DBScholar

Back to papers

ANN Softmax: Acceleration of Extreme Classification Training

Summary: ANN Softmax combines binary-quantized inverted-file retrieval, GPU kernels, and sample grouping to select high-recall classes while avoiding full extreme-classification softmax. It matches full-softmax accuracy with 1/10 classes, delivers 4.3× faster training, and scales to 300M classes. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID
12801
Venue
VLDB
Year
2022
Pagerank
5.3573227e-05
Overall Rank
8,855 | 39.25%
DOI
10.14778/3485450.3485451

Incoming Non-self Citations Over Time

Authors

BibTeX Citation

@article{zhao_vldb22,
        title = {{ANN Softmax: Acceleration of Extreme Classification Training}},
        author = {Zhao, Kang and Song, Liuyihan and Zhang, Yingya and Pan, Pan and Xu, Yinghui and Jin, Rong},
        journal = {PVLDB},
        series = {{VLDB} '22},
        volume = {15},
        number = {1},
        pages = {1--10},
        doi = {10.14778/3485450.3485451},
        url = {https://doi.org/10.14778/3485450.3485451},
        year = {2022}
}

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank Citing Paper Year Venue Pagerank
6,691 Dumpy: A Compact and Adaptive Index for Large Data Series Collections 2023 SIGMOD 5.8011086e-05
Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 13 of 13 cited papers.

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

Rank Cited Paper Year Venue Pagerank
21 Similarity Search in High Dimensions via Hashing 1999 VLDB 0.00056760516
93 Fast Approximate Nearest Neighbor Search With The Navigating Spreading-out Graph 2019 VLDB 0.00034701237
287 Multi-Probe LSH: Efficient Indexing for High-Dimensional Similarity Search 2007 VLDB 0.00022323585
397 TURL: Table Understanding through Representation Learning 2021 VLDB 0.00019278189
521 PyTorch Distributed: Experiences on Accelerating Data Parallel Training 2020 VLDB 0.0001713368
705 HD-Index: Pushing the Scalability-Accuracy Boundary for Approximate kNN Search in High-Dimensional Spaces 2018 VLDB 0.00014829964
805 Cache locality is not enough: High-Performance Nearest Neighbor Search with Product Quantization Fast Scan 2016 VLDB 0.00013891999
2,219 Chimera: Large-Scale Classification using Machine Learning, Rules, and Crowdsourcing 2014 VLDB 8.9303727e-05
2,673 DSH: Data Sensitive Hashing for High-Dimensional k-NN Search 2014 SIGMOD 8.2730112e-05
2,702 Panorama: A Data System for Unbounded Vocabulary Querying over Video 2020 VLDB 8.2342712e-05
3,782 CROSSBOW: Scaling Deep Learning with Small Batch Sizes on Multi-GPU Servers 2019 VLDB 7.1298979e-05
4,924 ODIN: Automated Drift Detection and Recovery in Video Analytics 2020 VLDB 6.4429959e-05
8,869 A Distributed System for Large-scale n-gram Language Models at Tencent 2019 VLDB 5.3543688e-05
Previous Page 1 / 1 Next

Semantically Similar Papers

# Overall Rank Paper Year Venue
1 2,688 Accelerating Recommendation System Training by Leveraging Popular Choices 2022 VLDB
2 9,729 Scalable Graph Convolutional Network Training on Distributed-Memory Systems 2023 VLDB
3 8,883 FlexMoE: Scaling Large-scale Sparse Pre-trained Model Training via Dynamic Device Placement 2023 SIGMOD
4 8,313 Fast Neural Ranking on Bipartite Graph Indices 2022 VLDB
5 10,800 Two Birds with One Stone: Efficient Deep Learning over Mislabeled Data through Subset Selection 2025 SIGMOD
6 7,374 ADGNN: Towards Scalable GNN Training with Aggregation-Difference Aware Sampling 2023 SIGMOD
7 10,336 Accelerating High-Dimensional ANN Search via Skipping Redundant Distance Computations 2026 SIGMOD
8 3,253 Jointly Optimizing Preprocessing and Inference for DNN-based Visual Analytics 2021 VLDB
9 5,813 Efficient Algorithms for Crowd-Aided Categorization 2020 VLDB
10 10,456 FlashANNS: GPU-Driven Asynchronous I/O Pipelining for Eliminating Storage-Compute Bottlenecks in Billion-Scale Similarity Search 2026 SIGMOD