Back to results
US 12,591,825 B2Granted
Computing system and method for progress tracking using a large language model
AssigneeProcore Technologies, Inc.
InventorsMatt Man, Farid Mirahadi, Hunter Buckhorn, Mohammad Mostafa Soltani, Patavee Meemeng
Life of this patent
Estimating the term…
Abstract
A computing platform is configured to (i) load source data comprising (a) a set of log entries of a construction project and (b) a schedule of the construction project, (ii) preprocess the source data, (iii) provide the preprocessed source data as input to a large-language model (LLM) that is configured to generate insights related to progress tracking and thereby determine one or more insights related to progress tracking for the construction project, and (iv) cause the one or more insights to be presented to one or more individuals involved in the construction project.
Claims20
A computing platform comprising:
- at least one processor;
- at least one non-transitory computer-readable medium;
- and
- program instructions stored on the at least one non-transitory computer-readable medium that, when executed by the at least one processor, cause the computing platform to:
- receive, from a client device, an indication of a user request for one or more types of insights related to progress tracking of a construction project having a schedule that includes a set of tasks, wherein the one or more types of insights comprise at least one of (i) one or more textual summaries of the set of tasks, or (ii) a status of a given type of material associated with the set of tasks, and wherein the user request is input via a graphical user interface (GUI) of the client device;
- and
- based on the received indication of the user request:
- load source data comprising (i) a set of log entries of the construction project and (ii) a representation of the schedule of the construction project, wherein each log entry in the set of log entries has an associated log entry type;
- preprocess the source data by:
- determining a set of log entry types that correspond to the one or more types of insights;
- based on an evaluation of the set of log entries and the set of tasks, determining, from the set of log entries, a subset of log entries that each (i) has an associated log entry type included in the set of log entry types and (ii) corresponds to at least one task in the set of tasks;
- and
- generating an input data structure according to a schema for input into a large language model (LLM), wherein the input data structure comprises a request to generate the one or more types of insights and a log dataset that is generated by, for each log entry in the subset of log entries:
- extracting log data from a log entry according to the associated log entry type of the log entry;
- and
- encoding the extracted log data from the log entry into the log dataset according to the schema;
- provide the generated input data structure as input to the LLM;
- obtain an output of the LLM that is responsive to the user request and comprises one or more insights related to the progress tracking of the construction project;
- format the output of the LLM into a form that is usable by the client device;
- and
- transmit the formatted output of the LLM to the client device and thereby cause the formatted output of the LLM to be presented via the GUI of the client device.
18 more elements
A non-transitory computer-readable medium, wherein the non-transitory computer-readable medium is provisioned with program instructions that, when executed by at least one processor, cause a computing platform to:
- receive, from a client device, an indication of a user request for one or more types of insights related to progress tracking of a construction project having a schedule that includes a set of tasks, wherein the one or more types of insights comprise at least one of (i) one or more textual summaries of the set of tasks, or (ii) a status of a given type of material associated with the set of tasks, and wherein the user request is input via a graphical user interface (GUI) of the client device;
- and
- based on the received indication of the user request:
- load source data comprising (i) a set of log entries of the construction project and (ii) a representation of the schedule of the construction project, wherein each log entry in the set of log entries has an associated log entry type;
- preprocess the source data by:
- determining a set of log entry types that correspond to the one or more types of insights:
- based on an evaluation of the set of log entries and the set of tasks, determining, from the set of log entries, a subset of log entries that each (i) has an associated log entry type included in the set of log entry types and (ii) corresponds to at least one task in the set of tasks;
- and
- generating an input data structure according to a schema for input into a large language model (LLM), wherein the input data structure comprises a request to generate the one or more types of insights and a log dataset that is generated by, for each log entry in the subset of log entries:
- extracting log data from a log entry according to the associated log entry type of the log entry;
- and
- encoding the extracted log data from the log entry into the log dataset according to the schema;
- provide the generated input data structure as input to the LLM;
- obtain an output of the LLM that is responsive to the user request and comprises one or more insights related to the progress tracking of the construction project;
- format the output of the LLM into a form that is usable by the client device;
- and
- transmit the formatted output of the LLM to the client device and thereby cause the formatted output of the LLM to be presented via the GUI of the client device.
14 more elements
A method implemented by a computing platform, the method comprising:
- receiving, from a client device, an indication of a user request for one or more types of insights related to progress tracking of a construction project having a schedule that includes a set of tasks, wherein the one or more types of insights comprise at least one of (i) one or more textual summaries of the set of tasks, or (ii) a status of a given type of material associated with the set of tasks, and wherein the user request is input via a graphical user interface (GUI) of the client device;
- and
- based on the received indication of the user request:
- loading source data comprising (i) a set of log entries of the construction project and (ii) a representation of the schedule of the construction project, wherein each log entry in the set of log entries has an associated log entry type;
- preprocessing the source data by:
- determining a set of log entry types that correspond to the one or more types of insights:
- based on an evaluation of the set of log entries and the set of tasks, determining, from the set of log entries, a subset of log entries that each (i) has an associated log entry type included in the set of log entry types and (ii) corresponds to at least one task in the set of tasks;
- and
- generating an input data structure according to a schema for input into a large language model (LLM), wherein the input data structure comprises a request to generate the one or more types of insights and a log dataset that is generated by, for each log entry in the subset of log entries:
- extracting log data from a log entry according to the associated log entry type of the log entry;
- and
- encoding the extracted log data from the log entry into the log dataset according to the schema;
- providing the generated input data structure as input to the LLM;
- obtaining an output of the LLM that is responsive to the user request and comprises one or more insights related to the progress tracking of the construction project;
- formatting the output of the LLM into a form that is usable by the client device;
- and
- transmitting the formatted output of the LLM to the client device, thereby causing the formatted output of the LLM to be presented via the GUI of the client device.
14 more elements
Dependencies are inferred from the claim text.