Signals, Waves, and Learning: A Data-Driven Paradigm for Wave-Based Inverse Problems

Informazioni relative al progetto

SWING

ID dell’accordo di sovvenzione: 852821

DOI

10.3030/852821

Progetto chiuso

Data della firma CE 30 Ottobre 2019

Data di avvio 1 Gennaio 2020

Data di completamento 31 Dicembre 2024

Finanziato da

• EXCELLENT SCIENCE - European Research Council (ERC)

Costo totale

€ 1 986 430,00

Contributo UE

€ 1 986 430,00

CORDIS fornisce collegamenti ai risultati finali pubblici e alle pubblicazioni dei progetti ORIZZONTE.

I link ai risultati e alle pubblicazioni dei progetti del 7° PQ, così come i link ad alcuni tipi di risultati specifici come dataset e software, sono recuperati dinamicamente da .OpenAIRE .

Pubblicazioni

Other (8)

Neural equilibria for long-term prediction of nonlinear conservation laws

Autori: J. Antonio Lara Benitez, Junyi Guo, Kareem Hegazy, Ivan Dokmanić, Michael W. Mahoney, Maarten V. de Hoop
Pubblicato in: 2025
Editore: arXiv
DOI: 10.48550/arxiv.2501.06933

Conditional Injective Flows for Bayesian Imaging

Autori: AmirEhsan Khorashadizadeh, Konik Kothari, Leonardo Salsi, Ali Aghababaei Harandi, Maarten de Hoop, Ivan Dokmanić
Pubblicato in: 2022
Editore: NA

A spring-block theory of feature learning in deep neural networks

Autori: Cheng Shi, Liming Pan, Ivan Dokmanić
Pubblicato in: 2024
Editore: arXiv

An Approximation Theory for Metric Space-Valued Functions With A View Towards Deep Learning

Autori: Kratsios, Anastasis; Liu, Chong; Lassas, Matti; de Hoop, Maarten V.; Dokmanić, Ivan
Pubblicato in: Numero 4, 2023
Editore: arXiv
DOI: 10.48550/arxiv.2304.12231

LoFi: Neural Local Fields for Scalable Image Reconstruction

Autori: AmirEhsan Khorashadizadeh, Tobías I. Liaudat, Tianlin Liu, Jason D. McEwen, Ivan Dokmanić
Pubblicato in: 2024
Editore: arXiv

End-to-end localized deep learning for Cryo-ET

Autori: Vinith Kishore, Valentin Debarnot, Ricardo D. Righetto, AmirEhsan Khorashadizadeh, Ivan Dokmanić
Pubblicato in: 2025
Editore: arXiv
DOI: 10.48550/arxiv.2501.15246

Orthogonal Matrix Retrieval with Spatial Consensus for 3D Unknown-View Tomography

Autori: Shuai Huang, Mona Zehni, Ivan Dokmanic, Zhizhen Zhao
Pubblicato in: arXiv, 2022
Editore: NA

High-Rate Phase Association with Travel Time Neural Fields

Autori: Cheng Shi, Giulio Poggiali, Chris Marone, Maarten V. de Hoop, Ivan Dokmanić
Pubblicato in: 2024
Editore: arXiv
DOI: 10.48550/arxiv.2307.07572

Peer reviewed articles (16)

Total Least Squares Phase Retrieval

Autori: Sidharth Gupta, Ivan Dokmanic
Pubblicato in: IEEE Transactions on Signal Processing, Numero 70, 2024, Pagina/e 536-549, ISSN 1053-587X
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tsp.2021.3128750

Ice-Tide: Implicit Cryo-ET Imaging and Deformation Estimation

Autori: Valentin Debarnot, Vinith Kishore, Ricardo D. Righetto, Ivan Dokmanić
Pubblicato in: IEEE Transactions on Computational Imaging, Numero 11, 2025, Pagina/e 24-35, ISSN 2333-9403
Editore: IEEE
DOI: 10.1109/tci.2024.3519805

Differentiable Uncalibrated Imaging

Autori: Sidharth Gupta, Konik Kothari, Valentin Debarnot, Ivan Dokmanić
Pubblicato in: IEEE Transactions on Computational Imaging, Numero 10, 2024, Pagina/e 1-16, ISSN 2333-9403
Editore: IEEE
DOI: 10.1109/tci.2023.3346294

Deep Injective Prior for Inverse Scattering

Autori: AmirEhsan Khorashadizadeh, Vahid Khorashadizadeh, Sepehr Eskandari, Guy A. E. Vandenbosch, Ivan Dokmanić
Pubblicato in: IEEE Transactions on Antennas and Propagation, Numero 71, 2023, Pagina/e 8894-8906, ISSN 0018-926X
Editore: Institute of Electrical and Electronics Engineers
DOI: 10.1109/tap.2023.3312818

Conditional Injective Flows for Bayesian Imaging

Autori: AmirEhsan Khorashadizadeh, Konik Kothari, Leonardo Salsi, Ali Aghababaei Harandi, Maarten de Hoop, Ivan Dokmanić
Pubblicato in: IEEE Transactions on Computational Imaging, Numero 9, 2024, Pagina/e 224-237, ISSN 2333-9403
Editore: IEEE
DOI: 10.1109/tci.2023.3248949

Injectivity of ReLU networks: Perspectives from statistical physics

Autori: Antoine Maillard, Afonso S. Bandeira, David Belius, Ivan Dokmanić, Shuta Nakajima
Pubblicato in: Applied and Computational Harmonic Analysis, Numero 76, 2025, Pagina/e 101736, ISSN 1063-5203
Editore: Academic Press
DOI: 10.1016/j.acha.2024.101736

Learning Multiscale Convolutional Dictionaries for Image Reconstruction

Autori: Tianlin Liu, Anadi Chaman, David Belius, Ivan Dokmanić
Pubblicato in: IEEE Transactions on Computational Imaging, 2022, ISSN 2333-9403
Editore: IEEE

WaveBench: Benchmarking Data-driven Solvers for Linear Wave Propagation PDEs

Autori: Liu, Tianlin; Lara Benitez, Jose Antonio; Faucher, Florian; Khorashadizadeh, Amirehsan; de Hoop, Maarten; Dokmanić, Ivan
Pubblicato in: Transactions on Machine Learning Research Journal, Numero 1, 2024, ISSN 2835-8856
Editore: OpenReview.net

Homophily modulates double descent generalization in graph convolution networks

Autori: Cheng Shi, Liming Pan, Hong Hu, Ivan Dokmanić
Pubblicato in: Proceedings of the National Academy of Sciences, Numero 121, 2024, ISSN 0027-8424
Editore: National Academy of Sciences
DOI: 10.1073/pnas.2309504121

Learning sub-patterns in piecewise continuous functions

Autori: Kratsios, Anastasis; Zamanlooy, Behnoosh
Pubblicato in: Neurocomputing, Numero 09252312, 2022, ISSN 0925-2312
Editore: Elsevier BV
DOI: 10.1016/j.neucom.2022.01.036

Orthogonal Matrix Retrieval with Spatial Consensus for 3D Unknown View Tomography

Autori: Shuai Huang, Mona Zehni, Ivan Dokmanić, Zhizhen Zhao
Pubblicato in: SIAM Journal on Imaging Sciences, Numero 16, 2024, Pagina/e 1398-1439, ISSN 1936-4954
Editore: Society for Industrial and Applied Mathematics
DOI: 10.1137/22m1498218

Globally Injective ReLU Networks

Autori: Michael Puthawala, Konik Kothari, Matti Lassas, Ivan Dokmanić, Maarten de Hoop
Pubblicato in: Journal of Machine Learning Research, 2022, ISSN 1532-4435
Editore: MIT Press

Learning Multiscale Convolutional Dictionaries for Image Reconstruction

Autori: Tianlin Liu, Anadi Chaman, David Belius, Ivan Dokmanic
Pubblicato in: IEEE Transactions on Computational Imaging, Numero 8, 2022, Pagina/e 425-437, ISSN 2333-9403
Editore: IEEE
DOI: 10.1109/tci.2022.3175309

Total Least Squares Phase Retrieval

Autori: Sidharth Gupta, Ivan Dokmanić
Pubblicato in: IEEE Transactions on Signal Processing, 2021, ISSN 1941-0476
Editore: IEEE

A Graph Dynamics Prior for Relational Inference

Autori: Liming Pan, Cheng Shi, Ivan Dokmanic
Pubblicato in: Proceedings of the AAAI Conference on Artificial Intelligence, Numero 38, 2024, Pagina/e 14508-14516, ISSN 2374-3468
Editore: AAAI Press
DOI: 10.1609/aaai.v38i13.29366

Designing universal causal deep learning models: The geometric (Hyper)transformer

Autori: Beatrice Acciaio, Anastasis Kratsios, Gudmund Pammer
Pubblicato in: Mathematical Finance, Numero 34, 2024, Pagina/e 671-735, ISSN 0960-1627
Editore: Blackwell Publishing Inc.
DOI: 10.1111/mafi.12389

Conference proceedings (14)

Truly shift-invariant convolutional neural networks

Autori: Anadi Chaman, Ivan Dokmanic
Pubblicato in: 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023, Pagina/e 3772-3782
Editore: IEEE
DOI: 10.1109/cvpr46437.2021.00377

Universal Joint Approximation of Manifolds and Densities by Simple Injective Flows

Autori: Michael Puthawala, Matti Lassas, Ivan Dokmanic, Maarten De Hoop
Pubblicato in: Proceedings of the 39th International Conference on Machine Learning, 2022
Editore: PMLR

Joint Graph Rewiring and Feature Denoising via Spectral Resonance

Autori: Jonas Linkerhägner, Cheng Shi, Ivan Dokmanić
Pubblicato in: The Thirteenth International Conference on Learning Representations (ICLR), 2025
Editore: OpenReview.net
DOI: 10.48550/arxiv.2408.07191

Deep Variational Inverse Scattering : (Invited Paper)

Autori: AmirEhsan Khorashadizadeh, Ali Aghababaei, Tin Vlašić, Hieu Nguyen, Ivan Dokmanić
Pubblicato in: 2023 17th European Conference on Antennas and Propagation (EuCAP), 2024, Pagina/e 1-5
Editore: IEEE
DOI: 10.23919/eucap57121.2023.10133365

Neural Link Prediction with Walk Pooling

Autori: Liming Pan, Cheng Shi, Ivan Dokmanic
Pubblicato in: International Conference on Learning Representations, 2022
Editore: Tenth International Conference on Learning Representations

Universal Approximation Under Constraints is Possible with Transformers

Autori: Anastasis Kratsios, Behnoosh Zamanlooy, Tianlin Liu, Ivan Dokmanic
Pubblicato in: International Conference on Learning Representations, 2022
Editore: ICLR

Truly Shift-Invariant Convolutional Neural Networks

Autori: Anadi Chaman, Ivan Dokmanic
Pubblicato in: IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021
Editore: IEEE

Trumpets: Injective Flows for Inference and Inverse Problems

Autori: Kothari, Konik; Khorashadizadeh, AmirEhsan; de Hoop, Maarten; Dokmanić, Ivan
Pubblicato in: Conference on Uncertainty in Artificial Intelligence (UAI), 2021
Editore: NA

Implicit Neural Representation for Mesh-Free Inverse Obstacle Scattering

Autori: Tin Vlašić, Hieu Nguyen, Ivan Dokmanić
Pubblicato in: Asilomar Conference on Signals, Systems, and Computers, 2022
Editore: NA

Truly shift-equivariant convolutional neural networks with adaptive polyphase upsampling

Autori: Anadi Chaman, Ivan Dokmanić
Pubblicato in: 2021 55th Asilomar Conference on Signals, Systems, and Computers, 2021
Editore: IEEE

Manifold Rewiring for Unlabeled Imaging

Autori: Valentin Debarnot, Vinith Kishore, Cheng Shi, Ivan Dokmanic
Pubblicato in: APSIPA, 2022
Editore: IEEE

Joint Cryo-ET Alignment and Reconstruction with Neural Deformation Fields

Autori: Valentin Debarnot, Sidharth Gupta, Konik Kothari, Ivan Dokmanić
Pubblicato in: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024, Pagina/e 1-5
Editore: IEEE
DOI: 10.1109/icassp49357.2023.10095454

Learning the Geometry of Wave-Based Imaging

Autori: Konik Kothari, Maarten de Hoop, Ivan Dokmanić
Pubblicato in: Advances in Neural Information Processing Systems, Numero 33, 2020
Editore: Advances in Neural Information Processing Systems

SeisLM: a Foundation Model for Seismic Waveforms

Autori: Tianlin Liu, Jannes Münchmeyer, Laura Laurenti, Chris Marone, Maarten V. de Hoop, Ivan Dokmanić
Pubblicato in: NeurIPS 2024 workshop on Foundation Models for Science: Progress, Opportunities, and Challenges, 2024
Editore: NA