Welcome to RS_c, the central platform for the RecSys community. We provide curated lists of recommender-systems datasets, algorithms, books, conferences and many resources more. Maybe most importantly, we publish the latest recommender-system news. If you want your news to be reported on RS_c, read here.
The Digital Services Act (DSA) introduces rules about how online platforms should manage their recommender systems. These systems suggest content to users. The DSA makes platforms called VLOPs/VLOSEs check and lessen their “systemic risks,” including risks from their recommender systems (Articles 34 and 35). Article 27, which applies to all online platforms, requires these platforms to clearly explain to users how their recommender systems work. Additionally, Article 38 says these platforms must offer a version of their recommender system that doesn’t use “profiling,” which involves collecting and analyzing personal data under the GDPR.
Other parts of the DSA also affect recommender systems. They deal with independent checks (Article 37), researchers getting data (Article 40), rules about platforms’ terms and conditions (Article 14), and explanations for decisions (Article 17). These include actions platforms take to lower the visibility of certain content.
However, there are questions about how effective these DSA rules are. How much detail will be added later by laws and advice from regulators? Do platforms still have too much freedom?
A recent study from Mozilla has suggested several actions to create a more accountable system for making recommendations online. These actions could help us judge the DSA’s effectiveness. They include making things clear to users, the public, and experts; letting third parties check recommender systems; giving researchers more access to data, documents, and tools; and giving users more control over their online experience and the data collected from them. Although each action alone won’t solve all the issues with platforms and their recommenders, together they could lead to better oversight and more research on these systems.
Now, the question is how does the DSA measure up to these suggestions? Will it give researchers enough power to really understand recommender systems? Will it make these systems transparent enough for real examination? And will it let users have real control?
I am the founder of Recommender-Systems.com and head of the Intelligent Systems Group (ISG) at the University of Siegen, Germany https://isg.beel.org. We conduct research in recommender-systems (RecSys), personalization and information retrieval (IR) as well as on automated machine learning (AutoML), meta-learning and algorithm selection. Domains we are particularly interested in include smart places, eHealth, manufacturing (industry 4.0), mobility, visual computing, and digital libraries.
We founded or maintain, among others, LensKit-Auto, Darwin & Goliath, Mr. DLib, and Docear, each with thousand of users; we contributed to TensorFlow, JabRef and others; and we developed the first prototypes of automated recommender systems (AutoSurprise and Auto-CaseRec) and Federated Meta Learning (FMLearn Server and Client).