Mixed Dimension Embeddings with Application to Memory-Efficient Recommendation Systems
| Autor: | Jiyan Yang, Antonio Ginart, Dheevatsa Mudigere, Maxim Naumov, James Zou |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
FOS: Computer and information sciences
Computer Science - Machine Learning Theoretical computer science Gigabyte Computer science Machine Learning (stat.ML) Recommender system Information theory Power (physics) Machine Learning (cs.LG) Memory management Dimension (vector space) Statistics - Machine Learning Embedding Layer (object-oriented design) |
| Zdroj: | ISIT |
| DOI: | 10.48550/arxiv.1909.11810 |
| Popis: | Embedding representations power machine intelligence in many applications, including recommendation systems, but they are space intensive - potentially occupying hundreds of gigabytes in large-scale 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 x fewer parameters for click-through rate prediction on the Criteo Kaggle dataset. They also train over 2x faster on a GPU. A full version of this paper is accessible at: https://arxiv.org/abs/1909.11810 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|