Publications

• AcceleGrad#: Adaptive Geometry-Aware Acceleration

  Hanka Goralija*, Francesco Tonin*, Kimon Antonakopoulos, Alp Yurtsever, Volkan Cevher

  NeurIPS 2026 [Paper]

• Demystifying Variance in Circuit Discovery of LLMs

  Frank Zhengqing Wu, Francesco Tonin, Volkan Cevher

  preprint [Paper]

• Sequence Modeling Architectures: Foundations

  Felix Sarnthein, Arshia Afzal, Razvan Pascanu, Antonio Orvieto, Caglar Gulcehre, Jonas Geiping, Francesco Tonin, Albert Gu, Volkan Cevher

  IEEE Signal Processing Magazine, 2026 [Paper]

• Keynotes on asymmetric kernel SVD

  Qinghua Tao, Francesco Tonin, Johan A. K. Suykens

  Cybernetics and Intelligence, 2026 [Paper]

• MaD-Mix: Multi-Modal Data Mixtures via Latent Space Coupling for Vision-Language Model Training

  Wanyun Xie°, Francesco Tonin°, Volkan Cevher

  ICLR 2026 DATA-FM Oral [Paper] [Slides]

• Efficient Large Language Model Inference with Neural Block Linearization

  Mete Erdogan°, Francesco Tonin°, Volkan Cevher

  NeurIPS 2025 [Paper] [Code]

• Linear Attention for Efficient Bidirectional Sequence Modeling

  Arshia Afzal, Elias Abad Rocamora, Leyla Naz Candogan, Pol Puigdemont, Francesco Tonin, Yongtao Wu, Mahsa Shoaran, Volkan Cevher

  NeurIPS 2025 [Paper]

• Accelerating Spectral Clustering under Fairness Constraints

  Francesco Tonin, Alex Lambert, Johan A. K. Suykens, Volkan Cevher

  ICML 2025 [Paper] [Poster]

• Chameleon: A Flexible Data-mixing Framework for Language Model Pretraining and Finetuning

  Wanyun Xie, Francesco Tonin, Volkan Cevher

  ICML 2025 [Paper] [Code]

• Chameleon: A Flexible Data-mixing Framework for Language Model Pretraining and Finetuning

  Wanyun Xie, Francesco Tonin, Volkan Cevher

  ICLR 2025 DATA-FM Workshop [Paper]

• Quantum-PEFT: Ultra parameter-efficient fine-tuning

  Toshiaki Koike-Akino*°, Francesco Tonin*°, Yongtao Wu, Frank Zhengqing Wu, Leyla Naz Candogan, Volkan Cevher

  ICLR 2025 [Paper] [Poster]

• HeNCler: Node Clustering in Heterophilous Graphs through Learned Asymmetric Similarity

  Sonny Achten, Zander Op de Beeck, Francesco Tonin, Volkan Cevher, Johan A. K. Suykens

  ICANN 2025 Workshop [Paper]

• Quantum-PEFT: Ultra parameter-efficient fine-tuning

  Toshiaki Koike-Akino*, Francesco Tonin*, Yongtao Wu, Leyla Naz Candogan, Volkan Cevher

  ICML 2024 ES-FOMO Workshop [Paper]

• Membership Inference Attacks against Large Vision-Language Models

  Zhan Li, Yongtao Wu, Yihang Chen, Francesco Tonin, Elias Abad Rocamora, Volkan Cevher

  NeurIPS 2024 [Paper]

• Learning in Feature Spaces via Coupled Covariances: Asymmetric Kernel SVD and Nyström method

  Qinghua Tao*°, Francesco Tonin*°, Alex Lambert, Yingyi Chen, Panagiotis Patrinos, Johan A. K. Suykens

  ICML 2024 [Paper] [Poster]

• Self-Attention through Kernel-Eigen Pair Sparse Variational Gaussian Processes

  Yingyi Chen, Qinghua Tao, Francesco Tonin, Johan A. K. Suykens

  ICML 2024 [Paper]

• Tensor-based Multi-view Spectral Clustering via Shared Latent Space

  Qinghua Tao°, Francesco Tonin°, Panagiotis Patrinos, Johan A. K. Suykens

  Information Fusion, 2024 [Paper] [Code]

• Unsupervised Neighborhood Propagation Kernel Layers for Semi-supervised Node Classification

  Sonny Achten, Francesco Tonin, Panagiotis Patrinos, Johan A. K. Suykens

  AAAI 2024 [Paper]

• Deep Kernel Principal Component Analysis for Multi-level Feature Learning

  Francesco Tonin, Qinghua Tao, Panagiotis Patrinos, Johan A. K. Suykens

  Neural Networks, 2024 [Paper] [Code]

• Primal-Attention: Self-attention through Asymmetric Kernel SVD in Primal Representation

  Yingyi Chen, Qinghua Tao, Francesco Tonin, Johan A. K. Suykens

  NeurIPS 2023 [Paper]

• Nonlinear SVD with Asymmetric Kernels: feature learning and asymmetric Nyström method

  Qinghua Tao*°, Francesco Tonin*°, Panagiotis Patrinos, Johan A. K. Suykens

  preprint [Paper]

• Combining Primal and Dual Representations in Deep Restricted Kernel Machines Classifiers

  Francesco Tonin, Panagiotis Patrinos, Johan A. K. Suykens

  ECML 2023 SCEFA Workshop [Paper]

• Extending Kernel PCA through Dualization: Sparsity, Robustness and Fast Algorithms

  Francesco Tonin*°, Alex Lambert*, Panagiotis Patrinos, Johan A. K. Suykens

  ICML 2023 [Paper] [Code] [Poster] [Project Page]

• Unsupervised Learning of Disentangled Representations in Deep Restricted Kernel Machines with Orthogonality Constraints

  Francesco Tonin, Panagiotis Patrinos, Johan A. K. Suykens

  Neural Networks, 2021 [Paper] [Code]

• Unsupervised Energy-based Out-of-distribution Detection using Stiefel-Restricted Kernel Machine

  Francesco Tonin, Arun Pandey, Panagiotis Patrinos, Johan A. K. Suykens

  IJCNN 2021 [Paper] [Slides] [Code]

* indicates equal contribution; ° indicates corresponding author.