Methods › Sequential › Recurrent Neural Networks › SRU
SRU
Introduced by Tao Lei et al. in Simple Recurrent Units for Highly Parallelizable Recurrence
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
SRU, or Simple Recurrent Unit, is a recurrent neural unit with a light form of recurrence. SRU exhibits the same level of parallelism as convolution and feed-forward nets. This is achieved by balancing sequential dependence and independence: while the state computation of SRU is time-dependent, each state dimension is independent. This simplification enables CUDA-level optimizations that parallelize the computation across hidden dimensions and time steps, effectively using the full capacity of modern GPUs.
SRU also replaces the use of convolutions (i.e., ngram filters), as in QRNN and KNN, with more recurrent connections. This retains modeling capacity, while using less computation (and hyper-parameters). Additionally, SRU improves the training of deep recurrent models by employing highway connections and a parameter initialization scheme tailored for gradient propagation in deep architectures.
A single layer of SRU involves the following computation:
𝐟ₜ =σ(𝐖_f 𝐱ₜ+𝐯_f ⊙𝐜ₜ₋₁+𝐛_f)
𝐜ₜ =𝐟ₜ ⊙𝐜ₜ₋₁+(1-𝐟ₜ) ⊙(𝐖 𝐱ₜ)
𝐫ₜ =σ(𝐖ᵣ 𝐱ₜ+𝐯ᵣ ⊙𝐜ₜ₋₁+𝐛ᵣ)
𝐡ₜ =𝐫ₜ ⊙𝐜ₜ+(1-𝐫ₜ) ⊙𝐱ₜ
where 𝐖, 𝐖_f and 𝐖ᵣ are parameter matrices and 𝐯_f, 𝐯ᵣ, 𝐛_f and 𝐛ᵥ are parameter vectors to be learnt during training. The complete architecture decomposes to two sub-components: a light recurrence and a highway network,
The light recurrence component successively reads the input vectors 𝐱ₜ and computes the sequence of states 𝐜ₜ capturing sequential information. The computation resembles other recurrent networks such as LSTM, GRU and RAN. Specifically, a forget gate 𝐟ₜ controls the information flow and the state vector 𝐜ₜ is determined by adaptively averaging the previous state 𝐜ₜ₋₁ and the current observation 𝐖 𝐱₊according to 𝐟ₜ.
Papers archive 2025-07-28
16 shown of 16, 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.
-
Selective-Stereo: Adaptive Frequency Information Selection for Stereo Matching 1 Mar 2024 · 1 repository · arXiv:2403.00486Syntology ran 9 of 13 samples · 4 unverified
-
On the Effectiveness of Unlearning in Session-Based Recommendation 22 Dec 2023 · 1 repository · arXiv:2312.14447Syntology ran 1 of 1 samples · 0 unverified · 1 pointer-only (licence)
-
Neural Machine Translation Models with Attention-Based Dropout Layer 1 May 2023 · 1 repository
-
SCConv: Spatial and Channel Reconstruction Convolution for Feature Redundancy 1 Jan 2023 · 1 repository
-
A Robust Approach for the Decomposition of High-Energy-Consuming Industrial Loads with Deep Learning 11 Mar 2022 · 0 repositories · arXiv:2203.07075
-
MOHAQ: Multi-Objective Hardware-Aware Quantization of Recurrent Neural Networks 2 Aug 2021 · 0 repositories · arXiv:2108.01192
-
Intelligent Reflecting Surface Enhanced Indoor Robot Path Planning: A Radio Map based Approach 27 Sep 2020 · 0 repositories · arXiv:2009.12804
-
ASAPP-ASR: Multistream CNN and Self-Attentive SRU for SOTA Speech Recognition 21 May 2020 · 0 repositories · arXiv:2005.10469
-
Multistream CNN for Robust Acoustic Modeling 21 May 2020 · 0 repositories · arXiv:2005.10470
-
Utterance-level Sequential Modeling For Deep Gaussian Process Based Speech Synthesis Using Simple Recurrent Unit 22 Apr 2020 · 0 repositories · arXiv:2004.10823
-
WaveCRN: An Efficient Convolutional Recurrent Neural Network for End-to-end Speech Enhancement 6 Apr 2020 · 5 repositories · arXiv:2004.04098
-
Economy Statistical Recurrent Units For Inferring Nonlinear Granger Causality 22 Nov 2019 · 1 repository · arXiv:1911.09879Syntology ran 1 of 1 samples · 0 unverified
-
FastFusionNet: New State-of-the-Art for DAWNBench SQuAD 28 Feb 2019 · 2 repositories · arXiv:1902.11291
-
Single Stream Parallelization of Recurrent Neural Networks for Low Power and Fast Inference 30 Mar 2018 · 0 repositories · arXiv:1803.11389
-
Training RNNs as Fast as CNNs 1 Jan 2018 · 2 repositories
-
Simple Recurrent Units for Highly Parallelizable Recurrence 8 Sep 2017 · 11 repositories · arXiv:1709.02755Syntology ran 1 of 1 samples · 0 unverified
Tasks archive 2025-07-28
20 shown of 27 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