% Key papers for RNN training dynamics and gradients (scime atlas, https://atlas.scime.ai). Authors are not included: add them from the paper pages.

@inproceedings{aaai2016_character_105208,
  title = {{Character-Aware Neural Language Models}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2016},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/10362},
}

@inproceedings{cvpr2017_full_12292,
  title = {{Full Resolution Image Compression With Recurrent Neural Networks}},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2017},
}

@inproceedings{iclr2017_quasi_28254,
  title = {{Quasi-Recurrent Neural Networks}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2017},
  url = {https://openreview.net/forum?id=H1zJ-v5xl},
}

@inproceedings{iclr2017_topicrnn_28365,
  title = {{TopicRNN: A Recurrent Neural Network with Long-Range Semantic Dependency}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2017},
  url = {https://openreview.net/forum?id=rJbbOLcex},
}

@inproceedings{aaai2018_a_106027,
  title = {{A Multilayer Convolutional Encoder-Decoder Neural Network for Grammatical Error Correction}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2018},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/12069},
}

@inproceedings{iclr2017_zoneout_28369,
  title = {{Zoneout: Regularizing RNNs by Randomly Preserving Hidden Activations}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2017},
  url = {https://openreview.net/forum?id=rJqBEPcxe},
}

@inproceedings{aaai2018_learning_105915,
  title = {{Learning to Compose Task-Specific Tree Structures}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2018},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/11975},
}

@inproceedings{icml2018_extracting_36680,
  title = {{Extracting Automata from Recurrent Neural Networks Using Queries and Counterexamples}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2018},
  url = {https://icml.cc/virtual/2018/poster/2276},
}

@inproceedings{neurips2017_dilated_56257,
  title = {{Dilated Recurrent Neural Networks}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2017},
  url = {https://nips.cc/virtual/2017/poster/8806},
}

@inproceedings{iclr2019_antisymmetricrnn_29228,
  title = {{AntisymmetricRNN: A Dynamical System View on Recurrent Neural Networks}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2019},
  url = {https://iclr.cc/virtual/2019/poster/696},
}

@inproceedings{neurips2016_architectural_55582,
  title = {{Architectural Complexity Measures of Recurrent Neural Networks}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2016},
  url = {https://nips.cc/virtual/2016/poster/7303},
}

@inproceedings{neurips2018_fastgrnn_57133,
  title = {{FastGRNN: A Fast, Accurate, Stable and Tiny Kilobyte Sized Gated Recurrent Neural Network}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2018},
  url = {https://nips.cc/virtual/2018/poster/11860},
}

@inproceedings{iclr2019_trellis_28969,
  title = {{Trellis Networks for Sequence Modeling}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2019},
  url = {https://iclr.cc/virtual/2019/poster/825},
}

@inproceedings{iclr2019_stable_28979,
  title = {{Stable Recurrent Models}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2019},
  url = {https://iclr.cc/virtual/2019/poster/658},
}

@inproceedings{icml2017_tunable_36435,
  title = {{Tunable Efficient Unitary Neural Networks (EUNN) and their application to RNNs}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2017},
  url = {https://icml.cc/virtual/2017/poster/813},
}

@inproceedings{icml2018_learning_36793,
  title = {{Learning Longer-term Dependencies in RNNs with Auxiliary Losses}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2018},
  url = {https://icml.cc/virtual/2018/poster/2127},
}

@inproceedings{aaai2020_particle_102090,
  title = {{Particle Filter Recurrent Neural Networks}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2020},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/5952},
}

@inproceedings{iclr2017_query_28203,
  title = {{Query-Reduction Networks for Question Answering}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2017},
  url = {https://openreview.net/forum?id=B1MRcPclx},
}

@inproceedings{iclr2017_capacity_28237,
  title = {{Capacity and Trainability in Recurrent Neural Networks}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2017},
  url = {https://openreview.net/forum?id=BydARw9ex},
}

@inproceedings{icml2018_dynamical_36659,
  title = {{Dynamical Isometry and a Mean Field Theory of RNNs: Gating Enables Signal Propagation in Recurrent Neural Networks}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2018},
  url = {https://icml.cc/virtual/2018/poster/2464},
}

@inproceedings{neurips2024_recurrent_70408,
  title = {{Recurrent neural networks: vanishing and exploding gradients are not the end of the story}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2024},
  url = {https://neurips.cc/virtual/2024/poster/96697},
}

@inproceedings{aaai2024_exploiting_5909,
  title = {{Exploiting Symmetric Temporally Sparse BPTT for Efficient RNN Training}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2024},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/29020},
}

@inproceedings{iclr2024_inverse_35649,
  title = {{Inverse Approximation Theory for Nonlinear Recurrent Neural Networks}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2024},
  url = {https://iclr.cc/virtual/2024/poster/17424},
}

@inproceedings{icml2024_universality_43417,
  title = {{Universality of Linear Recurrences Followed by Non-linear Projections: Finite-Width Guarantees and Benefits of Complex Eigenvalues}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2024},
  url = {https://icml.cc/virtual/2024/poster/35035},
}

@inproceedings{neurips2025_revisiting_78157,
  title = {{Revisiting Glorot Initialization for Long-Range Linear Recurrences}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=bd8kppxyB3},
}

@inproceedings{neurips2025_improving_78925,
  title = {{Improving Bilinear RNN with Closed-loop Control}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=jlJaRXDzCE},
}

@inproceedings{iclr2025_unlocking_82349,
  title = {{Unlocking State-Tracking in Linear RNNs Through Negative Eigenvalues}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2025},
  url = {https://iclr.cc/virtual/2025/poster/29442},
}

@inproceedings{aaai2025_do_8787,
  title = {{Do Transformer Interpretability Methods Transfer to RNNs?}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2025},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/34969},
}

@inproceedings{icml2024_a_43892,
  title = {{A Tensor Decomposition Perspective on Second-order RNNs}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2024},
  url = {https://icml.cc/virtual/2024/poster/34560},
}

@inproceedings{icml2025_implicit_46119,
  title = {{Implicit Language Models are RNNs: Balancing Parallelization and Expressivity}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2025},
  url = {https://icml.cc/virtual/2025/poster/46440},
}

@inproceedings{icml2026_why_49393,
  title = {{Why Are Linear RNNs More Parallelizable?}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=29sn1uqWn3},
}

@inproceedings{icml2026_a_54311,
  title = {{A unified theory of feature learning in RNNs and DNNs}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=nkVt0sY9HJ},
}

@inproceedings{icml2026_improving_54823,
  title = {{Improving the Performance and Learning Stability of Parallelizable RNNs Designed for Ultra-Low Power Applications}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=t2bQt6M7Qh},
}

@inproceedings{neurips2024_separations_70582,
  title = {{Separations in the Representational Capabilities of Transformers and Recurrent Architectures}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2024},
  url = {https://neurips.cc/virtual/2024/poster/96535},
}

@inproceedings{neurips2025_plstm_74960,
  title = {{pLSTM: parallelizable Linear Source Transition Mark networks}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=2sa13vyCn0},
}

@inproceedings{neurips2025_fixed_76539,
  title = {{Fixed-Point RNNs: Interpolating from Diagonal to Dense}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=KT8y9pFgJE},
}

@inproceedings{neurips2025_metric_77886,
  title = {{Metric Automata Theory: A Unifying Theory of RNNs}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=YeZnsJzjii},
}

@inproceedings{neurips2025_bridging_79704,
  title = {{Bridging Expressivity and Scalability with Adaptive Unitary SSMs}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=s4zitEu2R8},
}

@inproceedings{neurips2026_m_94226,
  title = {{M\$^2\$RNN: Non-Linear RNNs with Matrix-Valued States for Scalable Language Modeling}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2026},
  url = {https://neurips.cc/virtual/2026/poster/156010},
}

@inproceedings{iclr2025_scalable_81967,
  title = {{Scalable Mechanistic Neural Networks}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2025},
  url = {https://iclr.cc/virtual/2025/poster/29818},
}