Noisy Correspondence Learning With Self-reinforcing Errors Mitigation

Abstract

Cross-modal retrieval relies on well-matched large-scale datasets that are laborious in practice. Recently, to alleviate expensive data collection, co-occurring pairs from the Internet are automatically harvested for training. However, it inevitably includes mismatched pairs, \ie, noisy correspondences, undermining supervision reliability and degrading performance. Current methods leverage deep neural networks' memorization effect to address noisy correspondences, which overconfidently focus on *similarity-guided training with hard negatives* and suffer from self-reinforcing errors. In light of above, we introduce a novel noisy correspondence learning framework, namely \textbf\{S\}elf-\textbf\{R\}einforcing \textbf\{E\}rrors \textbf\{M\}itigation (SREM). Specifically, by viewing sample matching as classification tasks within the batch, we generate classification logits for the given sample. Instead of a single similarity score, we refine sample filtration through energy uncertainty and

Stats

• citations11
• S2 citations—
• github stars0
• HF likes0
• heat score8.09
• arxiv keydang2023noisy

Related papers

• PCSR: Pseudo-label Consistency-guided Sample Refinement For Noisy Correspondence Learning (2025)0.00
• Noisy Correspondence Learning With Meta Similarity Correction (2023)11.67
• Disentangled Noisy Correspondence Learning (2024)3.58
• Learning To Rematch Mismatched Pairs For Robust Cross-modal Retrieval (2024)13.82
• Neighbor-aware Instance Refining With Noisy Labels For Cross-modal Retrieval (2025)2.26
• Robust Remote Sensing Image-text Retrieval With Noisy Correspondence (2026)1.24
• Noisy Self-training With Synthetic Queries For Dense Retrieval (2023)0.00
• Robust Self-paced Hashing For Cross-modal Retrieval With Noisy Labels (2025)7.81