Task model training method and device, electronic equipment and storage medium

Abstract

The invention provides a task model training method and device, electronic equipment and a storage medium, and relates to the technical field of neural networks. The method comprises the steps that a collected sample behavior sequence set is input into a pre-trained feature extractor, an output result is obtained, the sample behavior sequence set comprises game behavior sequences of at least one user under a target game in different time periods, each game behavior sequence is composed of at least one game behavior basic unit, wherein the output result is feature information corresponding to the game behavior sequence; the output result is input into a preset recognition model, a recognition result is obtained, and the preset recognition model is an initial target task recognition model and is used for recognizing a result corresponding to the initial target task; according to the recognition result, the feature extractor and the recognition model are adjusted, a target feature extractor and a target recognition model are obtained, and the task model is formed by cascading the target feature extractor and the target recognition model. The method can improve the efficiency of task model training.

Description

technical field [0001] The present application relates to the technical field of neural networks, and in particular, to a task model training method, device, electronic equipment, and storage medium. Background technique [0002] In order to improve the game experience of game players, various intelligent services are usually deployed in the game, such as: plug-in detection, map pre-loading detection and other services, so as to supervise the player's game behavior during the game and ensure the fairness of the game . The execution of various intelligent services can be realized by deploying corresponding task models. [0003] In the prior art, for the training of different task models, it is usually necessary to independently select and label sample data, so as to perform model training according to the labeled sample data. [0004] However, since data annotation usually relies on expert experience, which is time-consuming and labor-intensive, and different task models ar...

Images