Research map / Adversarial robustness and security / Certified and empirical robustness
Robust Bayesian Regression via Hard Thresholding
Fan Zheyi, Zhaohui Li, Qingpei Hu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use BRHT (Bayesian reweighting plus hard thresholding) instead of CRR under adaptive adversarial attacks | Robust regression · Synthetic regression under AAA | L2 error of coefficients | Improved |
| Use CRR instead of BRHT under oblivious adversarial attacks | Robust regression · Synthetic regression under OAA | L2 error of coefficients | Mixed |
| Add a coefficient prior to hard-thresholding robust regression (TRIP) | Hard thresholding regression · Synthetic regression under AAA | L2 error of coefficients | Improved |
| Add a coefficient prior (TRIP) under oblivious attacks | Hard thresholding regression · Synthetic regression under OAA | L2 error of coefficients | Hurt |
| Replace TRIP with BRHT (Bayesian reweighting) under AAA | TRIP · Synthetic regression under AAA | convergence | Improved |
| Use Rob-ULA instead of BRHT under OAA | Robust Bayesian inference · Synthetic regression under OAA | L2 error of coefficients | Hurt |
| Use RRBR (no hard thresholding) instead of BRHT under AAA (ablation) | Reweighted robust Bayesian regression · Synthetic regression under AAA | L2 error of coefficients | Hurt |
| Vary the weight prior pr(r) of BRHT (ablation) | BRHT · Synthetic regression convergence | convergence | No clear effect |
| Use a log-normal rather than Gamma prior pr(r) when pw(w) is close to the real parameters (ablation) | BRHT · Synthetic regression convergence | convergence | Improved |
| Use a Gamma rather than log-normal prior when the coefficient prior is imprecise (ablation) | BRHT · Synthetic regression convergence | convergence | Improved |
Closest papers by meaning
- Robust Nonparametric Regression under Poisoning Attack (AAAI 2024 · similarity 0.88)
- Precise Accuracy / Robustness Tradeoffs in Regression: Case of General Norms (ICML 2024 · similarity 0.88)
- Adversarial Robustness of Nonparametric Regression (NeurIPS 2025 · similarity 0.87)
- Provably Cost-Sensitive Adversarial Defense via Randomized Smoothing (ICML 2025 · similarity 0.86)
- Treatment of Statistical Estimation Problems in Randomized Smoothing for Adversarial Robustness (NeurIPS 2024 · similarity 0.86)
- On the Adversarial Risk of Test Time Adaptation: An Investigation into Realistic Test-Time Data Poisoning (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
