Methods › General › Output Functions › Hierarchical Softmax
Hierarchical Softmax
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
Hierarchical Softmax is a is an alternative to softmax that is faster to evaluate: it is O(logn) time to evaluate compared to O(n) for softmax. It utilises a multi-layer binary tree, where the probability of a word is calculated through the product of probabilities on each edge on the path to that node. See the Figure to the right for an example of where the product calculation would occur for the word "I'm".
(Introduced by Morin and Bengio)
Image Credit: Steven Schmatz
Papers archive 2025-07-28
12 shown of 12, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.
-
VQ-Logits: Compressing the Output Bottleneck of Large Language Models via Vector Quantized Logits 15 May 2025 · 0 repositories · arXiv:2505.10202
-
Convergence Rates for Softmax Gating Mixture of Experts 5 Mar 2025 · 0 repositories · arXiv:2503.03213
-
Cross-lingual Embedding Clustering for Hierarchical Softmax in Low-Resource Multilingual Speech Recognition 29 Jan 2025 · 0 repositories · arXiv:2501.17615
-
Global Hierarchical Neural Networks using Hierarchical Softmax 2 Aug 2023 · 1 repository · arXiv:2308.01210
-
Hierarchical Softmax for End-to-End Low-resource Multilingual Speech Recognition 8 Apr 2022 · 1 repository · arXiv:2204.03855
-
Probabilistic Label Trees for Extreme Multi-label Classification 23 Sep 2020 · 2 repositories · arXiv:2009.11218
-
On the computational complexity of the probabilistic label tree algorithms 1 Jun 2019 · 0 repositories · arXiv:1906.00294
-
Effectiveness of Hierarchical Softmax in Large Scale Classification Tasks 13 Dec 2018 · 0 repositories · arXiv:1812.05737
-
A no-regret generalization of hierarchical softmax to extreme multi-label classification 27 Oct 2018 · 1 repository · arXiv:1810.11671
-
Self-organized Hierarchical Softmax 26 Jul 2017 · 0 repositories · arXiv:1707.08588
-
word2vec Parameter Learning Explained 11 Nov 2014 · 8 repositories · arXiv:1411.2738Syntology ran 4 of 9 samples · 5 unverified · 4 pointer-only (licence)
-
Distributed Representations of Words and Phrases and their Compositionality 16 Oct 2013 · 51 repositories · arXiv:1310.4546Syntology ran 6 of 26 samples · 20 unverified · 4 pointer-only (licence)
Tasks archive 2025-07-28
20 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.
Usage over time archive 2025-07-28
Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).
Categories archive 2025-07-28
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections