Special Issue on Deep Learning with Small Samples (Submission Deadline: April 15, 2020)

1. Summary and Scope

In the machine learning and computer vision fields, due to the rapid development of deep learning, recent years have witnessed breakthroughs for large-sample classification tasks. However, it remains a persistent challenge to learn a deep neural network with good generalizability from only a small number of training samples. In fact, humans can easily learn the concept of a class from a small amount of data rather than from millions of data. Moreover, many types of real-world data are small in quantity and are expensive to collect or label. Motivated by this fact, research on deep learning with small samples becomes more and more prevalent in the communities of machine learning and computer vision, for example, researches focusing on one-shot classification, few-shot classification, as well as classification with small training samples.

Recently, deep small-sample learning has achieved promising performance for certain small-sample problems, by transferring the “knowledge” learned from other datasets containing rich labelled data or generating synthetic samples to approximate the distribution of real data. However, many challenging topics remain with small-sample deep leaning techniques, such as data augmentation, feature learning, prior construction, meta-learning, and fine tuning. Therefore, the goal of this special issue is to collect and publish the latest developments in various aspects of deep learning with small samples.

The list of possible topics includes, but is not limited to:

• Survey/vision/review of deep learning with small samples

• Data augmentation methods for small-sample leaning

• Feature learning based methods small-sample leaning

• Regularization technology of deep model in small-sample leaning

• Ensemble learning based methods for small-sample learning

• Transfer learning methods for small-sample learning

• Semi-supervised learning methods for small-sample learning

• Prior based methods for few-shot learning

• Meta-leaning based methods for few-shot learning

• Fine-tuning based methods for small-sample learning

• Theoretical analysis for small-sample learning

• Applications of small-sample learning on person re-identification, object recognition, etc.

2. Submission Guidelines

Authors should prepare their manuscripts according to the "Instructions for Authors" guidelines of “Neurocomputing” outlined at the journal website https://www.elsevier.com/journals/neurocomputing/0925-2312/guide-for-authors. All papers will be peer-reviewed following a regular reviewing procedure. Each submission should clearly demonstrate evidence of benefits to society or large communities. Originality and impact on society, in combination with a media-related focus and innovative technical aspects of the proposed solutions will be the major evaluation criteria.

When submitting their manuscripts, authors must select "VSI:DLSS" as the article type.

3. Important Dates

Submission Deadline: 15 Apr. 2020

First Review Decision: 15 Aug. 2020

Revisions Due: 15 Oct. 2020

Final Manuscript: 15 Jan. 2021

Expected publication date: 15 Mar. 2021

4. Guest Editors

Jing-Hao Xue, Associate Professor, University College London, UK

Jufeng Yang, Professor, Nankai University, China

Xiaoxu Li, Associate Professor, Lanzhou University of Technology, China

Yan Yan, Associate Professor, Xiamen University, China

Yujiu Yang, Associate Professor, Tsinghua University, China

Zongqing Lu, Assistant Professor, Tsinghua University, China

Zhanyu Ma, Professor, Beijing University of Posts and Telecommunications, China