Doan, Khoa D.
16 publications
NeurIPS
2025
Mitigating Reward Over-Optimization in Direct Alignment Algorithms with Importance Sampling
ICLRW
2025
Synthesizing Physical Backdoor Datasets: An Automated Framework Leveraging Deep Generative Models
ICLR
2024
Understanding the Robustness of Randomized Feature Defense Against Query-Based Adversarial Attacks