Cost-Effective Active Learning for Deep Image Classification

Wang, Keze; Zhang, Dongyu; Li, Ya; Zhang, Ruimao; Li, Lin

doi:10.1109/tcsvt.2016.2589879

Cited by 594 publications

(397 citation statements)

References 35 publications

Supporting

Mentioning

387

Contrasting

Unclassified

Order By: Relevance

“…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]. In automatic speech recognition, self-training-style approaches have seen some success in hybrid, alignmentbased speech systems.…”

Section: Related Workmentioning

confidence: 99%

Self-Training for End-to-End Speech Recognition

Kahn

Lee

Hannun

2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

195

154

View full text Add to dashboard Cite

We revisit self-training in the context of end-to-end speech recognition. We demonstrate that training with pseudo-labels can substantially improve the accuracy of a baseline model by leveraging unlabelled data. Key to our approach are a strong baseline acoustic and language model used to generate the pseudo-labels, a robust and stable beam-search decoder, and a novel ensemble approach used to increase pseudo-label diversity. Experiments on the LibriSpeech corpus show that selftraining with a single model can yield a 21% relative WER improvement on clean data over a baseline trained on 100 hours of labelled data. We also evaluate label filtering approaches to increase pseudo-label quality. With an ensemble of six models in conjunction with label filtering, self-training yields a 26% relative improvement and bridges 55.6% of the gap between the baseline and an oracle model trained with all of the labels.

show abstract

Section: Related Workmentioning

confidence: 99%

Self-Training for End-to-End Speech Recognition

Kahn

Lee

Hannun

2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

195

154

View full text Add to dashboard Cite

show abstract

“…In [67] and [68], uncertainty-based active learning criteria for deep models are proposed. The authors offer several met-rics to estimate model uncertainty, including least confidence, margin or entropy sampling.…”

Section: Active Learning For Deep Architecturesmentioning

confidence: 99%

Active and Incremental Learning with Weak Supervision

2020

View full text Add to dashboard Cite

Note: This paper is an extended version of our previous work [4], from which certain parts of sections 2 to 5 (except novel YOLO-specific methods) were taken verbatim. Section 8 contains some verbatim parts from our previous work [33].Abstract Large amounts of labeled training data are one of the main contributors to the great success that deep models have achieved in the past. Label acquisition for tasks other than benchmarks can pose a challenge due to requirements of both funding and expertise. By selecting unlabeled examples that are promising in terms of model improvement and only asking for respective labels, active learning can increase the efficiency of the labeling process in terms of time and cost.In this work, we describe combinations of an incremental learning scheme and methods of active learning. These allow for continuous exploration of newly observed unlabeled data. We describe selection criteria based on model uncertainty as well as expected model output change (E-MOC). An object detection task is evaluated in a continuous exploration context on the PASCAL VOC dataset. We also validate a weakly supervised system based on active and incremental learning in a real-world biodiversity application where images from camera traps are analyzed. Labeling only 32 images by accepting or rejecting proposals gen-

show abstract

“…In addition, most of existing incremental approaches suffer from noisy samples or outliers in the model updating. In this work, we propose a novel active self-paced learning framework (ASPL) to handle the above difficulties, which absorbs powers of two recently rising techniques: active learning (AL) [12], [13] and self-paced learning (SPL) [14], [15], [16]. In particular, our framework tends to conduct a "Cost-less-Earn-more" working manner: as much as possible pursuing a high performance while reducing costs.…”

Section: Introductionmentioning

confidence: 99%

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Lin

Wang

Meng

et al. 2018

IEEE Trans. Pattern Anal. Mach. Intell.

Self Cite

151

View full text Add to dashboard Cite

Abstract-This paper aims to develop a novel cost-effective framework for face identification, which progressively maintains a batch of classifiers with the increasing face images of different individuals. By naturally combining two recently rising techniques: active learning (AL) and self-paced learning (SPL), our framework is capable of automatically annotating new instances and incorporating them into training under weak expert recertification. We first initialize the classifier using a few annotated samples for each individual, and extract image features using the convolutional neural nets. Then, a number of candidates are selected from the unannotated samples for classifier updating, in which we apply the current classifiers ranking the samples by the prediction confidence. In particular, our approach utilizes the high-confidence and low-confidence samples in the selfpaced and the active user-query way, respectively. The neural nets are later fine-tuned based on the updated classifiers. Such heuristic implementation is formulated as solving a concise active SPL optimization problem, which also advances the SPL development by supplementing a rational dynamic curriculum constraint. The new model finely accords with the "instructorstudent-collaborative" learning mode in human education. The advantages of this proposed framework are two-folds: i) The required number of annotated samples is significantly decreased while the comparable performance is guaranteed. A dramatic reduction of user effort is also achieved over other state-of-theart active learning techniques. ii) The mixture of SPL and AL effectively improves not only the classifier accuracy compared to existing AL/SPL methods but also the robustness against noisy data. We evaluate our framework on two challenging datasets, which include hundreds of persons under diverse conditions, and demonstrate very promising results. Please find the code of this project at: http://hcp.sysu.edu.cn/projects/aspl/

show abstract

Cost-Effective Active Learning for Deep Image Classification

Cited by 594 publications

References 35 publications

Self-Training for End-to-End Speech Recognition

Self-Training for End-to-End Speech Recognition

Active and Incremental Learning with Weak Supervision

Active Self-Paced Learning for Cost-Effective and Progressive Face Identification

Contact Info

Product

Resources

About