# Shuffling Recurrent Neural Networks Source: [Shuffling Recurrent Neural Networks]() Collection index: [llms.txt]() Data updated: 2025-10-25T22:25:59.816Z Authors: Michael Rotman, Lior Wolf Published: 2020-07-14T19:36:10.000Z ## Abstract We propose a novel recurrent neural network model, where the hidden state $h\_t$ is obtained by permuting the vector elements of the previous hidden state $h\_{t-1}$ and adding the output of a learned function $b(x\_t)$ of the input $x\_t$ at time $t$. In our model, the prediction is given by a second learned function, which is applied to the hidden state $s(h\_t)$. The method is easy to implement, extremely efficient, and does not suffer from vanishing nor exploding gradients. In an extensive set of experiments, the method shows competitive results, in comparison to the leading literature baselines. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Permuted Image Classification on pMNIST 784](): Permuted Image Classification; pMNIST 784; Accuracy: 96.43. - [Permuted Image Classification on pMNIST 3136](): Permuted Image Classification; pMNIST 3136; Accuracy: 90.31. - [Sequence Copying on Copy Length 120](): Sequence Copying; Copy Length 120; Accuracy: 100. - [Scattered Sequence Copying on Scattered Copy Length 5020](): Scattered Sequence Copying; Scattered Copy Length 5020; Accuracy: 0.303. - [Sequence Copying on Copy Length 5020](): Sequence Copying; Copy Length 5020; Accuracy: 70.8. - [Sequence Addition on Add Length 5000](): Sequence Addition; Add Length 5000; MSE: 0.01206. ## Original sources and implementation [Original paper]() [PDF]()