Methods › Natural Language Processing › Autoregressive Transformers › Universal Transformer
Universal Transformer
Introduced by Mostafa Dehghani et al. in Universal Transformers
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
The Universal Transformer is a generalization of the Transformer architecture. Universal Transformers combine the parallelizability and global receptive field of feed-forward sequence models like the Transformer with the recurrent inductive bias of RNNs. They also utilise a dynamic per-position halting mechanism.
Papers archive 2025-07-28
18 shown of 18, 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.
-
PLUTO: Pathology-Universal Transformer 13 May 2024 · 0 repositories · arXiv:2405.07905
-
Investigating Recurrent Transformers with Dynamic Halt 1 Feb 2024 · 1 repository · arXiv:2402.00976
-
Self-Critical Alternate Learning based Semantic Broadcast Communication 3 Dec 2023 · 0 repositories · arXiv:2312.01423
-
Sparse Universal Transformer 11 Oct 2023 · 2 repositories · arXiv:2310.07096Syntology ran 3 of 5 samples · 2 unverified
-
UMMAFormer: A Universal Multimodal-adaptive Transformer Framework for Temporal Forgery Localization 28 Aug 2023 · 1 repository · arXiv:2308.14395
-
Beyond Universal Transformer: block reusing with adaptor in Transformer for automatic speech recognition 23 Mar 2023 · 0 repositories · arXiv:2303.13072
-
Semantic Communication with Memory 22 Mar 2023 · 0 repositories · arXiv:2303.12335
-
Towards Autoformalization of Mathematics and Code Correctness: Experiments with Elementary Proofs 5 Jan 2023 · 1 repository · arXiv:2301.02195Syntology ran 0 of 7 samples · 7 unverified
-
Universal Transformer Hawkes Process with Adaptive Recursive Iteration 29 Dec 2021 · 0 repositories · arXiv:2112.14479
-
The Devil is in the Detail: Simple Tricks Improve Systematic Generalization of Transformers 26 Aug 2021 · 2 repositories · arXiv:2108.12284Syntology ran 0 of 2 samples · 2 unverified
-
Using BERT Encoding and Sentence-Level Language Model for Sentence Ordering 24 Aug 2021 · 0 repositories · arXiv:2108.10986
-
Semantic Communication with Adaptive Universal Transformer 20 Aug 2021 · 0 repositories · arXiv:2108.09119
-
Automatically Ranked Russian Paraphrase Corpus for Text Generation 17 Jun 2020 · 0 repositories · arXiv:2006.09719
-
Universal Transforming Geometric Network 2 Aug 2019 · 1 repository · arXiv:1908.00723
-
Latent Universal Task-Specific BERT 16 May 2019 · 0 repositories · arXiv:1905.06638
-
Self-Attentive Model for Headline Generation 23 Jan 2019 · 0 repositories · arXiv:1901.07786
-
Attending to Mathematical Language with Transformers 5 Dec 2018 · 3 repositories · arXiv:1812.02825
-
Universal Transformers 10 Jul 2018 · 8 repositories · arXiv:1807.03819Syntology ran 14 of 25 samples · 11 unverified · 24 pointer-only (licence)
Tasks archive 2025-07-28
20 shown of 43 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