paper-with-me

홈 › Papers

SkipPLUS: Skip the First Few Layers to Better Explain Vision Transformers

2024-06-17 · Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops 2024 6 · Faridoun Mehri, Mohsen Fayyaz, Mahdieh Soleymani Baghshah, Mohammad Taher Pilehvar

Despite their remarkable performance, the explainability of Vision Transformers (ViTs) remains a challenge. While forward attention-based token attribution techniques have become popular in text processing, their suitability for ViTs hasn’t been extensively explored. In this paper, we compare these methods against state-of-the-art input attribution methods from the Vision literature, revealing their limitations due to improper aggregation of information across layers. To address this, we introduce two general techniques, PLUS and SkipPLUS, that can be composed with any input attribution method to more effectively aggregate information across layers while handling noisy layers. Through comprehensive and quantitative evaluations of faithfulness and human interpretability on a variety of ViT architectures and datasets, we demonstrate the effectiveness of PLUS and SkipPLUS, establishing a new state-of-the-art in white-box token attribution. We conclude with a comparative analysis highlighting the strengths and weaknesses of the best versions of all the studied methods. The code used in this paper is freely available at https://github.com/NightMachinery/SkipPLUS-CVPR-2024.

📄 PDF Abstract BibTeX

Code (1)

NightMachinery/SkipPLUS-CVPR-2024 pytorch

Similar Papers 제목 키워드 기반

Image Restoration Using Very Deep Convolutional Encoder-Decoder Networks with Symmetric Skip Connections

2016-03-30 · NeurIPS 2016 12 · Xiao-Jiao Mao, Chunhua Shen, Yu-Bin Yang

In this paper, we propose a very deep fully convolutional encoding-decoding framework for image restoration such as denoising and super-resolution. The network is composed of multiple layers of convolution and de-convolu…

DecoderDenoisingImage RestorationImage Super-Resolution+1

Transformer Layers as Painters

2024-07-12 · Qi Sun, Marc Pickett, Aakash Kumar Nain, Llion Jones

Despite their nearly universal adoption for large language models, the internal workings of transformers are not well understood. We aim to better understand the impact of removing or reorganizing information throughout …

Log-DenseNet: How to Sparsify a DenseNet

2017-10-30 · ICLR 2018 1 · Hanzhang Hu, Debadeepta Dey, Allison Del Giorno, Martial Hebert 외

Skip connections are increasingly utilized by deep neural networks to improve accuracy and cost-efficiency. In particular, the recent DenseNet is efficient in computation and parameters, and achieves state-of-the-art pre…

Semantic Segmentation

Skip a Layer or Loop It? Learning Program-of-Layers in LLMs

2026-06-04 · Ziyue Li, Yang Li, Tianyi Zhou arxiv

Large language models (LLMs) perform inference by following a fixed depth and order, non-recurrent execution of all layers. We reveal the wide existence of training-free, flexible, dynamic program-of-layers (PoLar), wher…

Mathematical Reasoning

Improving Machine Translation by Searching Skip Connections Efficiently

2021-01-01 · Chen Yang, Houfeng Wang

As a widely used neural network model in NLP (neural language processing), transformer model achieves state-of-the-art performance in several translation tasks. Transformer model has a fixed skip connection architecture …

Machine TranslationTranslation