The knowledge of a deep learning model may be transferred to a student model, leading to intellectual property infringement or vulnerability propagation. Detecting such knowledge reuse is nontrivial because the suspect models may not be white-box accessible and/or may serve different tasks. In this paper, we propose ModelDiff, a testing-based approach to deep learning model similarity comparison. Instead of directly comparing the weights, activations, or outputs of two models, we compare their behavioral patterns on the same set of test inputs. Specifically, the behavioral pattern of a model is represented as a decision distance vector (DDV), in which each element is the distance between the model's reactions to a pair of inputs. The knowledge similarity between two models is measured with the cosine similarity between their DDVs. To evaluate ModelDiff, we created a benchmark that contains 144 pairs of models that cover most popular model reuse methods, including transfer learning, model compression, and model stealing. Our method achieved 91.7% correctness on the benchmark, which demonstrates the effectiveness of using ModelDiff for model reuse detection. A study on mobile deep learning apps has shown the feasibility of ModelDiff on real-world models.
Protecting Deep Learning Model Copyrights with Adversarial Example-Free Reuse Detection
School of Mathematical Sciences, Peking University, China(北京大学数学科学学院);School of Computer Science, University of Bristol, UK(英国布里斯托大学计算机科学学院);School of Control Science and Engineering, Zhejiang University, China(浙江大学控制科学与工程学院)
提出 NFARD,利用神经元功能与正常样本检测 DNN 复用关系,支持白盒、黑盒及异构场景,并构建 Reuse Zoo,检测 F1 最高达 1.0。
Similarity-based context aware continual learning for spiking neural networks
Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);Chinese Academy of Sciences(中国科学院);Center for Long-term Artificial Intelligence(长期人工智能中心)
Deep neural network based adaptive learning for switched systems
ShanghaiTech University(上海科技大学);Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences(中国科学院上海微系统与信息技术研究所);University of Chinese Academy of Sciences(中国科学院大学)
The Hong Kong University of Science and Technology(香港科技大学);Singapore Management University(新加坡管理大学);University of Alberta(阿尔伯塔大学)
由于它们在异构硬件设备上的广泛使用,深度学习(DL)模型由DL编译器编译成可执行文件,以充分利用低级硬件原语。这种方法允许在各种计算平台上以低成本进行DL计算,包括CPU、GPU和各种硬件加速器。我们提出了BTD(Bin to DNN),这是一个针对深度神经网络(DNN)可执行文件的反编译器。BTD获取DNN可执行文件并输出完整的模型规范,包括DNN操作符的类型、网络拓扑、维度和参数,这些参数与输入模型的参数(几乎)相同。BTD提供了一个实用的框架来处理由不同DL编译器编译的DNN可执行文件,并在x86平台上启用了全面优化。它采用基于学习的技术来推断DNN运算符,动态分析来揭示网络架构,符号执行来帮助推断DNN运算符的维度和参数。我们的评估表明,BTD能够准确恢复具有数百万参数的复杂DNN的完整规范(例如,ResNet)。恢复的DNN规范可以被重新编译为新的DNN可执行文件,该新的DNN可执行文件表现出与输入可执行文件相同的行为。我们表明,BTD可以提高两个代表性的攻击,对抗性的例子生成和知识窃取,对DNN可执行文件。我们还展示了使用BTD的跨架构遗留代码重用,并设想将BTD用于其他关键的下游任务,如DNN安全加固和修补。
Toward a Robust R2D2 Paradigm for Radio-interferometric Imaging: Revisiting Deep Neural Network Training and Architecture
Institute of Sensors, Signals and Systems, Heriot-Watt University, Edinburgh EH14 4AS, United Kingdom(传感器、信号与系统研究所,赫里奥特-瓦特大学);EPCC, University of Edinburgh, Potterrow, Edinburgh EH8 9BT, United Kingdom(爱丁堡大学埃平中心)