Workshop Home

Welcome to the 2013 KDD Mutimedia Data Mining (MDMKDD'13) workshop homepage!

The workshop will bring together experts in the analysis of digital media content, multimedia databases, knowledge engineers and domain experts from different applied disciplines with potential in multimedia data mining.  A new theme of this edition of the workshop is to present and discuss how multimedia is integrated into people’s daily life. With the large growth of devices and data, we want to focus this workshop on mining large scale rich content in a networked society.
The latest program schedule is available here.

We have two invited talks: a keynote speech by Professor Bing Liu and an invited presentation by Professor Qiaozhu Mei.
Professor Bing Liu  (University of Illinois at Chicago)

Title: Modeling Sentiment and Debates: Connecting Computer Science and Social Science

Sentiment analysis or opinion mining is the computational study of people’s opinions, sentiments, attitudes, and emotions expressed in written language. It is one of the most active research areas in natural language processing in recent years due to many challenging research problems and a wide range of practical applications. In this talk, I will start with discussing the mainstream sentiment analysis research and then move on to describe some more recent work on modeling debates and comments, which represents another kind of analysis of sentiment. Here the goal is to model, to mine, and to summarize different types of disagreement and agreement expressions, points of contention, questions and answers, arguing natures and interactions of participants, and whether the participants exhibit tolerance in their discussions/debates. Tolerance is an important concept in the field of communication, and is a sub-facet of deliberation which refers to critical thinking and exchange of rational arguments on an issue among participants that seek to achieve consensus/solution. This research naturally connects computer science and social science, especially communication and political sciences, in the analysis of social media. In the talk, I will also discuss a sentiment centric approach to social media analysis.  
 Professor Qiaozhu Mei  (University of Michigan, Ann Arbor)

Title:  Rumors, Information Needs, and Serendipity: three new text mining problems in Social Media. 

Abstract:  The wide availability of social media, such as the Tweet stream has presented unprecedented challenges and opportunities for text miners. In this talk, we will introduce our recent explorations of three new research problems concerned with microblogging: the detection and analysis of rumors, the analysis of users' information needs, and a quantitative study of the phenomenon of serendipity. Results of these novel research directions provide new insights on how to enhance user experience, how to increase user engagement, and how to achieve a “healthier” life cycle for information. 

For more information on KDD, please visit:


  • 07/21/2013: Professor Qiaozhu Mei (University of Michigan, Ann Arbor) will give an invited presentation at the workshop.
  • 07/18/2013: Our keynote speaker, Professor Bing Liu, will talk about "Modeling Sentiment and Debates: Connecting Computer Science and Social Science". 
  • 07/03/2013: Professor Bing Liu of University of Illinois at Chicago (UIC) has accepted our invitation to be the keynote speaker of the MDMKDD 2013 workshop.
  • 06/16/2013: Due to some administrative issues, we had to push back the acceptance notification date to June 19, 2013 (Wednesday).  We apologize for the delay.
  • 05/28/2013: We extended the submission deadline to June 7, 2013 (Friday)!  Submission site:
  • 03/23/2013: The MDM KDD 2013 workshop was accepted as a half day workshop at ACM KDD 2013!.

Workshop at a Glance:

The theme of this edition of the workshop is "mining largescale rich content in a networked society." Vast amount of multimedia are produced, shared, and accessed everyday in various social platforms. These multimedia objects (images, videos, texts, tags, etc.) represent rich, multifaceted recordings of human
behavior in the networked society, which lead to a range of social applications such as, (a) consumer behavior forecasting and socialdriven advertising / business, (b) local knowledge discovery (e.g., for tourism or shopping), and (c) detection of emergent news events and trends, and so on. In addition to techniques for mining single media items, all these applications require new methods for discovering robust features and stable relationships among the content of different media modalities and the users, in a dynamic, social contextrich, and likely noisy environment.

Mobile devices with multimedia sensors, such as cameras and geolocation sensors, have further integrated multimedia into people's daily life. New features, algorithms, and applications for mining the multimedia data collected at mobile devices can make these data of multiple modalities (image, video, geo, mobile data, etc.) accessible and useful in people’s daily life. Examples of such applications include (a) personal assistant, (b) augmented reality, (c) social applications, (d) entertainment, and so on. In addition to the research themes mentioned above, this workshop also welcomes submissions on various research topics of multimedia data mining, include but are not limited to the following:

● Measurement of relevance and user engagement in social media retrieval.
● Evaluation framework for the quality of rich content mining algorithms.
● Emerging technology of multimedia data mining for mobile applications.
● Emergent semantics analysis and topic detection from interrelated
multimedia objects.
● Social based business leveraging multimedia (e.g., social graph mining, social tagging, etc.).
● Data mining for location enhanced
services (e.g., maps, navigation and GIS systems).
● Fusion of multimedia features, metadata, user generated
content, and social context.
● Scalable or realtime architecture for largescale
media content processing and mining (e.g., parallel
computing, big data engineering, etc.).
● Multimedia data mining across platforms, including web and mobile devices.
● Scalable mobile multimedia computing (e.g., visual search, etc.).
● Predictive and prescriptive multimedia data modeling.
● Anomaly and outlier detection in multimedia databases.

Important Dates:

  • Submissions Due:  June 7, 2013 (Friday) 
  • Acceptance Notification: June 19, 2013 (Wednesday)  June 16, 2013 (Sunday) 
  • Camera-ready Due: July 28, 2013 (Friday)