Publications

• Karhadkar, Kedar and Murray, Michael and Tseran, Hanna and Montúfar, Guido. “Mildly Overparameterized ReLU Networks Have a Favorable Loss Landscape.” TMLR 2024. arxiv
• Tseran, Hanna, and Guido Montúfar. “Expected Gradients of Maxout Networks and Consequences to Parameter Initialization.” International Conference on Machine Learning, ICML 2023. arxiv
• Tseran, Hanna, and Guido Montúfar. “On the Expected Complexity of Maxout Networks.” Advances in Neural Information Processing Systems, NeurIPS 2021. arxiv
• Tseran, Hanna and Khan, Mohammad Emtiyaz and Harada, Tatsuya and Bui, Thang D. “Natural Variational Continual Learning.” NeurIPS Workshop on Continual Learning, 2018. pdf
• Tseran, Hanna, and Tatsuya Harada. “Memory augmented neural network with Gaussian embeddings for one-shot learning.” NeurIPS Workshop on Bayesian Deep Learning, 2017. pdf

Theses

• “Expected Complexity and Gradients of Deep Maxout Neural Networks and Implications to Parameter Initialization.”, Doctoral Thesis, Supervisor: Prof. Guido Montufar, 2023. pdf
• “Variational Inference for Continual Learning by using Weight-Perturbation in Adam.” Master’s Thesis, Supervisor: Prof. Tatsuya Harada, 2018. pdf
• “Algorithms for recognition of circular objects and elements on them (in case of coins).” Specialist Thesis, Supervisor: Prof. Yuri Svirid, 2015. (In Russian)

You can also find my publications in my Google Scholar profile.