 Loss Spike in Training Neural NetworksJournal of Computational Mathematics (JCM), 2023 |
 Understanding the Initial Condensation of Convolutional Neural NetworksCSIAM Transactions on Applied Mathematics (TCAM), 2023 |
 Functional Equivalence and Path Connectivity of Reducible Hyperbolic
Tangent NetworksNeural Information Processing Systems (NeurIPS), 2023 |
 Embedding Principle in Depth for the Loss Landscape Analysis of Deep Neural NetworksCSIAM Transactions on Applied Mathematics (TCAM), 2022 |
 Embedding Principle of Loss Landscape of Deep Neural NetworksNeural Information Processing Systems (NeurIPS), 2021 |
 Geometry of the Loss Landscape in Overparameterized Neural Networks:
Symmetries and InvariancesInternational Conference on Machine Learning (ICML), 2021 |
 Towards Understanding the Condensation of Neural Networks at Initial
TrainingNeural Information Processing Systems (NeurIPS), 2021 |
 When Are Solutions Connected in Deep Networks?Neural Information Processing Systems (NeurIPS), 2021 |
