paper-with-me

홈 › Papers

A Survey on Data Curation for Visual Contrastive Learning: Why Crafting Effective Positive and Negative Pairs Matters

2025-02-12 · Shasvat Desai, Debasmita Ghose, Deep Chakraborty

Visual contrastive learning aims to learn representations by contrasting similar (positive) and dissimilar (negative) pairs of data samples. The design of these pairs significantly impacts representation quality, training efficiency, and computational cost. A well-curated set of pairs leads to stronger representations and faster convergence. As contrastive pre-training sees wider adoption for solving downstream tasks, data curation becomes essential for optimizing its effectiveness. In this survey, we attempt to create a taxonomy of existing techniques for positive and negative pair curation in contrastive learning, and describe them in detail.

📄 PDF Abstract BibTeX arXiv:2502.08134

Code (0)

등록된 구현이 없습니다.

Tasks

Contrastive Learning

Methods 이 논문이 사용한 방법론

Contrastive Learning 설명 없음
SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Learning to Highlight Audio by Watching Movies

2025-05-17 · CVPR 2025 1 · Chao Huang, Ruohan Gao, J. M. F. Tsang, Jan Kurcius 외

Recent years have seen a significant increase in video content creation and consumption. Crafting engaging content requires the careful curation of both visual and audio elements. While visual cue curation, through techn…

Batch Curation for Unsupervised Contrastive Representation Learning

2021-08-19 · Michael C. Welle, Petra Poklukar, Danica Kragic

The state-of-the-art unsupervised contrastive visual representation learning methods that have emerged recently (SimCLR, MoCo, SwAV) all make use of data augmentations in order to construct a pretext task of instant disc…

Representation Learning

Quality Over Quantity? LLM-Based Curation for a Data-Efficient Audio-Video Foundation Model

2025-03-12 · Ali Vosoughi, Dimitra Emmanouilidou, Hannes Gamper

Integrating audio and visual data for training multimodal foundational models remains challenging. We present Audio-Video Vector Alignment (AVVA), which aligns audiovisual (AV) scene content beyond mere temporal synchron…

AudioCapsContrastive LearningLarge Language ModelRetrieval+1

Seizing the Means of Production: Exploring the Landscape of Crafting, Adapting and Navigating Generative AI Models in the Visual Arts

2024-04-26 · Ahmed M. Abuzuraiq, Philippe Pasquier

In this paper, we map out the landscape of options available to visual artists for creating personal artworks, including crafting, adapting and navigating deep generative models. Following that, we argue for revisiting m…

Active Data Curation Effectively Distills Large-Scale Multimodal Models

2024-11-27 · CVPR 2025 1 · Vishaal Udandarao, Nikhil Parthasarathy, Muhammad Ferjad Naeem, Talfan Evans 외

Knowledge distillation (KD) is the de facto standard for compressing large-scale models into smaller ones. Prior works have explored ever more complex KD strategies involving different objective functions, teacher-ensemb…

DecoderImage CaptioningKnowledge DistillationQuestion Answering+3