Publications

* denotes equal contribution.

Preprints

1. Tight certified robustness via min-max representations of ReLU neural networks, 2023, under review
        B. G. Anderson, S. Pfrommer, and S. Sojoudi

2. Towards optimal branching of linear and semidefinite relaxations for neural network robustness certification, 2023, under review
        B. G. Anderson, Z. Ma, J. Li, and S. Sojoudi

3. Asymmetric certified robustness via feature-convex neural networks, 2023, under review
        S. Pfrommer*, B. G. Anderson*, J. Piet, and S. Sojoudi

4. Projected randomized smoothing for certified adversarial robustness, 2023, under review
        INFORMS Data Mining Best Student Paper Award Runner-Up
        S. Pfrommer, B. G. Anderson, and S. Sojoudi

5. Improving the accuracy-robustness trade-off of classifiers via adaptive smoothing, 2023, under review
        Second place method on RobustBench CIFAR-100 \(\ell_\infty\)-leaderboard as of May 2023
        Y. Bai, B. G. Anderson, A. Kim, and S. Sojoudi

Journal Papers

1. Data-driven certification of neural networks with random input noise, IEEE Transactions on Control of Network Systems (TCNS), 2022
        B. G. Anderson and S. Sojoudi

Conference Papers

1. Mixing classifiers to alleviate the accuracy-robustness trade-off, 7th IEEE Conference on Control Technology and Applications (CCTA), 2023
        Y. Bai, B. G. Anderson, and S. Sojoudi

2. An overview and prospective outlook on robust training and certification of machine learning models, IFAC Symposium on System Structure and Control (SSSC), 2022
        B. G. Anderson*, T. Gautam*, and S. Sojoudi

3. A sequential greedy approach for training implicit deep models, 61st IEEE Conference on Decision and Control (CDC), 2022
        T. Gautam, B. G. Anderson, S. Sojoudi, and L. El Ghaoui

4. Towards optimal randomized smoothing: A semi-infinite linear programming approach, ICML Workshop on Formal Verification of Machine Learning (WFVML), 2022
        B. G. Anderson, S. Pfrommer, and S. Sojoudi

5. Certified robustness via locally biased randomized smoothing, 4th Annual Learning for Dynamics and Control Conference (L4DC), 2022
        B. G. Anderson and S. Sojoudi

6. Node-variant graph filters in graph neural networks, IEEE Data Science and Learning Workshop (DSLW), 2022
        F. Gama, B. G. Anderson, and S. Sojoudi

7. Tightened convex relaxations for neural network robustness certification, 59th IEEE Conference on Decision and Control (CDC), 2020
        B. G. Anderson, Z. Ma, J. Li, and S. Sojoudi

8. Global optimality guarantees for nonconvex unsupervised video segmentation, 57th Annual Allerton Conference on Communication, Control, and Computing, 2019
        B. G. Anderson and S. Sojoudi

9. Quantitative assessment of robotic swarm coverage, 15th International Conference on Informatics in Control, Automation and Robotics (ICINCO), 2018
        B. G. Anderson, E. Loeser, M. Gee, F. Ren, S. Biswas, O. Turanova, M. Haberland, and A. L. Bertozzi

Book Chapters

1. Quantifying robotic swarm coverage, Lecture Notes in Electrical Engineering (LNEE), volume 613, pp. 276–301, Springer, 2019
        B. G. Anderson, E. Loeser, M. Gee, F. Ren, S. Biswas, O. Turanova, M. Haberland, and A. L. Bertozzi

Theses

1. Towards optimality and robustness guarantees for data-driven learning and decision making, M.S. Thesis, University of California, Berkeley, 2020