English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This paper aims to develop a novel cost-effective framework for faceidentification, which progressively maintains a batch of classifiers with theincreasing face images of different individuals. By naturally combining tworecently rising techniques: active learning (AL) and self-paced learning (SPL),our framework is capable of automatically annotating new instances andincorporating them into training under weak expert re-certification. We firstinitialize the classifier using a few annotated samples for each individual,and extract image features using the convolutional neural nets. Then, a numberof candidates are selected from the unannotated samples for classifierupdating, in which we apply the current classifiers ranking the samples by theprediction confidence. In particular, our approach utilizes the high-confidenceand low-confidence samples in the self-paced and the active user-query way,respectively. The neural nets are later fine-tuned based on the updatedclassifiers. Such heuristic implementation is formulated as solving a conciseactive SPL optimization problem, which also advances the SPL development bysupplementing a rational dynamic curriculum constraint. The new model finelyaccords with the "instructor-student-collaborative" learning mode in humaneducation. The advantages of this proposed framework are two-folds: i) Therequired number of annotated samples is significantly decreased while thecomparable performance is guaranteed. A dramatic reduction of user effort isalso achieved over other state-of-the-art active learning techniques. ii) Themixture of SPL and AL effectively improves not only the classifier accuracycompared to existing AL/SPL methods but also the robustness against noisy data.We evaluate our framework on two challenging datasets, and demonstrate verypromising results. (http://hcp.sysu.edu.cn/projects/aspl/)

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