The 1st workshop on Emotions and Personality in Personalized Services will be organized in conjunction with the UMAP 2013 conference and will be held in Rome.
While a lot of discussion has been made on filtering algorithms, and evaluation measures, few studies have stood to consider the role of emotions and personality in user models and personalized services.
Characterizing the user model and the whole user experience with personalized service, by means of affective traits, is an important issue which merits attention from researchers and practitioners in both web technology and human factor fields.
Some questions motivate this workshop:
- Do affective traits (personality, emotions, and mood) influence and determine the acceptance of the
- How personality traits should be included in the user model?
- How the personalized services should be adapted to emotions and mood to increase user satisfaction?
In the pursuit of increasing the quality of personalized services, researchers started to turn to more user-centric descriptors of content and services in recent years. The advances made in affective computing, especially in automatic emotion detection techniques, paved the way for the exploitation of emotions and personality as
descriptors that account for a larger part of variance in user behavior than the generic descriptors (e.g. genre of a multimedia content) used so far.
Emotions, users’ responses, can be characterized in different ways. The two most common approaches are (i) the discrete basic emotions (discrete classes, e.g. joy, sadness, fear, disgust, surprise, anger) and (ii) the continuous values, in the valence-arousal-dominance space. The affective computing community has been very
active in the past decade and has developed several methods for the automatic non-invasive detection of emotions via several modalities (Zeng et al., 2009).
While emotions can change pretty quickly, personality, on the other hand, describes long-lasting human traits. The most common way of describing personality is the five-factor model (openness, conscientiousness, extraversion, agreeableness and neuroticism).
Emotions and personality in personalized services (e.g., recommender systems) can be exploited in different ways at different stages in the service-usage (e.g. content consumption) chain (Tkalčič et. al, 2011). In the entry stage they can be used as a contextual parameter, as additional information to predict, assist and influence decision-making (Kahneman, 2011) or a way to diversify the personalization via the detection of serendipitous services. In the consumption stage, emotions can be used as additional tags for the characterization of the services, content and users (Jiao and Pantid, 2011), opening new research areas for modeling services and
content with different lengths. Finally, emotions can be exploited also for the non-invasive acquisition of the implicit user feedback as well as for novel evaluation metrics.
So far, research on emotions and personality in personalized services has been carried out in a scattered fashion. The goal of this workshop is to provide a venue for researchers to present their work, discuss it and benefit from the interaction.
- Affective modeling
- Emotions as context
- Emotions in the decision-making process for recommender systems
- Role of personality on user similarities
- Emotion detection in recommended content consumption
- Emotion detection as non-invasive feedback
- Affective tagging of multimedia content and services
- Emotion-based evaluation metrics (satisfaction…)
- Lifestyle recommender systems
- Personality and mood for group decision making
- Incorporating personality and emotions in user models
- Models based on personality
- Datasets for affective modeling (Collecting, Available)
- Personality traits acquisition (explicit vs. implicit)
- Assessing personality traits implicitly from users’ activities/ratings/behavior
- Personality and interfaces/control/bubble-control
- Could interfaces/control/bubble-control be personalized based on personality traits? Should they be?
- Personality and users’ tasks/goals
- Do personality traits influence users’ goals?
- Social signal processing for personalized services
- Strategies for modeling emotions and personality
- Recognizing triggers and causes of emotion
- Theories about the relationship between reasoning and affect, between decision-making and affect
- Methods for evaluating the utility of adaptation to affective factors
Tkalčič, M., Košir, A., & Tasič, J. (2011). Affective recommender systems : the role of emotions in recommender systems. Proceedings of the RecSys 2011 Workshop on Human Decision Making in Recommender Systems (Decisions@RecSys’11), 9–13.
Arapakis, I. (2010). Affect-Based Information Retrieval. Signal Processing. University of Glasgow. Retrieved from http://theses.gla.ac.uk/1867/01/2010arapakisphd.pdf
Kahneman, D. (2011). Thinking, Fast and Slow. Book (Vol. 1, p. 512). Farrar, Straus and Giroux. Retrieved from http://www.amazon.com/Thinking-Fast-Slow-Daniel-Kahneman/dp/0374275637
Zeng, Z., Pantic, M., Roisman, G. I., & Huang, T. S. (2009). A Survey of Affect Recognition Methods: Audio, Visual, and Spontaneous Expressions. IEEE Trans. Pattern Analysis & Machine Intelligence, Vol. 31, No., 1pp, 39–58.
Jiao, J., & Pantic, M. (2010). Implicit image tagging via facial information. Proceedings of the 2nd international workshop on Social signal processing – SSPW ’10 (p. 59). New York, New York, USA: ACM Press. doi:10.1145/1878116.1878133
Nunes, M. A. S. N., & Hu, R. (2012). Personality-based recommender systems. Proceedings of the sixth ACM conference on Recommender systems – RecSys ’12 (p. 5). New York, New York, USA: ACM Press.