A Distributed System for Large-scale n-gram Language Models at Tencent

Summary: Tencent's distributed system for large n-gram LMs; indexing, batching, caching reduce network traffic and speed nodes. Cascade fault-tolerance switches to smaller n-grams on failure; shown on 9 ASR datasets and deployed to WeChat at 100M messages/min. (summarized by gpt-5-nano on Feb 09 2026)

Paper ID: 12127
Venue: VLDB
Year: 2019
Pagerank: 5.3543688e-05
Overall Rank: 8,869 | 39.16%
DOI: 10.14778/3352063.3352136

Incoming Non-self Citations Over Time

Authors

1. Qiang Long (Tencent)
2. Wei Wang (National University of Singapore)
3. Jinfu Deng (Tencent)
4. Song Liu (Tencent)
5. Wenhao Huang (Tencent)
6. Fangying Chen (Tencent)
7. Sifan Liu (Tencent)

BibTeX Citation

@article{long_vldb19,
        title = {{A Distributed System for Large-scale n-gram Language Models at Tencent}},
        author = {Long, Qiang and Wang, Wei and Deng, Jinfu and Liu, Song and Huang, Wenhao and Chen, Fangying and Liu, Sifan},
        journal = {PVLDB},
        series = {{VLDB} '19},
        volume = {12},
        number = {12},
        pages = {2206--2217},
        doi = {10.14778/3352063.3352136},
        url = {https://doi.org/10.14778/3352063.3352136},
        year = {2019}
}

Incoming Citations (Sorted by Pagerank)

Showing 2 of 2 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
8,855	ANN Softmax: Acceleration of Extreme Classification Training	2022	VLDB	5.3573227e-05
11,196	A Universal Sketch for Estimating Heavy Hitters and Per-Element Frequency Moments in Data Streams with Bounded Deletions	2024	SIGMOD	5.093636e-05

Outgoing Citations (Sorted by Pagerank)

Showing 5 of 5 cited papers.

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

Rank	Cited Paper	Year	Venue	Pagerank
284	NoScope: Optimizing Neural Network Queries over Video at Scale	2017	VLDB	0.00022370521
295	Accelerating Machine Learning Inference with Probabilistic Predicates	2018	SIGMOD	0.00022238183
1,792	MacroBase: Prioritizing Attention in Fast Data	2017	SIGMOD	9.7436856e-05
4,083	SketchML: Accelerating Distributed Machine Learning with Data Sketches	2018	SIGMOD	6.9160949e-05
7,218	Fast Failure Recovery in Distributed Graph Processing Systems	2015	VLDB	5.6683391e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	20	Distributed GraphLab: A Framework for Machine Learning and Data Mining in the Cloud	2012	VLDB
2	2,640	Scalable and Efficient Full-Graph GNN Training for Large Graphs	2023	SIGMOD
3	8,454	D3-GNN: Dynamic Distributed Dataflow for Streaming Graph Neural Networks	2024	VLDB
4	14,018	Communication-Efficient Distributed Mining of Association Rules	2001	SIGMOD
5	4,912	HET-GMP: A Graph-based System Approach to Scaling Large Embedding Model Training	2022	SIGMOD
6	2,162	Heterogeneity-aware Distributed Parameter Servers	2017	SIGMOD
7	8,034	Angel-PTM: A Scalable and Economical Large-scale Pre-training System in Tencent	2023	VLDB
8	3,109	Managing Large Dynamic Graphs Efficiently	2012	SIGMOD
9	7,900	Distributed Graph Embedding with Information-Oriented Random Walks	2023	VLDB
10	1,875	Large-Scale Distributed Graph Computing Systems: An Experimental Evaluation	2015	VLDB