Accelerating Recommendation System Training by Leveraging Popular Choices

Summary: FAE exploits extreme popularity skew in embedding accesses, placing hot entries in scarce GPU memory and accelerating their updates. A hot-embedding-aware layout cuts CPU–GPU transfers, delivering 2.3× speedup over CPU-only and 1.52× over hybrid training without accuracy loss. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID: 12824
Venue: VLDB
Year: 2022
Pagerank: 8.2564305e-05
Overall Rank: 2,688 | 81.56%
DOI: 10.14778/3485450.3485462

Incoming Non-self Citations Over Time

Authors

1. Muhammad Adnan (University of British Columbia)
2. Yassaman Ebrahimzadeh Maboud (University of British Columbia)
3. Divya Mahajan (Microsoft)
4. Prashant J. Nair (University of British Columbia)

BibTeX Citation

@article{adnan_vldb22,
        title = {{Accelerating Recommendation System Training by Leveraging Popular Choices}},
        author = {Adnan, Muhammad and Maboud, Yassaman Ebrahimzadeh and Mahajan, Divya and Nair, Prashant J.},
        journal = {PVLDB},
        series = {{VLDB} '22},
        volume = {15},
        number = {1},
        pages = {127--140},
        doi = {10.14778/3485450.3485462},
        url = {https://doi.org/10.14778/3485450.3485462},
        year = {2022}
}

Incoming Citations (Sorted by Pagerank)

Showing 10 of 10 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
6,954	DLRover-RM: Resource Optimization for Deep Recommendation Models Training in the Cloud	2024	VLDB	5.7303405e-05
6,958	Efficient Fault Tolerance for Recommendation Model Training via Erasure Coding	2023	VLDB	5.7303405e-05
8,907	Scheduling Data Processing Pipelines for Incremental Training on MLP-based Recommendation Models	2025	SIGMOD	5.3483178e-05
8,949	Modyn: Data-Centric Machine Learning Pipeline Orchestration	2025	SIGMOD	5.3462965e-05
9,164	FEC: Efficient Deep Recommendation Model Training with Flexible Embedding Communication	2023	SIGMOD	5.3101785e-05
9,475	BladeDISC: Optimizing Dynamic Shape Machine Learning Workloads via Compiler Approach	2023	SIGMOD	5.2634238e-05
9,881	The Image Calculator: 10x Faster Image-AI Inference by Replacing JPEG with Self-designing Storage Format	2024	SIGMOD	5.2040783e-05
10,752	CARINA: An Efficient CXL-Oriented Embedding Serving System for Recommendation Models	2025	SIGMOD	5.093636e-05
10,844	GPEmu: A GPU Emulator for Faster and Cheaper Prototyping and Evaluation of Deep Learning System Research	2025	VLDB	5.093636e-05
10,943	Lighter-X: An Efficient and Plug-and-play Strategy for Graph-based Recommendation through Decoupled Propagation	2025	VLDB	5.093636e-05

Outgoing Citations (Sorted by Pagerank)

Showing 4 of 4 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
803	MRShare: Sharing Across Multiple Queries in MapReduce	2010	VLDB	0.00013899943
1,446	Analyzing and Mitigating Data Stalls in DNN Training	2021	VLDB	0.0001076818
1,666	HippogriffDB: Balancing I/O and GPU Bandwidth in Big Data Analytics	2016	VLDB	0.00010068964
3,682	In-RDBMS Hardware Acceleration of Advanced Analytics	2018	VLDB	7.2035518e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	11,108	Efficient Graph Embedding Generation and Update for Large-Scale Temporal Graph	2025	VLDB
2	13,321	The Limits of Graph Samplers for Training Inductive Recommender Systems	2025	VLDB
3	9,520	Experimental Analysis of Large-scale Learnable Vector Storage Compression	2024	VLDB
4	6,958	Efficient Fault Tolerance for Recommendation Model Training via Erasure Coding	2023	VLDB
5	6,954	DLRover-RM: Resource Optimization for Deep Recommendation Models Training in the Cloud	2024	VLDB
6	8,907	Scheduling Data Processing Pipelines for Incremental Training on MLP-based Recommendation Models	2025	SIGMOD
7	1,234	Large Graph Convolutional Network Training with GPU-Oriented Data Communication Architecture	2021	VLDB
8	5,433	Supercharging Recommender Systems using Taxonomies for Learning User Purchase Behavior	2012	VLDB
9	9,556	CAFE: Towards Compact, Adaptive, and Fast Embedding for Large-scale Recommendation Models	2024	SIGMOD
10	9,164	FEC: Efficient Deep Recommendation Model Training with Flexible Embedding Communication	2023	SIGMOD