Deep learning has shown great performance in various domains of computer vision and machine learning. Affective Computing is still far behind these advances, despite its large number of practical applications in health, robotics, entertainment and marketing, among others. Traditional models for automated analysis and recognition of human affect (emotions, moods, social interactions, and so on) rely mainly on well-understood analytical models (e.g., SVMs, CRFs and DBNs). This is in part due to (i) the lack of suitable and (annotated) affect data, and (ii) limited ability of existing 'deep' models to explain the cause-effect relationships in such data. The latter is particularly important when, for instance, modeling the health data. Fortunately, over the last few years, the number of (annotated) affect datasets has started growing considerably. This paves the way for the design of a new class of models for affect analysis and recognition: Deep Affective Models. 

The goal of the DAL-COM 2017 workshop is to bridge the present gap between the affective computing and deep learning by exploring, in a highly focused manner, the most recent advances in these emerging fields. With the keynote speakers who are the top experts in deep learning (Ruslan Salakhutdinov), and social/affective computing (Tanzeem Choudhury), as well as with the expected high-quality oral and poster presentations, the DAL-COM 2017 workshop will provide you with the latest insights into the cutting-edge deep learning for Affective Computing and Context Modeling.

See you in Hawaii! Aloha!

Organizers