AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

DeepTracer: Tracing Stolen Model via Deep Coupled Watermarks

Yunfei Yang, Xiaojun Chen, Yuexin Xuan, Zhendong Zhao, Xin Zhao, He Li

AAAI 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Neural network watermarking · 2 citations (OpenAlex)

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 changedWhereMeasuredEffect
adaptive source class selection instead of random (ablation)DeepTracer · CIFAR10WSR on stolen model: +9.35%Improved
add intra-class coupling loss to Lwm (ablation)DeepTracer · CIFAR10WSR on stolen model: 82.30 -> 89.62Improved
add inter-class coupling loss to Lwm (ablation)DeepTracer · CIFAR10WSR on stolen model: 82.30 -> 84.35Improved
combine intra- and inter-class loss with Lwm (ablation)DeepTracer · CIFAR10WSR on stolen model: 82.30 -> 94.61Improved
remove Lwm (coupling losses only) (ablation)DeepTracer · CIFAR10WSR on victim model: 100.00 -> 64.38Hurt
add two-stage key sample filtering (ablation)DeepTracer · CIFAR10WSR on stolen model: 94.61 -> 98.75Improved
filtering stage one on benign model WSR (ablation)DeepTracer · CIFAR10WSR on benign model: 0.07 -> 0Improved
use DeepTracer watermark instead of prior watermarksstolen models under hard-label attacks · CIFAR100 hard-label stealingWSR: 100.00 (Knockoff soft label)Improved
reduce quantization to 4 bits or fewer on stolen modelDeepTracer · FMNIST/CIFAR10/CIFAR100Acc and WSR: WSR 37.55 at 3 bits (CIFAR10), 0.00 at 2 bitsHurt
DeepTracer vs baselines on JBDA FMNISTJBDA soft label · FMNISTWSR: 91.65 vs 96.17 (Unrelated)No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy