Publications

2024

1. 

   NeurIPS

   MambaLRP: Explaining Selective State Space Sequence Models

   Rezaei Jafari, F., Montavon, G., Müller, K-R., and Eberle, O.

   Conference on Neural Information Processing Systems (NeurIPS) 2024

   Abs PDF Website

   <p text-align: justify;>We propose MambaLRP, a novel algorithm within the LRP framework, which ensures a more stable and reliable relevance propagation through these components. Our proposed method is theoretically sound and excels in achieving state-of-the-art explanation performance across a diverse range of models and datasets. Moreover, MambaLRP facilitates a deeper inspection of Mamba architectures, uncovering various biases and evaluating their significance. It also enables the analysis of previous speculations regarding the long-range capabilities of Mamba models.</p>

2022

1. 

   ECCV Oral

   Adaptive Token Sampling For Efficient Vision Transformers

   Fayyaz, M.*, Abbasi Kouhpayegani, S.*,  Rezaei Jafari, F.*, Sengupta, S., Sommerlade, E., Vaezi Joze, H., Pirsiavash, H., and Gall, J.

   European Conference on Computer Vision (ECCV) 2022

   * denotes equal contribution.

   Abs PDF Website

   <p text-align: justify;>In this work, we therefore introduce a differentiable parameter-free Adaptive Token Sampler (ATS) module, which can be plugged into any existing vision transformer architecture. ATS empowers vision transformers by scoring and adaptively sampling significant tokens. As a result, the number of tokens is not constant anymore and varies for each input image. By integrating ATS as an additional layer within the current transformer blocks, we can convert them into much more efficient vision transformers with an adaptive number of tokens. Since ATS is a parameter-free module, it can be added to the off-the-shelf pre-trained vision transformers as a plug and play module, thus reducing their GFLOPs without any additional training. Moreover, due to its differentiable design, one can also train a vision transformer equipped with ATS. We evaluate the efficiency of our module in both image and video classification tasks by adding it to multiple SOTA vision transformers. Our proposed module improves the SOTA by reducing their computational costs (GFLOPs) by 2 times, while preserving their accuracy on the ImageNet, Kinetics-400, and Kinetics-600 datasets.</p>