Database Paper Browser

Back to papers

ANN Softmax: Acceleration of Extreme Classification Training

Summary: ANN Softmax: GPU-optimized sampling-based softmax for extreme classification with millions of classes, using an inverted-file index and binary quantization to boost recall. GPU kernels plus sample-grouping preserve full-softmax accuracy with 1/10 sampling, achieving 4.3x speedup and enabling 300M-class training on large GPU clusters. (summarized by gpt-5-nano on Feb 09 2026)

Paper ID
12613
Venue
VLDB
Year
2022
Pagerank
4.4626362e-05
Overall Rank
8,712 | 39.40%
DOI
10.14778/3485450.3485451

Incoming Non-self Citations Over Time

Authors

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank Citing Paper Year Venue Pagerank
7,095 Dumpy: A Compact and Adaptive Index for Large Data Series Collections 2023 SIGMOD 4.8350023e-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.

Previous Page 1 / 1 Next

Semantically Similar Papers