Cost-Effective Active Learning for Deep Image Classification

By Keze Wang, Dongyu Zhang, Ya Li, Ruimao Zhang, Liang Lin. 2017

750 citations recorded by the source corpus.

Recent successes in learning-based image classification, however, heavily rely on the large number of annotated training samples, which may require considerable human effort. In this paper, we propose a novel active learning (AL) framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner. Our approach advances the existing AL methods in two aspects. First, we incorporate deep convolutional neural networks into AL. Through the properly designed framework, the feature representation and the classifier can be simultaneously updated with progressively annotated informative samples. Second, we present a cost-effective sample selection strategy to improve the classification performance with less manual annotations. Unlike traditional methods focusing on only the uncertain samples of low prediction confidence, we especially discover the large amount of high-confidence samples from the unlabeled set for feature learning. Specifically, these high-confidence samples are automatically selected and iteratively assigned pseudolabels. We thus call our framework cost-effective AL (CEAL) standing for the two advantages. Extensive experiments demonstrate that the proposed CEAL framework can achieve promising results on two challenging image classification data sets, i.e., face recognition on the cross-age celebrity face recognition data set database and object categorization on Caltech-256.

Subfields: Artificial Intelligence, Radiology, Nuclear Medicine and Imaging

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Deep Active Learning with a Neural Architecture Search (2018, methodology): “[28] applied the well-known softmax response (SR) idea supplemented with pseudo-labeling (self-labeling of highly confident points) for active learning.”
Replay in Deep Learning: Current Approaches and Missing Biological Elements (2021, neutral): “The sample selection problem is also closely related to active learning strategies (Cohn, Atlas, & Ladner, 1994; Lin & Parikh, 2017; Wang, Zhang, Li, Zhang, & Lin, 2016; Settles, 2009; Yoo & So Kweon, 2019; Wei, Iyer, & Bilmes, 2015), with the most common selection methods using uncertainty…”
Reliable Uncertainty Estimates in Deep Neural Networks using Noise Contrastive Priors (2018, background): “Approaches include maximum entropy (Joshi et al., 2009), distance from the decision boundary (Tong and Koller, 2001), pseudo labelling high conﬁdence examples (Wang et al., 2017), and mixtures of information density and uncertainty measures (Li and Guo, 2013).”
Learning to Play with Intrinsically-Motivated Self-Aware Agents (2018, methodology): “Recent methods obtain diversiﬁed sets of hard examples [8, 39], or use conﬁdence-based heuristics to determine when to query for more labels [45].”
Learning Loss for Active Learning (2019, methodology): “Do these methods, advanced with small models and data, well scale to large deep networks [23, 17] and data? Fortunately, the uncertainty approach [28, 55] for classification tasks still performs well despite its simplicity.”
Adversarial Active Learning for Deep Networks: a Margin Based Approach (2018, methodology): “Uncertainty selection has been improved in a pseudo-labeling method called CEAL (Wang et al., 2016): CEAL performs uncertainty selection, but also adds highly confident samples into the increased training set.”
Comparison of Different Classifiers with Active Learning to Support Quality Control in Nucleus Segmentation in Pathology Images (2018, methodology): “11,12,13 In the field of biomedical image classification, it was utilized for either generating better feature representation for images 14,15 or creating classification models 16,17,18,19 iteratively.”
A Survey of Deep Active Learning (2020, methodology): “ut in testing for better confidence estimation, the analysis presented in [51] contends that the performance of this method is similar to the performance of using neural network softmax response (SR) [166] as uncertainty sampling, which requires vigilance. In addition, DFAL [36] pointed out that the”
Self-Training for End-to-End Speech Recognition (2019, methodology): “Self-training has been applied to tasks in natural language processing including word-sense disambiguation [2], noun identification [23] and parsing [3], in addition to tasks in computer vision such as object detection [24, 25] and image classification [26].”
T-CNN: Tubelets With Convolutional Neural Networks for Object Detection From Videos (2016, background): “When they are directly applied to videos in a frame-by-frame manner, they might miss some positive samples because objects might not be of their best poses at certain frames of videos.”
Variational Adversarial Active Learning (2019, methodology): “Pool-based methods can be grouped into three major categories as follows: uncertainty-based methods [19, 51, 1], representation-based models [40], and their combination [53, 38].”
Artificial Intelligence and Marketing: Pitfalls and Opportunities (2020, methodology): “Likewise, in computer vision, a convolutional neural network will typically be trained on a set of pictures of identical sizes (even if it means padding or resizing the original pictures to similar dimensions) (Wang, Zhang, Li, Zhang, & Lin, 2016; Jordan & Mitchell, 2015).”
Private Model Compression via Knowledge Distillation (2018, result): “We compare it with three other selection methods: random, margin (Wang et al. 2017), and diverse (Wang and Ye 2013).”
Selective Annotation Makes Language Models Better Few-Shot Learners (2022, methodology): “Most active learning methods involve iterative parameter updates (e.g., Wang et al., 2017; Kasai et al., 2019), which are computationally expensive for large language models used in in-context learning.”
BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learning (2019, background): “[35], Sener and Savarese [29], Samarth Sinha [28] explore combining it with AL.”
Selection Via Proxy: Efficient Data Selection For Deep Learning (2019, methodology): “To make active learning more feasible for deep learning, Sener and Savarese [2018] investigated the batch setting and proposed a core-set selection approach to active learning that outperformed existing techniques [Wang and Ye, 2015, Wang et al., 2016, Gal et al., 2017].”
Active Learning for Convolutional Neural Networks: A Core-Set Approach (2017, methodology): “Wang et al. (2016) propose an heuristic based algorithm which directly assigns labels to the data points with high confidence and queries labels for the ones with low confidence.”
Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision (2020, methodology): “Also the entropy of the model outputs [29] is used as a confidence measurement for transferring samples across domains [24].”