Research map / Adversarial robustness and security / Certified and empirical robustness
Neural Population Geometry Reveals the Role of Stochasticity in Robust Perception
Joel Dapello, Jenelle Feather, Hang Le, Tiago Marques, David Daniel Cox, Josh Mcdermott, James J. DiCarlo, SueYeon Chung
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 |
|---|---|---|---|
| add Gaussian noise with σ=0.125 at the cochleagram representation (ablation) | StochCochResNet50 · Word-Speaker-Noise held-out speech, L∞ attacks | adversarial performance | Improved |
| add inference-time stochasticity (ablation) | StochCochResNet50 · Word-Speaker-Noise held-out speech, L∞ attacks | adversarial robustness | Improved |
| add inference-time stochasticity (ablation) | StochCochResNet50 · Word-Speaker-Noise held-out speech, L2 attacks | adversarial robustness | Improved |
| replace Poisson-like noise with Gaussian noise (ablation) | VOneResNet50 · ImageNet validation, L∞ PGD | adversarial robustness | No clear effect |
| increase Gaussian noise level (ablation) | CIFAR-VOneNet · CIFAR-10 class and exemplar manifolds | manifold capacity | Hurt |
| increase Gaussian noise level (ablation) | CIFAR-VOneNet · CIFAR-10 clean and adversarial exemplar manifolds | manifold overlap | Hurt |
| increase Gaussian noise level (ablation) | StochCochResNet50 · Word-Speaker-Noise class and adversarial exemplar manifolds | manifold capacity | Hurt |
| increase Poisson noise ratio (ablation) | CIFAR-10 VOneNet · CIFAR-10 adversarial performance | adversarial performance | Mixed |
| replace Gaussian noise with Poisson-like noise (ablation) | CIFAR-VOneNet · CIFAR-10 adversarial performance across attack strengths | adversarial performance | Improved |
| adversarially train the model (ablation) | CochResNet50 · Word-Speaker-Noise, L2 attacks | adversarial robustness | Improved |
| increase Gaussian noise level (ablation) | CIFAR-VOneNet · CIFAR-10 VOneBlock stochastic layer | manifold width | Hurt |
| use random shell perturbations instead of FGSM perturbations to sample exemplar manifolds (ablation) | VOneNets · ImageNet exemplar manifolds | manifold capacity trends | No clear effect |
Closest papers by meaning
- Biologically Inspired Mechanisms for Adversarial Robustness (NeurIPS 2020 · similarity 0.88)
- Representational Dissimilarity Metric Spaces for Stochastic Neural Networks (ICLR 2023 · similarity 0.87)
- Adversarially trained neural representations are already as robust as biological neural representations (ICML 2022 · similarity 0.86)
- Finding Biological Plausibility for Adversarially Robust Features via Metameric Tasks (ICLR 2022 · similarity 0.86)
- Comparing noisy neural population dynamics using optimal transport distances (ICLR 2025 · similarity 0.86)
- Enhancing Adversarial Robustness in SNNs with Sparse Gradients (ICML 2024 · similarity 0.85)
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.
