Scholar
Daouda Sow
Google Scholar ID: 2kpMGWIAAAAJ
PhD candidate at The Ohio State University
Citations & Impact
Citations
150
H-index
6
i10-index
4
Publications
12
Co-authors
6
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Paper 'Doubly Robust Instance-Reweighted Adversarial Training' accepted by ICLR 2024
- Paper 'Online Bilevel Optimization' accepted by NeurIPS 2023
- Paper 'Dynamic Online Meta-Learning' to appear at CPAL 2024
- Preprint: 'Non-Convex Bilevel Optimization with Time-Varying Objective Functions'
- Preprint: 'Doubly Robust Instance-Reweighted Adversarial Training'
- Invited talk at the 2022 INFORMS Annual Meeting in the session 'Bilevel Stochastic Methods for Optimization and Learning'
Background
- Ph.D. candidate in the Department of Electrical and Computer Engineering at The Ohio State University
- Research interests include optimization methods for modern machine learning (ML), robust ML, and deep learning theory
- Theoretically focuses on building new frameworks—especially through bilevel optimization—for modern ML applications such as meta-learning, adversarial training, and hyperparameter optimization
- Practically aims to design new algorithms with provable guarantees and develop efficient implementations
- Recently focuses on optimization for task-specific adaptation of large language models (LLMs), via direct fine-tuning (when allowed) or instruction optimization for black-box LLMs
- Affiliated with the NSF AI Institute for Future Edge Networks and Distributed Intelligence (AI-EDGE)
