T1 [Morning]: Information Visualization for Interactive Information Retrieval

  • Orland Hoeber (University of Regina)

As search tasks move beyond targeted search and into the domain of complex search, a substantial cognitive burden is placed on the searcher to craft and refine their queries, evaluate and explore among the search results, and ultimately make use of what is found. In such cases, information visualization techniques may be leveraged to enable searchers to perceive, interpret, and make sense of the information available throughout the search process. This tutorial will establish the fundamental principles and theories of information visualization, explain how information visualization can support interactive information retrieval, and survey search interfaces from my own research that leverage information visualization techniques. The goal of this tutorial will be to encourage researchers to make informed design decisions for how to integrate information visualization into their own interactive information retrieval projects.

T2 [Morning]: Practical deep learning for recommender systems

  • Oleksandr Zakharchuk (Cograma)

The ability to provide high quality personalized recommendations is among the most significant types of competitive advantage an online business can have. However, even having vast amounts of data, creating a recommender system is far from being trivial. This tutorial covers applying deep learning models for creating personalized recommendations, as well as some of the typical problems encountered when working on production recommender systems and possible solutions for these problems.

T3 [Afternoon]: Estimating models combining latent and measured variables: A tutorial on basics, applications and current developments in Structural Equation Models and their estimation using PLS Path Modeling

  • Markus Kattenbeck (University of Regensburg)
  • David Elsweiler (University of Regensburg)

Structural Equation Modeling is a powerful statistical approach where measured variables and those which are latent can be combined in a single model. We propose a half-day tutorial where participants will learn about the statistical technique, its theoretical underpinnings and will gain enough insight to apply this technique in a practical sense to their own research problems.

T4 [Afternoon]: Visualizing and Exploring Scientific Literature with CiteSpace

  • Chaomei Chen (Drexel University)

This half-day tutorial aims to introduce the fundamental concepts, principles and methods of visualizing and exploring the development of a scientific knowledge domain. The tutorial explains the design rationale and various applications of CiteSpace – a freely available tool for interactive and exploratory analysis of the evolution of a scientific domain, ranging from a single specialty to multiple interrelated scientific frontiers. The tutorial demonstrates the analytic procedure of applying CiteSpace to a diverse range of examples and how one may interpret various patterns and trends revealed by interactive visual analytics.magnetic field, applied along the easy axis of the elements.