Sihong Xie

Securing Behavior-based Opinion Spam Detection

2018 · arXiv (Cornell University)

Reviews spams are prevalent in e-commerce to manipulate product ranking and customers decisions maliciously. While spams generated based on simple spamming strategy can be detected effectively, hardened spammers can evade regular detectors via more advanced …

Provable Robust Saliency-based Explanations

2022 · arXiv (Cornell University)

To foster trust in machine learning models, explanations must be faithful and stable for consistent insights. Existing relevant works rely on the $\ell_p$ distance for stability assessment, which diverges from human perception. Besides, existing adversarial …

Robust Conformal Prediction under Distribution Shift via Physics-Informed Structural Causal Model

2024 · arXiv (Cornell University)

Uncertainty is critical to reliable decision-making with machine learning. Conformal prediction (CP) handles uncertainty by predicting a set on a test input, hoping the set to cover the true label with at least $(1-α)$ confidence. …