PDN Pro-Social: Ranking by User-Value w/ Smitha Milli

​It is now a truism that ranking content on social media by engagement algorithms alone leads to no good (e.g. flame wars and social divisiveness). But what are the alternatives? One approach, recommended by the Neely code among others, is to rank content by what users indicate to be valuable content.

Smitha Milli is a postdoc at Cornell Tech who has thought extensively about how to build such algorithms. Most recently, they have co-authored papers that review approaches to using non-engagement signals in content ranking and that propose how to strike the right balance of value, strategy, and noise in recommender systems.

We hope you can join us for a conversation with Smitha on how to integrate user value indicators into platform algorithms.