AI Model Training Using Clustered Key Samples for Game Actions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing game AI training methods struggle to effectively improve game experience by adequately training AI models, leading to deviations and suboptimal performance due to uniform weighting of all game samples, including those that are not critical to the game outcome.
Innovation Solution
A method involving offline game recording to obtain training samples, clustering these samples to identify key or significant samples, and assigning higher weights to these samples during network training, enhancing their impact on the network parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If uniform weighting is applied to all training samples, then the training process is simple to implement, but the model training effect deteriorates due to insufficient focus on critical game moments
Solution Approach 1:
The patent applies local quality by assigning different weights to different training samples based on their importance. Key samples that capture critical game moments are identified through clustering and given higher weights, while less important samples receive lower weights. This differentiated weighting strategy improves model training effectiveness without significantly complicating the implementation process.
2Loss of time
If all training samples are processed equally, then the training time is reduced, but the convergence speed deteriorates due to lack of focus on significant samples
Solution Approach 1:
The patent implements partial action by focusing computational resources on a subset of key training samples identified through clustering. Rather than processing all samples uniformly, the method identifies and prioritizes samples that contain critical game moments, applying higher weights to these partial samples. This approach accelerates convergence while maintaining reasonable overall training time.
3Manufacturing precision
If clustering and weighting of key samples is implemented, then the model training effect is improved, but the computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the training samples into different clusters based on their characteristics and importance. The clustering process groups similar samples together, and key samples are identified from these clusters. This segmentation approach improves model training effectiveness while managing computational complexity through organized processing of sample groups rather than individual samples.
Data Source
AI summary
A model training method and apparatus, a storage medium, and a device, relating to the field of artificial intelligence (AI) technologies are provided. The method includes: obtaining a plurality of training samples, each training sample including an interaction screen and an action label, the action label indicating an interaction action adopted by a character object in the interaction screen; extracting features from the interaction screens included in the plurality of training samples, and performing clustering according to the extracted features, to obtain a clustering result; determining at least one key sample from the plurality of training samples according to the clustering result; and setting a weight for each training sample, and updating a network parameter of a deep network based on the plurality of training samples with the weights, the weight of each key sample being greater than the weight of another sample in the plurality of training samples.