 Robustness of Nonlinear Representation LearningInternational Conference on Machine Learning (ICML), 2025 |
 Parameter-Efficient Orthogonal Finetuning via Butterfly FactorizationInternational Conference on Learning Representations (ICLR), 2023 |
 Controlling Text-to-Image Diffusion by Orthogonal FinetuningNeural Information Processing Systems (NeurIPS), 2023 |
 Generalizing and Decoupling Neural Collapse via Hyperspherical
Uniformity GapInternational Conference on Learning Representations (ICLR), 2023 |
 TANGOS: Regularizing Tabular Neural Networks through Gradient
Orthogonalization and SpecializationInternational Conference on Learning Representations (ICLR), 2023 |
 Hyperspherical Quantization: Toward Smaller and More Accurate ModelsIEEE Workshop/Winter Conference on Applications of Computer Vision (WACV), 2022 |
 Iterative Teaching by Data HallucinationInternational Conference on Artificial Intelligence and Statistics (AISTATS), 2022 |
 Utilizing Excess Resources in Training Neural NetworksInternational Conference on Information Photonics (ICIP), 2022 |
 Data-Efficient Learning via Minimizing Hyperspherical EnergyIEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2022 |
 Hyperspherical Consistency RegularizationComputer Vision and Pattern Recognition (CVPR), 2022 |
 Feedback Gradient Descent: Efficient and Stable Optimization with
Orthogonality for DNNsAAAI Conference on Artificial Intelligence (AAAI), 2022 |
 projUNN: efficient method for training deep networks with unitary
matricesNeural Information Processing Systems (NeurIPS), 2022 |
 Towards Principled Disentanglement for Domain GeneralizationComputer Vision and Pattern Recognition (CVPR), 2021 |
 A Geometric Analysis of Neural Collapse with Unconstrained FeaturesNeural Information Processing Systems (NeurIPS), 2021 |
 Learning with Hyperspherical UniformityInternational Conference on Artificial Intelligence and Statistics (AISTATS), 2021 |
 Learning by Turning: Neural Architecture Aware OptimisationInternational Conference on Machine Learning (ICML), 2021 |
 ShiftAddNet: A Hardware-Inspired Deep NetworkNeural Information Processing Systems (NeurIPS), 2020 |
 Angular Visual HardnessInternational Conference on Machine Learning (ICML), 2019 |
 Regularizing Neural Networks via Minimizing Hyperspherical EnergyComputer Vision and Pattern Recognition (CVPR), 2019 |
