CN-122022431-A - Task execution method and electronic device

Abstract

The application discloses a task execution method and electronic equipment, wherein the task execution method comprises the steps of determining target reference information from a plurality of historical cases and a plurality of pieces of historical knowledge according to target task information input by a user, wherein the target task information comprises instruction information and time sequence data to be processed, the historical cases represent task execution examples of historical tasks related to the historical time sequence data, the historical knowledge represents task execution experience related to the historical time sequence data, the target reference information comprises at least one of a target historical case and target historical knowledge, generating a target execution plan according to the instruction information, the time sequence data to be processed and the target reference information, and executing the target execution plan to obtain task execution results corresponding to the target task information.

Inventors

• ZHANG GUANYU
• NIE YIWEN

Assignees

• 联想(北京)有限公司

Dates

Publication Date

20260512

Application Date

20260130

Claims (10)

1. 1. A method of task execution, comprising: Determining target reference information from a plurality of historical cases and a plurality of pieces of historical knowledge according to target task information input by a user, wherein the target task information comprises instruction information and time sequence data to be processed, the historical cases represent task execution examples of historical tasks related to the historical time sequence data, the historical knowledge represents task execution experience related to the historical time sequence data, and the target reference information comprises at least one of target historical cases and target historical knowledge; Generating a target execution plan according to the instruction information, the time sequence data to be processed and the target reference information; And executing the target execution plan to obtain a task execution result corresponding to the target task information.
2. 2. The method of claim 1, wherein determining target reference information from a plurality of historical cases and a plurality of historical knowledge based on target task information entered by a user comprises: Determining target metadata and text description according to the instruction information; determining at least one of a first historical case and a target historical knowledge from the plurality of historical cases and the plurality of historical knowledge based on the target metadata and the textual description; And under the condition that the first historical case is determined, determining the target historical case from the first historical case according to the time sequence data to be processed.
3. 3. The method of claim 2, the determining the target historical case from the first historical case according to the time series data to be processed, comprising: determining target time sequence characteristics corresponding to the time sequence data to be processed; and determining the target historical case from the first historical cases according to the target time sequence characteristics and the historical time sequence characteristics corresponding to the historical time sequence data in each first historical case.
4. 4. The method of claim 2, further comprising: determining data anomaly information and statistical feature information according to the time sequence data to be processed; The determining at least one of a first historical case and a target historical knowledge from the plurality of historical cases and the plurality of historical knowledge based on the target metadata and the textual description, comprising: determining at least one of the first historical case and the target historical knowledge from the plurality of historical cases and the plurality of historical knowledge based on the data anomaly information, the statistical feature information, the target metadata, and the textual description.
5. 5. A method according to any one of claims 1 to 3, the target execution plan characterizing a linear execution plan consisting of at least a model selection plan and a model training plan; the task execution plan is executed to obtain a task execution result corresponding to the target task information, and the task execution result comprises the following steps: determining a first model according to the model selection scheme; Training the first model by using the time sequence data to be processed according to the model training scheme to obtain a second model; and generating a task execution result corresponding to the target task information by using the second model according to the target task information.
6. 6. The method of any of claims 1 to 5, the generating a target execution plan from the instruction information, the timing data to be processed, and the target reference information, comprising: Generating target prompt information according to the instruction information, the time sequence data to be processed and the target reference information; and generating the target execution plan by using a third model according to the target prompt information, wherein the third model at least comprises a large model.
7. 7. The method of any one of claims 1 to 6, further comprising: Determining a target case corresponding to the target task information according to the target task information, the execution process of executing the target execution plan and the task execution result; Storing the target case in a case database, wherein the case database stores the plurality of historical cases.
8. 8. The method of claim 7, the case database comprising second historical cases belonging to a same context type as the target case, the context type being a result of clustering cases in the case database; the method further comprises the steps of: Generating first knowledge by using a fourth model according to the target case and the second historical case; performing utility verification on the first knowledge; And in response to the first knowledge passing the utility verification, updating the first knowledge to a knowledge database, wherein the knowledge database stores the historical knowledge pieces.
9. 9. The method of claim 8, the updating the first knowledge to a knowledge database comprising: inputting the first knowledge and the second knowledge into the fourth model to generate third knowledge under the condition that the first knowledge and the second knowledge in the plurality of historical knowledge have semantic conflict; performing utility verification on the third knowledge; and updating the third knowledge to the knowledge database in response to the third knowledge passing the utility verification.
10. 10. An electronic device comprising a memory and at least one processor, wherein, The memory stores a computer program executable on a processor for, when executed by the processor: Determining target reference information from a plurality of historical cases and a plurality of pieces of historical knowledge according to target task information input by a user, wherein the target task information comprises instruction information and time sequence data to be processed, the historical cases represent task execution examples of historical tasks related to the historical time sequence data, the historical knowledge represents task execution experience related to the historical time sequence data, and the target reference information comprises at least one of target historical cases and target historical knowledge; Generating a target execution plan according to the instruction information, the time sequence data to be processed and the target reference information; And executing the target execution plan to obtain a task execution result corresponding to the target task information.

Description

Task execution method and electronic device Technical Field The present application relates to, but not limited to, the field of computer technologies, and in particular, to a task execution method and an electronic device. Background In the field of time series prediction (e.g., scenarios such as retail demand prediction, energy load prediction, and financial index prediction), building a high-performance and robust prediction model is a complex and empirically intensive task. The task usually involves a plurality of links such as data preprocessing, feature engineering, model selection, parameter adjustment and the like, so that the decision space is huge, the dependence on professional knowledge is strong, and the modeling efficiency is low and the cost is high. Disclosure of Invention In view of this, the present application at least provides a task execution method and an electronic device. The technical scheme of the application is realized as follows: in one aspect, the present application provides a task execution method, including: Determining target reference information from a plurality of historical cases and a plurality of pieces of historical knowledge according to target task information input by a user, wherein the target task information comprises instruction information and time sequence data to be processed; Generating a target execution plan according to the instruction information, the time sequence data to be processed and the target reference information; and executing the target execution plan to obtain a task execution result corresponding to the target task information. In some embodiments, determining target reference information from a plurality of historical cases and a plurality of historical knowledge based on target task information entered by a user includes: determining target metadata and text description according to the instruction information; Determining at least one of a first historical case and a target historical knowledge from a plurality of historical cases and a plurality of historical knowledge based on the target metadata and the textual description; In the case of determining the first history case, a target history case is determined from the first history case according to the time series data to be processed. In some embodiments, determining a target historical case from the first historical case according to the time series data to be processed includes: Determining target time sequence characteristics corresponding to time sequence data to be processed; and determining a target historical case from the first historical cases according to the target time sequence characteristics and the historical time sequence characteristics corresponding to the historical time sequence data in each first historical case. In some embodiments, the method further comprises: Determining data anomaly information and statistical feature information according to the time sequence data to be processed; determining at least one of the first historical case and the target historical knowledge from the plurality of historical cases and the plurality of historical knowledge based on the target metadata and the textual description, comprising: at least one of the first historical case and the target historical knowledge is determined from the plurality of historical cases and the plurality of historical knowledge based on the data anomaly information, the statistical feature information, the target metadata, and the text description. In some implementations, the target execution plan characterizes a linear execution scheme consisting of at least a model selection scheme and a model training scheme; executing the target execution plan to obtain a task execution result corresponding to the target task information, including: Determining a first model according to a model selection scheme; training a first model by utilizing time sequence data to be processed according to a model training scheme to obtain a second model; And generating a task execution result corresponding to the target task information by using the second model according to the target task information. In some embodiments, generating a target execution plan based on instruction information, timing data to be processed, and target reference information includes: generating target prompt information according to the instruction information, the time sequence data to be processed and the target reference information; and generating a target execution plan by using a third model according to the target prompt information, wherein the third model at least comprises a large model. In some embodiments, the method further comprises: Determining a target case corresponding to the target task information according to the target task information, an execution process of an execution target execution plan and a task execution result; Storing the target case in a case database, wherein the case database stores a plurality of hist