Robust machine learning
This work is from 2018-present. It is supported by an AFOSR grant as well as NSF RAISE grant. Part of the work is in collaboration with Mitsubishi Electric Researc Labs (MERL), Cambrige, MA. The sub-areas include adversarial robustness, domain adaptation, domain generalization, and contrastive representation learning.
Boyang Lyu, Thuan Nguyen, Prakash Ishwar, Matthias Scheutz, and Shuchin Aeron. “Barycentric-Alignment and Reconstruction Loss Minimization for Domain Generalization”. In: IEEE Access 11 (2023), pp.49226–49240. doi: 10.1109/ACCESS.2023.3276775.
Boyang Lyu, Thuan Nguyen, Matthias Scheutz, Prakash Ishwar, and Shuchin Aeron. “APrincipled Approach to Model Validation in Domain Generalization”. In: ICASSP 2023 - 2023 IEEEInternational Conference on Acoustics, Speech and Signal Processing (ICASSP). 2023, pp. 1–5. doi:10.1109/ICASSP49357.2023.10094659.
Thuan Nguyen, Matthias Scheutz, and Shuchin Aeron. “On The Failure of Invariant Risk Minimization and an Effective fix via Classification Error Control”. In: 2023 IEEE 33rd International Workshop on Machine Learning for Signal Processing (MLSP). 2023, pp. 1–6. doi:10.1109/MLSP55844.2023.10286000.
Xi Yu, Niklas Smedemark-Margulies, Shuchin Aeron, Toshiaki Koike-Akino, Pierre Moulin, Matthew Brand, Kieran Parsons, and Ye Wang. “Improving adversarial robustness by learning shared information”. In: Pattern Recognition 134 (2023), p. 109054.
Thuan Nguyen, Boyang Lyu, Prakash Ishwar, Matthias Scheutz, and Shuchin Aeron. “Conditional entropy minimization principle for learning domain invariant representation features”. In: 2022 26th International Conference on Pattern Recognition (ICPR). 2022, pp. 3000– 3006. doi:10.1109/ICPR56361.2022.9956548.
Thuan Nguyen, Boyang Lyu, Prakash Ishwar, Matthias Scheutz, and Shuchin Aeron. “Jointcovariate-alignment and concept-alignment: a framework for domain generalization”. In: 2022 IEEE 32nd International Workshop on Machine Learning for Signal Processing (MLSP). IEEE. 2022, pp. 1–6.
James Webber, Erika Hussey, Eric Miller, and Shuchin Aeron. “On non-parametric density estimation on linear and non-linear manifolds using generalized Radon transforms”. In: Communications in Statistics - Theory and Methods 51.23 (2022), pp. 8406–8426. doi: 10.1080/ 03610926.2021.1897143.
Michael T Wojnowicz, Shuchin Aeron, Eric L Miller, and Michael Hughes. “Easy Variational Inference for Categorical Models via an Independent Binary Approximation”. In: International Conference on Machine Learning. PMLR. 2022, pp. 23857–23896.
Adnan Siraj Rakin, Ye Wang, Shuchin Aeron, Toshiaki Koike-Akino, Pierre Moulin, and Kieran Parsons. “Towards Universal Adversarial Examples and Defenses”. In: 2021 IEEE Information Theory Workshop(ITW). IEEE. 2021, pp. 1–6.
Ye Wang, Shuchin Aeron, Adnan Siraj Rakin, Toshiaki Koike-Akino, and Pierre Moulin. “Robust Machine Learning via Privacy/ Rate-Distortion Theory”. In: 2021 IEEE Interna- tional Symposium on Information Theory (ISIT). 2021, pp. 1320–1325. doi: 10.1109/ISIT45174. 2021.9517751.
Note: Shuchin Aeron is a paid consultant with Mitsubishi Electric Research Labs (MERL).