DeepRecSys, лекция 5: Нейросетевое ранжирование

Lecture: Kirill Khrylchenko Seminar: Artem Matveev This week we begin a two-part discussion on deep ranking: 1. When and why to use neural networks for ranking 2. How to process categorical features, numerical features, and outputs of other models (including embeddings) During the seminar, we discuss two research papers: 1. Unified Embedding: Battle-Tested Feature Representations for Web-Scale ML Systems 2. On Embeddings for Numerical Features in Tabular Deep Learning