Call for Papers: 3rd International Workshop on Deep Learning Practice for High-Dimensional Sparse Data (DLP‑KDD 2021)

DLP‑KDD Introduction

With the rapid growth of the Internet and mobile services, providing high‑quality personalized experiences has become essential for improving user satisfaction and revenue. Advances in hardware and massive data generation have propelled deep learning forward in image and language tasks, yet sparse, high‑dimensional data in core Internet applications such as advertising, recommendation, and search present unique challenges.

These challenges include severe over‑fitting, model redundancy, lack of interpretability, and heavy demands on computation, storage, and communication. Although models like Wide&Deep, DeepFM, DIN, DIEN, and MIMN have shown promising results, adapting deep learning to industrial‑scale sparse data remains difficult.

The DLP‑KDD workshop aims to systematically explore practical deep‑learning solutions for large‑scale industrial sparse data, fostering research and real‑world applications.

Organized jointly with leading industry partners (Alibaba, Microsoft, Huawei, Roku) and academic institutions (Shanghai Jiao‑Tong University, University of Utah), the workshop welcomes contributions on a broad range of topics.

Call for Papers Topics

1. Deep Learning System Construction and Optimization · Large‑Scale Recommendation and Retrieval Systems · High‑Throughput, Low‑Latency Real‑Time Serving · Scalable Distributed and Parallel Training for Deep Learning · Auto‑Machine Learning · Automatic Feature Selection

2. Data Representation and Mining · Representation Learning for High‑Dimensional Sparse Data · Embedding Techniques · Manifold and Dictionary Learning · Transfer Learning Applications · Meta‑Learning for Sparse Data · Explainable Deep Learning for High‑Dimensional Data · Data Augmentation · Anomaly Detection for High‑Dimensional Sparse Data · Generative Adversarial Networks for Sparse Data

3. User Modeling · Large‑Scale User Response Prediction · User Behavior Understanding · User Interest Mining

Other relevant topics, such as practical challenges and experiences of applying deep learning in industry, are also welcome.

Workshop Organization

Chair and program committee members include senior researchers from both academia and industry.

Important Dates and Information

Workshop website: https://dlp-kdd.github.io

Paper submission system: easychair.org/conferences/?conf=dlpkdd2021

Paper length: short papers (2–4 pages) or long papers (up to 9 pages).

Submission deadline: May 10 2021

Notification of acceptance: June 10 2021

Workshop dates: August 10–14 2021 (virtual), with an on‑site satellite session in Beijing (venue to be announced).

Accepted papers will be indexed in the ACM Digital Library or Springer, according to author preference.

For any questions, contact the program chairs via Zhihu private messages: @Zhu Xiaoqiang, @Zhou Guorui, @Wang Zhe, @Weinan Zhang, etc.