Research map / Adversarial robustness and security / Certified and empirical robustness
Smoothed Embeddings for Certified Few-Shot Learning
Mikhail Pautov, Olesya Kuznetsova, Nurislam Tursynbek, Aleksandr Petiushko, Ivan Oseledets
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 |
|---|---|---|---|
| Increase number of noise samples n from 10^3 to 10^5 (ablation) | certification Algorithm · Cub-200-2011 | time per image (s): 0.044 -> 4.744 | Hurt |
| Increase n from 10^3 to 10^4 (ablation) | certification Algorithm · Cub-200-2011 | time per image (s): 0.044 -> 0.509 | Hurt |
| Decrease alpha from 1e-2 to 1e-4 (ablation) | smoothed ProtoNet · CIFAR-FS 5-shot | non-certified objects (%): 3.0 -> 3.8 | Hurt |
| Decrease alpha from 1e-2 to 1e-4 (ablation) | smoothed ProtoNet · Cub-200-2011 1-shot | non-certified objects (%): 1.6 -> 1.6 | No clear effect |
| Decrease alpha from 1e-2 to 1e-4 (ablation) | smoothed ProtoNet · miniImageNet 1-shot | non-certified objects (%): 1.9 -> 2.4 | Hurt |
| Decrease alpha from 1e-2 to 1e-4 (ablation) | smoothed ProtoNet · miniImageNet 5-shot | non-certified objects (%): 2.9 -> 3.0 | No clear effect |
| Use a smaller sample size n for certification (ablation) | smoothed ProtoNet · CIFAR-FS 5-shot | certified accuracy | No clear effect |
| Train the baseline encoder with Gaussian noise augmentation (ablation) | smoothed ProtoNet · Cub-200-2011, CIFAR-FS, miniImageNet | empirical robust accuracy | Improved |
| Random additive noise attack on plain ProtoNet | plain baseline · three datasets | empirical robust accuracy | No clear effect |
| FGSM adversarial attack on smoothed model | smoothed ProtoNet · three datasets | empirical robust accuracy | No clear effect |
| Replace one-step FGSM with 20-step PGD attack (ablation) | smoothed ProtoNet · three datasets 1-shot | empirical robust accuracy | Hurt |
Closest papers by meaning
- Provably Robust Deep Learning via Adversarially Trained Smoothed Classifiers (NeurIPS 2019 · similarity 0.90)
- A FRAMEWORK FOR ROBUSTNESS CERTIFICATION OF SMOOTHED CLASSIFIERS USING F-DIVERGENCES (ICLR 2020 · similarity 0.90)
- Improving l1-Certified Robustness via Randomized Smoothing by Leveraging Box Constraints (ICML 2023 · similarity 0.88)
- The Lipschitz-Variance-Margin Tradeoff for Enhanced Randomized Smoothing (ICLR 2024 · similarity 0.87)
- Robust Representation Consistency Model via Contrastive Denoising (ICLR 2025 · similarity 0.86)
- Provably Cost-Sensitive Adversarial Defense via Randomized Smoothing (ICML 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.
