Scholar
Alex Bie
Google Scholar ID: AUJuK3AAAAAJ
University of Waterloo
Citations & Impact
Citations
290
H-index
8
i10-index
8
Publications
16
Co-authors
0
Publications
Cited
0
Resume
Academic Achievements
- Published multiple papers at top venues including NeurIPS, ICML, EMNLP, and TMLR
- Selected publications:
- • 'Clustering and median aggregation improve differentially private inference' (Preprint, 2025)
- • 'Escaping collapse: The strength of weak data for large language model training' (NeurIPS 2025)
- • 'On the learnability of distribution classes with adaptive adversaries' (ICML 2025)
- • 'Foundation Models Meet Federated Learning: A One-shot Feature-sharing Method with Privacy and Performance Guarantees' (TMLR, 2025)
- • 'RenderAttack: Hundreds of adversarial attacks through differentiable texture generation' (AdvML Frontiers @ NeurIPS 2024)
- • 'Private prediction for large-scale synthetic text generation' (EMNLP 2024 Findings)
- • 'Distribution learnability and robustness' (NeurIPS 2023, spotlight)
- • 'Private distribution learning with public data: The view from sample compression' (NeurIPS 2023)
- • 'Private GANs, revisited' (TMLR, 2023, with survey certification)
- • 'Private estimation with public data' (NeurIPS 2022)
- • 'Don't generate me: Training differentially private generative models with Sinkhorn divergence' (NeurIPS 2021)
- • 'Fully quantizing Transformer-based ASR for edge deployment' (Hardware Aware Efficient Training @ ICLR 2021)