| Paper ID | D6-S5-T3.3 |
| Paper Title |
Mixed Dimension Embeddings with Application to Memory-Efficient Recommendation Systems |
| Authors |
Antonio Ginart, Stanford University, United States; Maxim Naumov, Dheevatsa Mudigere, Jiyan Yang, Facebook, United States; James Zou, Stanford University, United States |
| Session |
D6-S5-T3: Dimension Reduction |
| Chaired Session: |
Monday, 19 July, 23:20 - 23:40 |
| Engagement Session: |
Monday, 19 July, 23:40 - 00:00 |
| Abstract |
Embedding representations power machine intelligence in many applications, including recommendation systems, but they are space intensive — potentially occupying hundreds of gigabytes in largescale settings. To help manage this outsized memory consumption, we explore mixed dimension embeddings, an embedding layer architecture in which a particular embedding vector’s dimension scales with its query frequency. Through theoretical analysis and systematic experiments, we demonstrate that using mixed dimensions can drastically reduce the memory usage, while maintaining and even improving the ML performance. Empirically, we show that the proposed mixed dimension layers improve accuracy by 0.1% using half as many parameters or maintain it using 16× fewer parameters for click-through rate prediction task on the Criteo Kaggle dataset.
|
