# US 11,122,100 B2 · Detecting events from ingested data > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,122,100 B2 - Document ID: US11122100B2 - Status: Granted - Kind: B2 - Application number: 16/679,665 - Filed: 2019-11-11 - Published: 2020-03-26 - Granted: 2021-09-14 - Earliest claimed: 2017-08-28 - Assignees: Banjo, Inc. - Applicants: Banjo, Inc. - Inventors: Rish Mehta; Damien Patton - Web page: https://patentgrep.com/patent/US11122100B2 ## CPC - G06F 9/542: {Event management; Broadcasting; Multicasting; Notifications} - G06F 9/5027: {the resource being a machine, e.g. CPUs, Servers, Terminals} - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract The present invention extends to methods, systems, and computer program products for detecting events from ingested data. Data is ingested. A first score is computed from a portion of the data. The first score indicates a likelihood of the data including event information. Computing resources are allocated to further process the data based on the first score indicating that further processing is warranted. A second score is computed from parameters of the data. An event is detected based on the second score. ## Claims ### Claim 1 A method comprising: ingesting data associated with a geographic region; computing a first score from a portion of the data, the first score indicating a likelihood of the data including event information related to an event type that occurs with less than a threshold frequency in the geographic region; determining further processing of the data is warranted based on the likelihood indicated by the first score; allocating computing resources to further process the data; extracting parameters from the data portion utilizing the allocated computing resources; computing a second score from the extracted parameters utilizing the allocated computing resources; and detecting a previously unidentified event of the event type in the geographic area based on the second score. ### Claim 2 (depends on 1) The method of claim 1, wherein ingesting data comprises ingesting a sensor data stream from a sensor selected from among: an air quality sensor or a smart city sensor. ### Claim 3 (depends on 1) The method of claim 1, further comprising identifying the data portion including: detecting a beginning condition in the data; and detecting an ending condition in the data. ### Claim 4 (depends on 3) The method of claim 3, wherein identifying the data portion comprises stitching different parts of the data together into a conversation. ### Claim 5 (depends on 1) The method of claim 1, wherein ingesting data comprises ingesting one of: a communication stream, a video stream, or an audio stream. ### Claim 6 A system comprising: a processor; system memory coupled to the processor and storing instructions configured to cause the processor to: ingest data associated with a geographic region; computing a first score from a portion of the data, the first score indicating a likelihood of the data including event information related to an event type that occurs with less than a threshold frequency in the geographic region; determine further processing of the data is warranted based on the likelihood indicated by the first score; allocate computing resources to further process the data; extract parameters from the data portion utilizing the allocated computing resources; compute a second score from the extracted parameters utilizing the allocated computing resources; and detect a previously unidentified event of the event type in the geographic area based on the second score. ### Claim 7 (depends on 6) The system of claim 6, wherein instructions configured to ingest data comprise instructions configured to ingest a sensor data stream from a sensor selected from among: an air quality sensor or a smart city sensor. ### Claim 8 (depends on 6) The system of claim 6, further comprising instructions configured to identify the data portion including: detect a beginning condition in the data; and detect an ending condition in the data. ### Claim 9 (depends on 8) The system of claim 8, wherein instructions configured to identify the data portion comprise instructions configured to stitch different parts of the data together into a conversation. ### Claim 10 (depends on 6) The system of claim 6, wherein instructions configured to ingest data comprise instructions configured to ingesting one of: a communication stream, a video stream, or an audio stream. ### Claim 11 A method comprising: ingesting an audio stream; computing a first score from a portion of the audio stream, the first score indicating a likelihood of the audio stream including event information related to an event type; determining further processing of the audio stream portion is warranted based on the likelihood indicated by the first score; allocating computing resources to further process the audio stream portion; computing a second score from parameters of the audio stream portion utilizing the allocated computing resources; and detecting a previously unidentified event of the event type based on the second score. ### Claim 12 (depends on 11) The method of claim 11, wherein computing a first score comprises computing a first score indicating a likelihood of the audio stream including event information related to an event type that occurs with less than a threshold frequency. ### Claim 13 (depends on 11) The method of claim 11, wherein computing a first score comprises computing a first score indicating a likelihood of the audio stream including event information related to an event type occurring within a geographic region; and wherein detecting a previously unidentified event comprises detecting a previously unidentified event within the geographic region. ### Claim 14 (depends on 11) The method of claim 11, wherein identifying the audio stream portion comprises: detecting a beginning condition in the audio stream; and detecting an ending condition in the audio stream. ### Claim 15 (depends on 14) The method of claim 14, wherein identifying the audio stream portion comprises stitching different parts of the audio stream together into a conversation. ### Claim 16 A system comprising: a processor; system memory coupled to the processor and storing instructions configured to cause the processor to: ingest an audio stream; compute a first score from a portion of the audio stream, the first score indicating a likelihood of the audio stream including event information related to an event type; determine further processing of the audio stream portion is warranted based on the likelihood indicated by the first score; allocate computing resources to further process the audio stream portion; compute a second score from parameters of the audio stream portion utilizing the allocated computing resources; and detect a previously unidentified event of the event type based on the second score. ### Claim 17 (depends on 16) The computer system of claim 16, wherein instructions configured to compute a first score comprise instructions configured to compute a first score indicating a likelihood of the audio stream including event information related to an event type that occurs with less than a threshold frequency. ### Claim 18 (depends on 11) The computer system of claim 11, wherein instructions configured to compute a first score comprise instructions configured to compute a first score indicating a likelihood of the audio stream including event information related to an event type occurring within a geographic region; and wherein instructions configured to detect a previously unidentified event comprise instructions configured to detect a previously unidentified event within the geographic region. ### Claim 19 (depends on 16) The computer system of claim 16, wherein instructions configured to identify the audio stream portion comprise instructions configured to: detect a beginning condition in the audio stream; and detect an ending condition in the audio stream. ### Claim 20 (depends on 19) The computer system of claim 19, wherein instructions configured to identify the audio stream portion comprise instructions configured to stitch different parts of the audio stream together into a conversation.