The second workshop on Theory-Informed User Modeling for Tailoring and Personalizing Interfaces (HUMANIZE)1 took place in conjunction with the 23rd annual meeting of the intelligent user interfaces (IUI)2 community in Tokyo, Japan on March 11, 2018. The goal of the workshop was to attract researchers from different fields by accepting contributions on the intersection of practical data mining methods and theoretical knowledge for personalization. A total of eight papers were accepted for this edition of the workshop.
Author Keywords User modeling, personalization, tailoring, user interfaces INTRODUCTION When designing interfaces practitioners often rely on knowledge and experience about the interface’s intended users and ©2018. Copyright for the individual papers remains with the authors. Copying permitted for private and academic purposes.
HUMANIZE ’18, March 11, 2018, Tokyo, Japan their needs in order to provide the optimal interface for its users. When creating user interfaces that can be personalized, quite often a more data-driven approach is taken, where practitioners rely on methods that use implicit or explicit feedback to prescribe how to alter an interface.
The current workshop aims at soliciting work that investigates the potential of combining the more practical data mining/machine learning methods with a more theory-driven approach. Three aspects play an important role in taking a more theorydriven approach to personalization: 1. How to consider the users of a system and their individual differences. 2. How to infer these individual differences from interaction data. 3. How to translate individual differences into interface designs.
The characteristics that play a role in what a user needs or wants from a system need to be investigated. Knowing what users differ on allows us to alter the interface. These characteristics can then be used to construct a user model containing this information. Examples of characteristics that may play a role in how to design an optimal interface are cognitive style, personality, and susceptibility to persuasive strategies. Secondly, there is a challenge of profiling a user in terms of these characteristic based on interaction data. Several approaches exist for this more computational challenge, for example mining data from social media and clickstream analysis. A third aspect is knowing how to adapt an interface to match a certain type of user. When a user’s characteristics are known, the interface can be altered to match this user. For example by reducing the number of search results for users under high cognitive load, or manipulating diversity.
These challenges are interconnected and there is no natural order in which these aspects need to be addressed when personalizing an interface. For example, by analyzing behavior data we can identify potential individual characteristics that play a role in people’s needs.
HUMANIZE provides scholars and practitioners in the field of personalized user interfaces with a venue to discuss and explore the commonalities between the sub-problems involved with user interface personalization.
An non-exhaustive list of topics for this workshop:
Identifying models that are (expected to be) useful for
personalizing user interfaces (e.g., personality, level of domain
knowledge, need for cognition, cognitive styles)
Data mining methods to infer user profiles in terms of
cognitive/psychological user characteristics from data (e.g., how
to infer personality from social media or domain knowledge
from clickstreams)
Theory on how to tailor interfaces to better match certain
user profiles (e.g., altering the number of search results,
ordering of interface elements, visual versus textual
representations)
User studies investigating one or more of the aspects
mentioned above
CONTRIBUTIONS
A total of eight papers were accepted: 3 long papers, 4 short
papers, and 1 position paper. Papers were categorized into one
of the three sessions: 1) Personality, 2) Social, or 3) Health &
Wellbeing. Below a description of the accepted papers:
Personality. Lay and Ferwerda [
Social. Xu and Lee [