ProReGen: Progressive Residual Generation under Attribute Correlations
Proposes a progressive residual learning approach that improves the generation of underrepresented attribute combinations across VAEs, GANs, and diffusion models.
My publications are also listed on Google Scholar . Click Citation to expand a formatted reference.
Proposes a progressive residual learning approach that improves the generation of underrepresented attribute combinations across VAEs, GANs, and diffusion models.
Introduces a difficulty-aware task sampler that improves the accuracy of meta-learned physics-informed neural networks.
Toloubidokhti, M., Ye, Y., Missel, R., Jiang, X., Kumar, N., Shrestha, R., & Wang, L. (2024). DATS: Difficulty-Aware Task Sampler for Meta-Learning Physics-Informed Neural Networks. International Conference on Learning Representations (ICLR). https://openreview.net/forum?id=EvyYFSxdgB
* Equal contribution.
Examines the limitations of robust optimization and invariant learning, then introduces an approach that combines them to reduce subgroup disparities in medical image classification.
Kumar, N., Shrestha, R., Li, Z., & Wang, L. (2023). Distributionally Robust Optimization and Invariant Representation Learning for Addressing Subgroup Underrepresentation: Mechanisms and Limitations. In Clinical Image-Based Procedures, Fairness of AI in Medical Imaging, and Ethical and Philosophical Issues in Medical Imaging (pp. 183–193). Springer. https://arxiv.org/abs/2308.06434
Addresses conflicting assumptions across data augmentation, explainable AI, and causal modeling by systematically examining how feature correlations affect minority image generation and whether existing approaches can mitigate their impact.
Presents a machine-learning pipeline with Isolation Forest outlier rejection, β-VAE denoising, and Self-Organizing-Map clustering for individual pulsar pulses to support analysis of the 2024 Vela glitch and other pulsars.
Lousto, C. O., Shrestha, R., Zubieta, E., Araujo Furlan, S. B., Gancio, G., García, F., del Palacio, S., Missel, R., & Wang, L. (2026). Single-pulse reanalysis of the 2024 Vela glitch and new observations of PSR J0437−4715 and PSR J1644−4559. arXiv:2609.03056. https://arxiv.org/abs/2609.03056