Recommender Systems - Beyond Machine Learning with Joe Konstan

Explore the intricacies of recommender systems beyond traditional machine learning approaches in this insightful ACM conference talk. Delve into the successes and failures of combining human-centered evaluation with data mining techniques to improve user experience. Learn about sophisticated technologies for modeling user preferences, item properties, and leveraging community experiences. Discover the challenges of improving recommendations beyond accuracy and precision metrics. Gain valuable insights from Joseph A. Konstan, a distinguished professor and ACM Software System Award recipient, as he discusses personalization, collaborative filtering, and the importance of human factors in recommender systems. Examine topics such as eliciting online participation, designing systems for public health, and the evolution of recommender system metrics. Understand the balance between marketing goals and user needs, and explore innovative concepts like novelty, personality-based recommendations, and giving users more control over their recommendations.

Welcome
Housekeeping
Presentation
Personalization
Types of recommendations
User collaborative filtering
Latent factor models
Why I love computing
What is useful
Metrics history
Challenges
Marketing
A horrible reality
Giving people control
Novelty
Personality
Top Hat Lists
Purple Rain
Oliver
Cycling
Second Best
Explorer
Recommender
Machine Learning
Message
Questions Answers
Collaborative Filtering