Cross-batch Memory For Embedding Learning

Abstract

Mining informative negative instances are of central importance to deep metric learning (DML), however this task is intrinsically limited by mini-batch training, where only a mini-batch of instances is accessible at each iteration. In this paper, we identify a "slow drift" phenomena by observing that the embedding features drift exceptionally slow even as the model parameters are updating throughout the training process. This suggests that the features of instances computed at preceding iterations can be used to considerably approximate their features extracted by the current model. We propose a cross-batch memory (XBM) mechanism that memorizes the embeddings of past iterations, allowing the model to collect sufficient hard negative pairs across multiple mini-batches - even over the whole dataset. Our XBM can be directly integrated into a general pair-based DML framework, where the XBM augmented DML can boost performance considerably. In particular, without bells and whistles, a simple

Stats

• citations191
• S2 citations—
• github stars316
• HF likes0
• heat score22.13
• arxiv keywang2019cross

Related papers

• Rethinking Deep Contrastive Learning With Embedding Memory (2021)0.00
• Adaptive Cross Batch Normalization For Metric Learning (2023)0.00
• Learning With Memory-based Virtual Classes For Deep Metric Learning (2021)9.92
• Learning Intra-batch Connections For Deep Metric Learning (2021)0.00
• Improve Multi-modal Embedding Learning Via Explicit Hard Negative Gradient Amplifying (2025)2.80
• Unime-v2: Mllm-as-a-judge For Universal Multimodal Embedding Learning (2025)0.00
• Loop: Looking For Optimal Hard Negative Embeddings For Deep Metric Learning (2021)8.82
• Llave: Large Language And Vision Embedding Models With Hardness-weighted Contrastive Learning (2025)3.58