2024/10/03 by Nikos Giakoumoglou, Nikolaos Giakoumoglou, Giakoumoglou, Nikolaos +2 · 1 citation
Computer Science · #Advanced Image and Video Retrieval Techniques #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Face recognition and analysis #I.2 #I.4 #cs.AI #cs.CV #msc:I.2 #msc:I.4
paper · pdf · doi:10.48550/arxiv.2410.02401
Preprint. Code: https://github.com/giakoumoglou/synco, Supplementary: https://giakoumoglou.com/src/synco_suppl.pdf
openalex publication_date 2024/10/03 · openalex created_date 2024/11/01 · arxiv created 2025/02/17 · arxiv updated 2026/07/30 · openalex updated_date 2026/08/01
Contrastive learning has become a dominant approach in self-supervised visual representation learning, but efficiently leveraging hard negatives, which are samples closely resembling the anchor, remains challenging. We introduce SynCo (Synthetic negatives in Contrastive learning), a novel approach that improves model performance by generating synthetic hard negatives on the representation space. Building on the MoCo framework, SynCo introduces six strategies for creating diverse synthetic hard negatives on-the-fly with minimal computational overhead. SynCo achieves faster training and strong representation learning, surpassing MoCo-v2 by +0.4% and MoCHI by +1.0% on ImageNet ILSVRC-2012 linear evaluation. It also transfers more effectively to detection tasks achieving strong results on PASCAL VOC detection (57.2% AP) and significantly improving over MoCo-v2 on COCO detection (+1.0% AP) and instance segmentation (+0.8% AP). Our synthetic hard negative generation approach significantly enhances visual representations learned through self-supervised contrastive learning.