发布时间:2026年11月4日

News

* News, Online Zoom Link is available! (VERY IMPORTANT:  this link is for your actual Workshop Day ONLY)

https://zoom.us/j/97994459923?pwd=d2FTbzdEM2tGWEc3SldQc1UrUHd6Zz09

Time: August 24, 2020 Afternoon 1:00PM - 5:00PM (San Diego, CA, GMT-7)

Beijing Time:  August 25, 2020 Morning 4:00AM - 8:00AM (Beijing, GMT+8)

 

What is Intelligent Information Feed

Information feed is a linear form of displaying content including text, image, audio, video, or their combination.  It gained popularity on social media platforms such as Twitter and Facebook, and enjoys explosive growth outside of US especially in China.  There are a handful of mega Apps with over 100 million DAU (daily active users) in China, feeding news, articles, and short videos to billions of users.  Feeds are designed to fill in fragmented spare time with content that is easy to consume just like fast food.  When the amount of feeds exceeds the limit of display canvas or human attention capacity, intelligent mechanisms have to step in to reduce information overload.  As the sources of feeds become more abundant, Intelligent Information Feed (IIF) becomes ubiquitous.

The enabling technologies of Intelligent Information Feed touch upon areas such as recommendation, content creation, content understanding, user profiling, model prediction, and big data analytics, which are highly relevant to KDD.  In contrast to the red-hot feed industry and enthusiastic users, the response from the KDD research community has been sporadic. Meanwhile, the rapid development of the industry leaves behind deep technology debts that cannot be filled without fully engaging the research community.  The organizers of this workshop believe KDD is an ideal venue to call for immediate attention to this emerging field, which deserves a dedicated forum to study unique algorithmic aspects induced by the nature of fast consumption etc.  Following the success of the first workshop on IIF in conjunction with KDD 2019, we have officially named the second workshop as KDDFeed from now on, to emphasize it's connection with the KDD community.

Program

Date: August 24, 2020 Afternoon

All time below is at local time (GMT-7). You may check different time zones in the table below.
 

Session One

13:00--13:05    Opening

13:05--13:50    Jie Tang: Cognitive Graph and Its Applications  

13:50--14:40    Jun Xu: Beyond Probability Ranking Principle: Modeling the Dependencies among Documents 

Session Two

14:40--16:00 Paper presentation (20 mins * 4 papers)

  1. Zhengxiao Du, Xiaowei Wang and Hongxia Yang. Sequential Scenario-Specific Meta Learner for Online Recommendation. (Tsinghua University, Alibaba Group)
  2. Chen Li, Linmei Hu, Chuan Shi and Guojie Song. Sequential Recommendation with Heterogeneous Graph Neural Network. (Beijing University of Posts and Telecommunications, Peking University)
  3. Yukuo Cen, Jianwei Zhang, Xu Zou, Chang Zhou and Hongxia Yang. Controllable Multi-Interest Sequential Recommendation. (Tsinghua University, Alibaba Group)
  4. Zhen Yang, Ming Ding, Chang Zhou, Hongxia Yang and Jingren Zhou. Understanding Negative Sampling in Graph Representation Learning. (Tsinghua University, Alibaba Group)

Session Three

16:00--16:50    Ji Liu: Communication Efficient Distributed Training

16:50--17:00    Closing

 

Beijing

Taipei (local)

San Diego

CA

Urbana

Chicago

Montreal

New York

London

Paris

Ljubljana

GMT+8

PDT/GMT-7

CDT/GMT-5

EDT/GMT-4

GMT+1

GMT+2

4:00 AM

1:00 PM 

3:00 PM

4:00PM

9:00 PM

10:00 PM

 

Call-for-Papers

The target audience includes students, professors, and researchers from academic institutes and corporate labs, data scientists, and engineers and applied researchers from industry.  The scope covers but is not limited to the following topics:

  • ​Creation Algorithms that build tools for content creation.  Content includes text, image, audio, video, and their combination.  Automatic generating news from raw event data belongs to this category
  • Pre-processing Algorithms that prepare content for further processing.  Enhancing the resolution of low quality images or videos belongs to this category
  • Review and filtration Algorithms that detect and get rid of content that are not appropriate for wide distribution
  • Content Understanding Algorithms that extract structural and semantic metadata in the raw content for further process
  • Analytics Methods, metrics and algorithms that provide deep insight to improve the feed performance of any aspects.  Better user profiling or predicting popular content belongs to this category
  • Recommendation Algorithms that distribute the right content to the right users at the right time
  • Monetization Algorithms that distribute the content that carry brand information to the right audience for marketing purpose
  • Social responsibility Algorithms that prevent users from being addicted to feeds
  • User Understanding Algorithms that understand intent based on user behavior
  • Representation Learning Algorithms that learn representations of both content and user, and match them using the learned representations

The submission link: https://easychair.org/conferences/?conf=kddfeed2020

Paper submissions: May 31, 2020
Paper notifications: June 20, 2020
Full workshop programs online: June 30, 2020
Workshop date: August 24 (tentative), 2020

 

Organizers

   Ying Shan | Tencent | China

   Changhu Wang | ByteDance | China

   Ji Liu | Kwai Inc. | USA

   Jie Tang | Tsinghua University | China

Program Committees

  • Smriti Bhagat, Facebook
  • Mohsen Jamali, Linkedin
  • Longin Jan Latecki, Temple University
  • Krishtina Lerman, USC
  • Shuanglong Li, Baidu
  • Peter Lofgren, Stanford University
  • Ruihua Song, Microsoft
  • Huan Sun, OSU
  • Hao Tian, Baidu
  • Zhen Wen, Tencent
  • Jun Xu, Renmin University
  • Xiaolin Shi, Science Manager, Snap Inc.