specialized neural network model trained to interpret the multimedia content; resolving, by the machine learning system, the predicted anomaly by (i) encoding a portion of the multimedia content containing the predicted anomaly, (ii) storing the serialized anomaly and a corresponding serialized function that specifies how to resolve the anomaly, and (iii) executing the serialized function based on the serialized anomaly, wherein executing the serialized function causes: extracting a portion of the multimedia content containing the anomaly, ignoring the anomaly, or correcting the anomaly; predicting, by the machine learning system, a future match state based on the current match state data and the multimedia content, including the resolved anomaly; and sending, by the transmitter, the future match state to a communicating device. ### Claim 20 (depends on 19) The non-transitory computer-readable storage medium in claim 19, wherein the machine learning system is trained to determine processor computing time or memory capacity required to correct the anomaly.