Research map / Adversarial robustness and security / Attacks on text and graphs
"SIMBA: Split Inference - Mechanisms, Benchmarks and Attacks"
Abhishek Singh, Vivek Sharma, Rohan Sukumaran, John J Mose, Jeffrey K Chiu, Justin Yu, Ramesh Raskar
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 data-driven obfuscation instead of data-independent noise or blur | obfuscation mechanisms · privacy-utility trade-off on image classification | privacy-utility trade-off | Improved |
| split the network at a deeper layer (ablation) | reconstruction attacks on ResNet-18 activations · CelebA and UTKFace | reconstruction quality (SSIM) | Hurt |
| use model-space/input-space optimization attack (A.2) instead of learned inversion (A.1) | feature inversion attack · ResNet-18 split layers | reconstruction quality | Mixed |
| give the attacker a larger share of training data (ablation) | attack A.1 · UTKFace, various split layers | reconstruction quality | Improved |
| train the attacker on a different face dataset (distribution shift) (ablation) | attack A.1 · FairFace to UTKFace, fifth split layer | reconstruction quality | No clear effect |
Closest papers by meaning
- Theoretical Insights in Model Inversion Robustness and Conditional Entropy Maximization for Collaborative Inference Systems (CVPR 2025 · similarity 0.87)
- Reconstruction Attacks on Machine Unlearning: Simple Models are Vulnerable (NeurIPS 2024 · similarity 0.86)
- DRAG: Data Reconstruction Attack using Guided Diffusion (ICML 2025 · similarity 0.86)
- A Sample-Level Evaluation and Generative Framework for Model Inversion Attacks (AAAI 2025 · similarity 0.86)
- DECO-Bench: Unified Benchmark for Decoupled Task-Agnostic Synthetic Data Release (NeurIPS 2024 · similarity 0.86)
- A Stealthy Wrongdoer: Feature-Oriented Reconstruction Attack against Split Learning (CVPR 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.
