Research map / Vision-language and multimodal / Human and embodied modeling
Discovering Clone Negatives via Adaptive Contrastive Learning for Image-Text Matching
Renjie Pan, Jihao Dong, Hua Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Identifies "clone negatives" in image-text matching and proposes Adaptive Contrastive Learning (AdaCL) with adaptive margins and a modulating anchor to separate them from positives.
Problem
Image-text matching suffers from clone negatives: negative image-text pairs that semantically resemble positive pairs, causing ambiguous and sub-optimal matching.
Approach
Introduces two margin parameters plus a modulating anchor in a contrastive loss to dynamically strengthen compactness between positives and enlarge separation from clone negatives, rather than treating all negatives uniformly.
Closest papers by meaning
- Expanding the Scope of Negatives: Boosting Image-Text Matching with Negatives Distribution Guided Learning (AAAI 2025 · similarity 0.91)
- TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives (NeurIPS 2024 · similarity 0.86)
- Aligning Composed Query with Image via Discriminative Perception from Negative Correspondences (AAAI 2025 · similarity 0.85)
- SuperCLIP: CLIP with Simple Classification Supervision (NeurIPS 2025 · similarity 0.85)
- Text as Partial Constraint: Core–Residual Alignment for Robust Vision–Language Learning (NeurIPS 2026 · similarity 0.85)
- BiMAC: Bidirectional Multimodal Alignment in Contrastive Learning (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
