← Back to all news
arXiv Machine Learning By Kim-Cuc Nguyen, Ngai-Man Cheung

Dissecting Representation Structure in Vision Transformers: A Rigorous Architectural Study

Read the original on arXiv Machine Learning →

The paper presents a detailed study of feature representations in Vision Transformers (ViTs), revealing that feature collapse at initialization causes redundancy and proposing a reduction scheme to address it. It demonstrates that entropy and minimum eigenvalue metrics reliably predict generalization, and that token‑space features and linear submodule outputs are particularly informative. The authors develop a proxy that improves ranking of ViT architectures by 18‑48% over previous baselines, enabling the selection of models with higher accuracy at comparable or lower computational cost.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Computer Vision
Sep 17

Decoder-Agnostic Token Merging for Vision Transformers: A Systematic Study of G2TM

The paper studies Graph-Guided Token Merging (G2TM), a module that reduces token count in Vision Transformers. It evaluates G2TM across multiple segmentation frameworks and decoder types, finding that its performance gains are tied to the encoder rather than the decoder. The authors report consistent reductions in GFLOPs (22‑47%) and throughput improvements (up to 74%) on ADE20K, with optimal hyperparameters depending mainly on backbone pre‑training and target dataset.

By Victor Bercy, Martyna Poreba, Michal Szczepanski, Samia Bouchafa
llmscomputer-visionefficiencybenchmarks
More like this →
arXiv AI
4d ago

Scaling Vision Transformers Without Training: Prune-and-Merge Token Compression

The paper introduces Prune-and-Merge ViT, a training‑free token compression framework for Vision Transformers that combines class‑attention pruning with similarity‑based merging to reduce computational cost while preserving semantic information. It evaluates robustness using two new metrics—Relative Accuracy Drop (RAD) and Prediction Flip Rate (PFR)—and shows that the method maintains or improves accuracy on pretrained ViT and DeiT models compared to existing token reduction techniques. The authors also argue that the hybrid approach does not induce hallucination‑like behavior and keeps predictions stable under common image transformations.

By Vatsal, Soumil Malhotra, Priyanshu Chahal, Sanjeev Kumar
llmsefficiencysafety
More like this →
arXiv AI
Sep 2

HiLRP: Toward One Trustworthy Explanation for Vision Transformer: Conservation-Valid Attribution via Attention Primitives

HiLRP introduces a unified attribution framework for Vision Transformers (ViTs) that addresses the challenges posed by diverse architectural designs. By decomposing ViT operations into four basic types—linear maps, bilinear mixing, normalization/gating, and reindexing—HiLRP applies conservation‑satisfying relevance rules, enabling reliable explanations across a wide range of backbones. The method outperforms 14 existing attribution techniques on 10 architectures, maintaining conservation and improving localization accuracy (0.97 Pointing) compared to competitors.

By Sathiyamohan Nishankar, Pubudu Sanjeewani, Asanka Perera, Selvarajah Thuseethan
llmscomputer-vision
More like this →
arXiv Machine Learning
Jul 14

Vertical Fusion: Condensing Internal Representations for Robust ViT Classification

arXiv:2607. 10391v1 Announce Type: cross Abstract: Despite exposing rich intermediate representations, Vision Transformers (ViTs) are almost exclusively utilized as black-box feature extractors, where only the last layer is considered for downstream tasks.

By Francesco Di Salvo, Shyam Nandan Rai, Hamed Damirchi, Ignacio Meza De la Jara, Sebastian Doerrich, Marco Lents, Christian Ledig
llmsdiffusion
More like this →