Dash: Semi-Supervised Learning with Dynamic Thresholding

TL;DR

Dash introduces a dynamic thresholding framework for semi-supervised learning that adaptively selects unlabeled data based on loss, improving training efficiency and model performance with theoretical guarantees.

Contribution

The paper proposes Dash, a novel semi-supervised learning method that dynamically adjusts data selection thresholds, providing theoretical convergence guarantees and empirical improvements.

Findings

Dash outperforms state-of-the-art SSL methods on benchmarks.

Theoretical convergence rate of Dash is established.

Adaptive data selection enhances SSL training effectiveness.

Abstract

While semi-supervised learning (SSL) has received tremendous attentions in many machine learning tasks due to its successful use of unlabeled data, existing SSL algorithms use either all unlabeled examples or the unlabeled examples with a fixed high-confidence prediction during the training progress. However, it is possible that too many correct/wrong pseudo labeled examples are eliminated/selected. In this work we develop a simple yet powerful framework, whose key idea is to select a subset of training examples from the unlabeled data when performing existing SSL methods so that only the unlabeled examples with pseudo labels related to the labeled data will be used to train models. The selection is performed at each updating iteration by only keeping the examples whose losses are smaller than a given threshold that is dynamically adjusted through the iteration. Our proposed approach, Dash, enjoys its adaptivity in terms of unlabeled data selection and its theoretical guarantee. Specifically, we theoretically establish the convergence rate of Dash from the view of non-convex optimization. Finally, we empirically demonstrate the effectiveness of the proposed method in comparison with state-of-the-art over benchmarks.