DAWN
Authors:
Open Access
- 17 October 2021
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 4417-4425
Abstract
Training machine learning (ML) models is expensive in terms of computational power, amounts of labeled data and human expertise. Thus, ML models constitute business value for their owners. Embedding digital watermarks during model training allows a model owner to later identify their models in case of theft or misuse. However, model functionality can also be stolen via model extraction, where an adversary trains a surrogate model using results returned from a prediction API of the original model. Recent work has shown that model extraction is a realistic threat. Existing watermarking schemes are ineffective against model extraction since it is the adversary who trains the surrogate model. In this paper, we introduce DAWN (Dynamic Adversarial Watermarking of Neural Networks), the first approach to use watermarking to deter model extraction theft. Unlike prior watermarking schemes, DAWN does not impose changes to the training process but operates at the prediction API of the protected model, by dynamically changing the responses for a small subset of queries (e.g., 0.5%) from API clients. This set is a watermark that will be embedded in case a client uses its queries to train a surrogate model. We show that DAWN is resilient against two state-of-the-art model extraction attacks, effectively watermarking all extracted surrogate models, allowing model owners to reliably demonstrate ownership (with confidence greater than 1-2-64), incurring negligible loss of prediction accuracy (0.03-0.5%).Keywords
This publication has 12 references indexed in Scilit:
- DeepSigns2019
- Model Extraction Warning in MLaaS Paradigm2018
- Copycat CNN: Stealing Knowledge by Persuading Confession with Random Non-Labeled Data2018
- Protecting Intellectual Property of Deep Neural Networks with Watermarking2018
- Trojaning Attack on Neural Networks2018
- MagNet2017
- Embedding Watermarks into Deep Neural Networks2017
- Practical Black-Box Attacks against Machine Learning2017
- The German Traffic Sign Recognition Benchmark: A multi-class classification competition2011
- Information hiding-a surveyProceedings of the IEEE, 1999