EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A SYSTEMS-AWARE FRAMEWORK FOR INTELLIGENT LARGE-SCALE RECOMMENDATION MODELS

AoWei Shen, Shan Huang

Journal of Computer Science and Electrical Engineering · 03.12.2022 · 4 (1) · 36-44

DOI: 10.61784/jcsee4153 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Computer Science and Electrical Engineering, online erschienen am 03.12.2022. Verfasst von AoWei Shen und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Embedding tables dominate the memory footprint of deep learning recommendation models, and their size already exceeds the capacity of the fast memory attached to accelerators. Existing compression methods reduce parameter counts without regard to where the parameters reside, while placement and caching systems move fixed tables between memory levels without changing their representation. This paper presents TierRec, a systems-aware framework that treats representation and placement as one decision. TierRec profiles the access skew of each categorical field together with an empirical fidelity profile of candidate representations, partitions the ID space into a full-dimensional hot tier, low-dimensional warm tiers with shared projections, and a compositional cold tier, and chooses the tier boundaries and dimensions by maximizing an access-weighted fidelity objective under a memory budget and, optionally, a slow-memory traffic constraint derived from a two-level memory model. Placement then follows access density so that compact rows occupy fast memory first. On MovieLens-20M with a DLRM-style ranking model, TierRec reaches a test AUC of 0.8199 at 8× compression and 0.8153 at 16× compression, matching or exceeding mixed-dimension embeddings and outperforming hashing, quotient-remainder and frequency-based double hashing by 0.0176 to 0.0560 AUC. The hierarchy-aware variant meets a slow-traffic bound of half the memory-only optimum at a cost of 0.0008 AUC. Trace-driven analyses on real and Criteo-scale synthetic workloads quantify the resulting gains in fast-memory hit ratio and access time.

Zitieren

APA

Shen, A., & Huang, S. (2022). A SYSTEMS-AWARE FRAMEWORK FOR INTELLIGENT LARGE-SCALE RECOMMENDATION MODELS. Journal of Computer Science and Electrical Engineering, 4(1), 36-44. https://doi.org/10.61784/jcsee4153

Vancouver

Shen A, Huang S. A SYSTEMS-AWARE FRAMEWORK FOR INTELLIGENT LARGE-SCALE RECOMMENDATION MODELS. Journal of Computer Science and Electrical Engineering. 2022;4(1):36-44. doi:10.61784/jcsee4153

BibTeX
@article{Shen2022,
  author = {Shen, AoWei and Huang, Shan},
  title = {A SYSTEMS-AWARE FRAMEWORK FOR INTELLIGENT LARGE-SCALE RECOMMENDATION MODELS},
  journal = {Journal of Computer Science and Electrical Engineering},
  year = {2022},
  volume = {4},
  number = {1},
  pages = {36--44},
  doi = {10.61784/jcsee4153}
}
Kostenlos starten →