Exploiting Counter-Examples for Active Learning with Partial labels

by   Fei Zhang, et al.

This paper studies a new problem, active learning with partial labels (ALPL). In this setting, an oracle annotates the query samples with partial labels, relaxing the oracle from the demanding accurate labeling process. To address ALPL, we first build an intuitive baseline that can be seamlessly incorporated into existing AL frameworks. Though effective, this baseline is still susceptible to the overfitting, and falls short of the representative partial-label-based samples during the query process. Drawing inspiration from human inference in cognitive science, where accurate inferences can be explicitly derived from counter-examples (CEs), our objective is to leverage this human-like learning pattern to tackle the overfitting while enhancing the process of selecting representative samples in ALPL. Specifically, we construct CEs by reversing the partial labels for each instance, and then we propose a simple but effective WorseNet to directly learn from this complementary pattern. By leveraging the distribution gap between WorseNet and the predictor, this adversarial evaluation manner could enhance both the performance of the predictor itself and the sample selection process, allowing the predictor to capture more accurate patterns in the data. Experimental results on five real-world datasets and four benchmark datasets show that our proposed method achieves comprehensive improvements over ten representative AL frameworks, highlighting the superiority of WorseNet. The source code will be available at <https://github.com/Ferenas/APLL>.


page 1

page 2

page 3

page 4


Single Shot Active Learning using Pseudo Annotators

Standard myopic active learning assumes that human annotations are alway...

Making Your First Choice: To Address Cold Start Problem in Vision Active Learning

Active learning promises to improve annotation efficiency by iteratively...

Active Learning in Incomplete Label Multiple Instance Multiple Label Learning

In multiple instance multiple label learning, each sample, a bag, consis...

Sampling Bias in Deep Active Classification: An Empirical Study

The exploding cost and time needed for data labeling and model training ...

Instance-Dependent Partial Label Learning

Partial label learning (PLL) is a typical weakly supervised learning pro...

Addressing practical challenges in Active Learning via a hybrid query strategy

Active Learning (AL) is a powerful tool to address modern machine learni...

Active Learning with Partial Feedback

In the large-scale multiclass setting, assigning labels often consists o...

Please sign up or login with your details

Forgot password? Click here to reset