paper-with-me

Papers

Embedding Shift Dissection on CLIP: Effects of Augmentations on VLM's Representation Learning

2025-03-30 · Ashim Dahal, Saydul Akbar Murad, Nick Rahimi

Understanding the representation shift on Vision Language Models like CLIP under different augmentations provides valuable insights on Mechanistic Interpretability. In this study, we show the shift on CLIP's embeddings on 9 common augmentation techniques: noise, blur, color jitter, scale and rotate, flip, elastic and perspective transforms, random brightness and contrast, and coarse dropout of pixel blocks. We scrutinize the embedding shifts under similarity on attention map, patch, edge, detail preservation, cosine similarity, L2 distance, pairwise distance and dendrogram clusters and provide qualitative analysis on sample images. Our findings suggest certain augmentations like noise, perspective transform and shift scaling have higher degree of drastic impact on embedding shift. This study provides a concrete foundation for future work on VLM's robustness for mechanical interpretation and adversarial data defense.

📄 PDF Abstract BibTeX arXiv:2503.23495

Code (0)

등록된 구현이 없습니다.

Tasks

Representation Learning

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음
CLIP Contrastive Language-Image Pre-training (CLIP), consisting of a simplified version of ConVIRT trained from scratch, is an efficient method of image representation learning…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Similar Papers 제목 키워드 기반

Configuring Data Augmentations to Reduce Variance Shift in Positional Embedding of Vision Transformers

2024-05-23 · Bum Jun Kim, Sang Woo Kim

Vision transformers (ViTs) have demonstrated remarkable performance in a variety of vision tasks. Despite their promising capabilities, training a ViT requires a large amount of diverse data. Several studies empirically …

Data Augmentation

Dissecting CLIP: Decomposition with a Schur Complement-based Approach

2024-12-24 · Azim Ospanov, Mohammad Jalali, Farzan Farnia

The use of CLIP embeddings to assess the alignment of samples produced by text-to-image generative models has been extensively explored in the literature. While the widely adopted CLIPScore, derived from the cosine simil…

Diversity

Learning dissection trajectories from expert surgical videos via imitation learning with equivariant diffusion

2025-06-05 · Hongyu Wang, Yonghao Long, Yueyao Chen, Hon-Chi Yip 외

Endoscopic Submucosal Dissection (ESD) is a well-established technique for removing epithelial lesions. Predicting dissection trajectories in ESD videos offers significant potential for enhancing surgical skill training …

Imitation LearningRepresentation LearningTrajectory Prediction

ETSM: Automating Dissection Trajectory Suggestion and Confidence Map-Based Safety Margin Prediction for Robot-assisted Endoscopic Submucosal Dissection

2024-11-28 · Mengya Xu, Wenjin Mo, Guankun Wang, Huxin Gao 외

Robot-assisted Endoscopic Submucosal Dissection (ESD) improves the surgical procedure by providing a more comprehensive view through advanced robotic instruments and bimanual operation, thereby enhancing dissection effic…

Decision MakingPredictionregressionTrajectory Prediction

Visual Word Sense Disambiguation with CLIP through Dual-Channel Text Prompting and Image Augmentations

2026-02-06 · Shamik Bhattacharya, Daniel Perkins, Yaren Dogan, Vineeth Konjeti 외 arxiv

Ambiguity poses persistent challenges in natural language understanding for large language models (LLMs). To better understand how lexical ambiguity can be resolved through the visual domain, we develop an interpretable …

Natural Language UnderstandingWord Sense DisambiguationImage Augmentation