Thursday February 7, 2019 Venue: Elings 1601
Interactive Music Recommendation: Applications and Design (PDF)
Abstract: In this workshop we will explore interactive music recommendation systems through discussion of their applications, possible designs and technical requirements for their implementation. A large amount of research in recommender systems focuses on algorithmic accuracy. This trend is reflected in popular commercial music recommender systems (e.g. Spotify, Pandora, Google Play Music) that rely on fast, sophisticated algorithms to offer list based recommendations, while accepting limited user feedback via track likes or skips. However, recent work has unveiled the importance of other aspects of recommendation process, including explanation, transparency, control and user experience in general. Developing these aspects can not only improve consumer recommenders, but also expand the application of music recommendation to therapy, art installations and performance.
Workshop will be divided into two parts: presentation (1 hour) and organized discussion (2 hours). First, participants will learn about the state-of-the-art research, challenges and opportunities in the domain of music recommendation. Our focus will be on interactive systems, with rich interfaces, that enable transparency and control of the recommendation process, and allow for serendipity and novelty in music discovery. Experimental evidence has shown a strong relation between emotion and music and previous research in affect-based recommender systems produced improvements over their non-contextual alternatives. Therefore, we will pay special attention to moods and emotions as a context for recommending music artists and songs, and for visually representing music collection to enhance user’s understanding and control of the recommender. As a proof of concept, we will explore an interactive, web based artists recommender system which integrates content and mood-based filtering in a novel interface. During organized discussion, we will investigate the design possibilities and challenges of developing interactive music recommenders, covering the topics of applications, interface design and implementation. Specifically, we will discuss ways to identify music listening context, person’s emotional state and music preference, visual representation of music databases, modes of interaction with the interface and potential sources of music data and metadata used by recommendation algorithms.
Saturday February 9, 2019 Venue: Elings 1601
Her Stories in the Age of AI (PDF)
Artist(s): Eunsu Kang, Haru Ji, Maja Petric, Ha Na Lee, Jinsil Hwaryoung Seo
Abstract: Where is SHE in the age of artificial intelligence (A.I.)? In the robot face, in the photo dataset, or in the voice of assistive system? No. She is, we are, in the group of artists leading the exploration of new possibilities of the cutting-edge technologies and embarking on an experimental yet considerable step in the development of media art.
This workshop shares five female media artists’ stories, skills, and experiences of their research and artworks, focusing on how “her stories” are woven into their works and how cutting-edge technologies such as machine learning, evolutionary algorithms, immersive technologies, and interactive technologies unfold these stories. It also offers a place for AWMAT attendees, women artists and researchers, to discuss issues that need to be addressed in the age of A.I.. That may include how their stories can be shaped through advanced technologies, how we can make our work and ourselves more visible at the as-yet-male-dominated frontline, how we actively acknowledge underlying inclusiveness and gender issues such as biased datasets.
Note: laptops optional