![[2602.22238] TT-SEAL: TTD-Aware Selective Encryption for Adversarially-Robust and Low-Latency Edge AI](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22238] TT-SEAL: TTD-Aware Selective Encryption for Adversarially-Robust and Low-Latency Edge AI
Summary
The paper presents TT-SEAL, a selective encryption framework designed for Tensor-Train Decomposed (TTD) networks, enhancing security and reducing latency in edge AI applications.
Why It Matters
As AI applications increasingly move to edge devices, ensuring both security and efficiency is critical. TT-SEAL addresses the challenge of maintaining model robustness while minimizing encryption overhead, making it relevant for developers and researchers focused on secure AI deployment.
Key Takeaways
- TT-SEAL enables selective encryption for TTD networks, improving security without significant latency.
- The framework uses a sensitivity-based metric to prioritize critical model components for encryption.
- TT-SEAL achieves robustness comparable to full encryption while encrypting only a fraction of model parameters.
Computer Science > Cryptography and Security arXiv:2602.22238 (cs) [Submitted on 24 Feb 2026] Title:TT-SEAL: TTD-Aware Selective Encryption for Adversarially-Robust and Low-Latency Edge AI Authors:Kyeongpil Min, Sangmin Jeon, Jae-Jin Lee, Woojoo Lee View a PDF of the paper titled TT-SEAL: TTD-Aware Selective Encryption for Adversarially-Robust and Low-Latency Edge AI, by Kyeongpil Min and 3 other authors View PDF HTML (experimental) Abstract:Cloud-edge AI must jointly satisfy model compression and security under tight device budgets. While Tensor-Train Decomposition (TTD) shrinks on-device models, prior selective-encryption studies largely assume dense weights, leaving its practicality under TTD compression unclear. We present TT-SEAL, a selective-encryption framework for TT-decomposed networks. TT-SEAL ranks TT cores with a sensitivity-based importance metric, calibrates a one-time robustness threshold, and uses a value-DP optimizer to encrypt the minimum set of critical cores with AES. Under TTD-aware, transfer-based threat models (and on an FPGA-prototyped edge processor) TT-SEAL matches the robustness of full (black-box) encryption while encrypting as little as 4.89-15.92% of parameters across ResNet-18, MobileNetV2, and VGG-16, and drives the share of AES decryption in end-to-end latency to low single digits (e.g., 58% -> 2.76% on ResNet-18), enabling secure, low-latency edge AI. Comments: Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI) Ci...