SCOPE AND FOCUS

This workshop focuses on analyzing and understanding both the quantitative and qualitative spread of derivative ideas. Of particular interest are linguistic memes such as memorable quotations and variants thereof, unique hashtags and phrases, visual memes and their variants, and identifiable elements of viral videos. Another core focus is the tracking and prediction of viral spread, and the identification of derivative works. The target audience includes social scientists specializing in computer-mediated communications, network scientists, computational linguists, and researchers in the psychology of visual and audiovisual media.

Active research areas that are relevant to viral memes include:
• Predictive analytics of viral sharing of content
• Tracing the origin, provenance, and modification of viral material
• Understanding the inductive causes of virality, deliberate and intrinsic (cf. Kleinberg)
• Community detection and formation modeling using the spread of viral content
• Viral marketing
• Modeling of link types and relationship strength
• Path-based similarity measures and relationship extraction
• Applications to modeling of weblogs, social media, social networks, and the semantic web

The emphasis of this workshop shall be approaches based on analysis of viral content and spread from sources including but not limited to: social media, social news, collaboration networks, and document collections.

Application areas that often exhibit a need for viral memetic analysis include:
• Trends: trending topics in social media; hype index; communities, networks, and wikis (e.g., in education research)
• Internet humor: viral images, memes (cf. Know Your Meme)
• Online political communications: opinion, position, or campaign memes, lampoons
• Marketing and social recommender systems: communities, experts, friends, products, reviewers, providers
• Information diffusion and sharing systems: social media (opinions and sentiments, meme propagation, viral content, political commentary, etc.)
• Other behavioral modeling: community dynamics, recruitment and mass activity, large-scale patterns, traffic, spatiotemporal effects
• Application areas: social informatics of meme adaptation, public service announcements, cybersecurity (PSAs, anthropological aspects of security, trust networks)

INTENDED AUDIENCE AND IMPACT

This workshop is intended for researchers and practitioners in information systems that can be modeled using networks that exhibit some heterogeneity (i.e., differentiation among the entities and potential relationships represented by graphical elements). Examples of such heterogeneous information networks include community graphs with roles such as moderators and members, including models of social media that differentiate content providers, critics, and consumers; graphs of web pages that are annotated with paths and relationship strength indicators; and blogs and tweets with links or co-occurrence data.

Analyzing heterogeneous information networks involves the application of diverse new approaches from information extraction and integration, graph theory and algorithms, machine learning, topic modeling, knowledge representation, and uncertain reasoning. Researchers with interests in big data, social media, knowledge discovery in databases, cyber-physical systems, and informatics will also find this workshop of interest. In particular, prominent challenges exemplified by real-world problems in this area include how to account for and make use of behavioral patterns (including some large to colossal patterns), social dynamics including information propagation, relational characteristics, organizational structure, dynamic topic modeling, and concept drift.

We welcome paper submissions from researchers in all areas of heterogeneous information networks listed in the above section describing the workshop scope. We also hope to attract IJCAI participants from industrial R&D with interesting current applications that showcase aspects of heterogeneity in social and other networks.