Training Machines to Humanize Dating

Synthetic Data: How Do We Manage Them? | Armando Vieira, Hazy

Synthetic Data: How Do We Manage Them? | Armando Vieira, Hazy

Users Floating in Space: A Study in Recommendations | Tyler Neylon, Unbox Research

Users Floating in Space: A Study in Recommendations | Tyler Neylon, Unbox Research

Are You Vulnerable to Adversarial Machine Learning? | Celeste Fralick, McAfee

Are You Vulnerable to Adversarial Machine Learning? | Celeste Fralick, McAfee

Convenient and Flexible ML Pipelines with Kubeflow | Mattias Arro, Subspace AI

Convenient and Flexible ML Pipelines with Kubeflow | Mattias Arro, Subspace AI

ABOUT

While machine learning recommender algorithms traditionally solve for the personalized preference problem, in online dating we must also consider the user base’s collective fulfillment. More generally, balancing individualism and collectivism is central to sustaining a product ecosystem as well as to using machine learning responsibly. In this video, Shanshan Ding discusses ways this tension manifests on Hinge, some of the existing solutions to address this tension, and the open problems that they continue to explore.

Shanshan Ding
Director of Data Science, Hinge

Shanshan Ding is the Director of Data Science at Hinge, where she leads the development of online dating recommender algorithms and anti-fraud measures. Shanshan holds a Ph.D. in probability theory from the University of Pennsylvania, and she relishes leveraging her professional expertise in mechanisms of chance to deliver experiences of serendipitous encounters to Hinge members. Being a long-time math educator prior to entering tech, Shanshan is passionate about democratizing data science and empowering everyone to make data-driven decisions.