Michael RoiznerThe Ultimate Question of Diversity, Exploration, and Value in Recommender SystemsReformulating the challenges of ‘beyond accuracy’ aspects through the lens of value optimizationFeb 15Feb 15
Michael RoiznerPersonalization and Popularity BiasIs popularity bias just a convenient excuse?Dec 26, 2023Dec 26, 2023
Michael RoiznerinTowards Data ScienceThe Principled Approach to Early Ranking StagesA systematic method for designing and evaluating candidate generation and early ranking stages in recommender systems.Dec 6, 20231Dec 6, 20231
Michael RoiznerBeyond External Embeddings: Integrating User Histories for Enhanced RecommendationsSome time ago, I wrote about how to use external embeddings. Today, let’s discuss a related (even overlapping) topic: how to use external…Nov 29, 2023Nov 29, 2023
Michael RoiznerinTowards Data ScienceTwo-Tower Networks and Negative Sampling in Recommender SystemsUnderstand the key elements that power advanced recommendation enginesNov 24, 20233Nov 24, 20233
Michael RoiznerRating Systems: The Math Behind StarsMany services feature ratings of objects: products, movies, applications, organizations on a map. How should these be accurately…Nov 16, 2023Nov 16, 2023
Michael RoiznerinTowards Data ScienceFrom Hacks to Harmony: Structuring Product Rules in RecommendationsDon’t let heuristics undermine your ML, learn to combine themSep 23, 2023Sep 23, 2023
Michael RoiznerLeveraging External Embeddings in Recommender Systems: A Practical GuideIn recommender systems, you frequently encounter situations where you have access to ‘external embeddings.’ These are embeddings trained…Sep 10, 20231Sep 10, 20231
Michael RoiznerSLIM: A Fast and Interpretable Baseline for Recommender AlgorithmsContinuing from our post about linear models, today we’ll delve into a specific case — Sparse Linear Methods (SLIM). Here’s what sets this…Aug 31, 2023Aug 31, 2023
Michael RoiznerBeyond Counters: Linear Models in RecommendationsLinear models are nearly the simplest tools in machine learning, but they shouldn’t be underestimated. They are remarkably adaptable, and…Aug 22, 2023Aug 22, 2023