In active learning, a query strategy is the method by which the algorithm chooses the next data points to query for labels. Rather than randomly labeling data, the model “asks questions” in the form of selecting unlabeled instances it finds most informative for improving its accuracy. The goal is to maximize model learning while minimizing the number of labeled examples required.Common query strategies include:Uncertainty sampling: select samples where the model is least certain of the prediction.Query-by-committee: use an ensemble of models and select samples where they disagree most.Diversity-based methods: select a set of points that are representative of the remaining pool.By employing a clever query strategy, active learning focuses labeling effort on the most valuable data, improving efficiency.
Data Selection & Data Viewer
Get data insights and find the perfect selection strategy
Learn MoreSelf-Supervised Pretraining
Leverage self-supervised learning to pretrain models
Learn MoreSmart Data Capturing on Device
Find only the most valuable data directly on devide
Learn More