Mirjam Augstein

Gustavo Escobedo, Christian Ganhör, Stefan Brandl et al.

In widely used neural network-based collaborative filtering models, users' history logs are encoded into latent embeddings that represent the users' preferences. In this setting, the models are capable of mapping users' protected attributes (e.g., gender or ethnicity) from these user embeddings even without explicit access to them, resulting in models that may treat specific demographic user groups unfairly and raise privacy issues. While prior work has approached the removal of a single protected attribute of a user at a time, multiple attributes might come into play in real-world scenarios. In the work at hand, we present AdvXMultVAE which aims to unlearn multiple protected attributes (exemplified by gender and age) simultaneously to improve fairness across demographic user groups. For this purpose, we couple a variational autoencoder (VAE) architecture with adversarial training (AdvMultVAE) to support simultaneous removal of the users' protected attributes with continuous and/or categorical values. Our experiments on two datasets, LFM-2b-100k and Ml-1m, from the music and movie domains, respectively, show that our approach can yield better results than its singular removal counterparts (based on AdvMultVAE) in effectively mitigating demographic biases whilst improving the anonymity of latent embeddings.

Thomas Neumayr, Banu Saatci, Sean Rintel et al.

Interest in hybrid collaboration and meetings (HCM), where several co-located participants engage in coordinated work with remote participants, is gaining unprecedented momentum after the rapid shift in working from home due to the COVID-19 pandemic. However, while the interest is new, researchers have been exploring HCM phenomena for decades, albeit dispersed across diverse research traditions, using different terms, definitions, and frameworks. In this article, we present a systematic literature review of the contexts and tools of HCM in the ACM Digital Library. We obtained approximately 1,200 results, which were narrowed down to 62 key articles. We report on the terms, citations, venues, authors, domains, study types, and data of these publications and present a taxonomic overview based on their reported hybrid settings' actual characteristics. We discuss why the SLR resulted in a relatively small number of publications, and then as a corollary, discuss how some excluded high-profile publications flesh out the SLR findings to provide important additional concepts. The SLR itself covers the ACM until November 2019, so our discussion also includes relevant 2020 and 2021 publications. The end result is a baseline that researchers and designers can use in shaping the post-COVID-19 future of HCM systems.