Understanding Transformer-based Vision Models through Inversion
This work addresses the problem of interpretability in transformer-based vision models for researchers and practitioners in machine learning and computer vision, though it is incremental as it builds on existing feature inversion techniques.
The study tackled the challenge of understanding deep neural networks by revisiting feature inversion, introducing a modular variation that enables more efficient application to transformer-based vision models like Detection Transformer and Vision Transformer, resulting in qualitative interpretations and quantitative evaluations that reveal insights into how these models encode contextual shape, image details, layer correlations, and robustness against color perturbations.
Understanding the mechanisms underlying deep neural networks remains a fundamental challenge in machine learning and computer vision. One promising, yet only preliminarily explored approach, is feature inversion, which attempts to reconstruct images from intermediate representations using trained inverse neural networks. In this study, we revisit feature inversion, introducing a novel, modular variation that enables significantly more efficient application of the technique. We demonstrate how our method can be systematically applied to the large-scale transformer-based vision models, Detection Transformer and Vision Transformer, and how reconstructed images can be qualitatively interpreted in a meaningful way. We further quantitatively evaluate our method, thereby uncovering underlying mechanisms of representing image features that emerge in the two transformer architectures. Our analysis reveals key insights into how these models encode contextual shape and image details, how their layers correlate, and their robustness against color perturbations. These findings contribute to a deeper understanding of transformer-based vision models and their internal representations. The code for reproducing our experiments is available at github.com/wiskott-lab/inverse-tvm.