% Key papers for Theory of self-attention dynamics (scime atlas, https://atlas.scime.ai). Authors are not included: add them from the paper pages.

@inproceedings{neurips2021_twins_61359,
  title = {{Twins: Revisiting the Design of Spatial Attention in Vision Transformers}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2021},
  url = {https://nips.cc/virtual/2021/poster/27970},
}

@inproceedings{icml2019_set_37737,
  title = {{Set Transformer: A Framework for Attention-based Permutation-Invariant Neural Networks}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2019},
  url = {https://icml.cc/virtual/2019/poster/3716},
}

@inproceedings{icml2021_perceiver_39780,
  title = {{Perceiver: General Perception with Iterative Attention}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2021},
  url = {https://icml.cc/virtual/2021/poster/10291},
}

@inproceedings{aaai2021_encoding_519,
  title = {{Encoding Syntactic Knowledge in Transformer Encoder for Intent Detection and Slot Filling}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2021},
  url = {https://aaai.org/papers/13943-encoding-syntactic-knowledge-in-transformer-encoder-for-intent-detection-and-slot-filling/},
}

@inproceedings{cvpr2024_learning_23949,
  title = {{Learning Correlation Structures for Vision Transformers}},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2024},
  url = {https://cvpr.thecvf.com/virtual/2024/poster/29995},
}

@inproceedings{neurips2020_limits_60252,
  title = {{Limits to Depth Efficiencies of Self-Attention}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2020},
  url = {https://nips.cc/virtual/2020/poster/18904},
}

@inproceedings{icml2021_synthesizer_39996,
  title = {{Synthesizer: Rethinking Self-Attention for Transformer Models}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2021},
  url = {https://icml.cc/virtual/2021/poster/9307},
}

@inproceedings{cvpr2024_facial_23445,
  title = {{Facial Identity Anonymization via Intrinsic and Extrinsic Attention Distraction}},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2024},
  url = {https://cvpr.thecvf.com/virtual/2024/poster/30728},
}

@inproceedings{neurips2023_causal_67361,
  title = {{Causal Interpretation of Self-Attention in Pre-Trained Transformers}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2023},
  url = {https://nips.cc/virtual/2023/poster/72367},
}

@inproceedings{aaai2021_explicitly_560,
  title = {{Explicitly Modeled Attention Maps for Image Classification}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2021},
  url = {https://aaai.org/papers/09799-explicitly-modeled-attention-maps-for-image-classification/},
}

@inproceedings{neurips2020_kalman_60123,
  title = {{Kalman Filtering Attention for User Behavior Modeling in CTR Prediction}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2020},
  url = {https://nips.cc/virtual/2020/poster/16869},
}

@inproceedings{neurips2023_the_68641,
  title = {{The emergence of clusters in self-attention dynamics}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2023},
  url = {https://nips.cc/virtual/2023/poster/71198},
}

@inproceedings{icml2019_area_37165,
  title = {{Area Attention}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2019},
  url = {https://icml.cc/virtual/2019/poster/3744},
}

@inproceedings{icml2024_attention_45717,
  title = {{Attention Meets Post-hoc Interpretability: A Mathematical Perspective}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2024},
  url = {https://icml.cc/virtual/2024/poster/32735},
}

@inproceedings{icml2022_dual_40479,
  title = {{Dual Decomposition of Convex Optimization Layers for Consistent Attention in Medical Images}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2022},
  url = {https://icml.cc/virtual/2022/poster/17537},
}

@inproceedings{icml2022_unraveling_41319,
  title = {{Unraveling Attention via Convex Duality: Analysis and Interpretations of Vision Transformers}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2022},
  url = {https://icml.cc/virtual/2022/poster/17171},
}

@inproceedings{neurips2024_clustering_71959,
  title = {{Clustering in Causal Attention Masking}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2024},
  url = {https://neurips.cc/virtual/2024/poster/95352},
}

@inproceedings{neurips2024_dissecting_72424,
  title = {{Dissecting the Interplay of Attention Paths in a Statistical Mechanics Theory of Transformers}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2024},
  url = {https://neurips.cc/virtual/2024/poster/94932},
}

@inproceedings{neurips2025_vision_78119,
  title = {{Vision Transformers Don't Need Trained Registers}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=bA02DmQN5d},
}

@inproceedings{iclr2021_on_29965,
  title = {{On the Dynamics of Training Attention Models}},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year = {2021},
  url = {https://iclr.cc/virtual/2021/poster/2692},
}

@inproceedings{neurips2024_the_72226,
  title = {{The Closeness of In-Context Learning and Weight Shifting for Softmax Regression}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2024},
  url = {https://neurips.cc/virtual/2024/poster/95116},
}

@inproceedings{aaai2024_the_7548,
  title = {{The Inhibitor: ReLU and Addition-Based Attention for Efficient Transformers (Student Abstract)}},
  booktitle = {AAAI Conference on Artificial Intelligence (AAAI)},
  year = {2024},
  url = {https://ojs.aaai.org/index.php/AAAI/article/view/30422},
}

@inproceedings{neurips2024_dissecting_73082,
  title = {{Dissecting Query-Key Interaction in Vision Transformers}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2024},
  url = {https://neurips.cc/virtual/2024/poster/94346},
}

@inproceedings{neurips2025_benign_79959,
  title = {{Benign Overfitting in Single-Head Attention}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=ud7VVZ693U},
}

@inproceedings{icml2025_on_46226,
  title = {{On Understanding Attention-Based In-Context Learning for Categorical Data}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2025},
  url = {https://icml.cc/virtual/2025/poster/46334},
}

@inproceedings{icml2026_kuramoto_49364,
  title = {{Kuramoto Oscillatory Phase Encoding: Neuro-inspired Synchronization for Improved Learning Efficiency}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=1rSqVUl7l3},
}

@inproceedings{icml2026_focus_51348,
  title = {{Focus and Dilution: The Multi-stage Learning Process of Attention}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=LF4YsCYfrf},
}

@inproceedings{icml2026_controlled_52466,
  title = {{Controlled Dynamics Attractor Transformer}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=WLRuaimgMw},
}

@inproceedings{icml2026_quaternion_52617,
  title = {{Quaternion Self-Attention with Shared Scores}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=XdyDNR5gtb},
}

@inproceedings{icml2026_softplus_53166,
  title = {{Softplus Attention with Re-weighting Boosts Length Extrapolation in Large Language Models}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=cmTjoiRyyr},
}

@inproceedings{icml2026_a_54671,
  title = {{A Capacity-Based Rationale for Multi-Head Attention}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=rIbK1X14la},
}

@inproceedings{icml2026_krause_54709,
  title = {{Krause Synchronization Transformers}},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
  url = {https://openreview.net/forum?id=rhX58pEFNd},
}

@inproceedings{neurips2025_attention_76842,
  title = {{Attention Mechanism, Max-Affine Partition, and Universal Approximation}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=NTjr79xpX0},
}

@inproceedings{neurips2025_understanding_78237,
  title = {{Understanding Differential Transformer Unchains Pretrained Self-Attentions}},
  booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
  year = {2025},
  url = {https://openreview.net/forum?id=cWlRC064GK},
}