ModelGiF: Gradient Fields for Model Functional Distance
This work addresses the problem of model comparison for researchers and practitioners dealing with diverse pre-trained models, offering a novel approach but is incremental in building on physics-inspired concepts.
The paper tackles the challenge of quantifying functional distances between heterogeneous pre-trained deep learning models by introducing ModelGiF, a gradient field representation, and demonstrates its effectiveness in tasks like task relatedness estimation and intellectual property protection with superior performance over state-of-the-art methods.
The last decade has witnessed the success of deep learning and the surge of publicly released trained models, which necessitates the quantification of the model functional distance for various purposes. However, quantifying the model functional distance is always challenging due to the opacity in inner workings and the heterogeneity in architectures or tasks. Inspired by the concept of "field" in physics, in this work we introduce Model Gradient Field (abbr. ModelGiF) to extract homogeneous representations from the heterogeneous pre-trained models. Our main assumption underlying ModelGiF is that each pre-trained deep model uniquely determines a ModelGiF over the input space. The distance between models can thus be measured by the similarity between their ModelGiFs. We validate the effectiveness of the proposed ModelGiF with a suite of testbeds, including task relatedness estimation, intellectual property protection, and model unlearning verification. Experimental results demonstrate the versatility of the proposed ModelGiF on these tasks, with significantly superiority performance to state-of-the-art competitors. Codes are available at https://github.com/zju-vipa/modelgif.