# Deep Reversible Consistency Learning for Cross-modal Retrieval

arXiv: https://arxiv.org/abs/2501.05686 · HTML: https://mlsanity.com/paper/2501.05686/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 53 references that are in the map.
- First posted on arXiv 10 Jan 2025.

## Builds on (most cited first, 12 of 12)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [An Empirical Evaluation of doc2vec with Practical Insights into Document Embedding Generation](https://mlsanity.com/paper/1607.05368/) (2016)
- [Exposing and Mitigating Spurious Correlations for Cross-Modal Retrieval](https://mlsanity.com/paper/2304.03391/) (2023)
- [Generative Cross-Modal Retrieval: Memorizing Images in Multimodal Language Models for Retrieval and Beyond](https://mlsanity.com/paper/2402.10805/) (2024)
- [An Overview of Cross-media Retrieval: Concepts, Methodologies, Benchmarks and Challenges](https://mlsanity.com/paper/1704.02223/) (2017)
- [Your Negative May not Be True Negative: Boosting Image-Text Matching with False Negative Elimination](https://mlsanity.com/paper/2308.04380/) (2023)
- [CL2CM: Improving Cross-Lingual Cross-Modal Retrieval via Cross-Lingual Knowledge Transfer](https://mlsanity.com/paper/2312.08984/) (2023)
- [Unifying Two-Stream Encoders with Transformers for Cross-Modal Retrieval](https://mlsanity.com/paper/2308.04343/) (2023)
- [Variational Autoencoder with CCA for Audio-Visual Cross-Modal Retrieval](https://mlsanity.com/paper/2112.02601/) (2021)
- [HGAN: Hierarchical Graph Alignment Network for Image-Text Retrieval](https://mlsanity.com/paper/2212.08281/) (2022)

## Built on by (most cited first, 2 of 2)

- [From Generator to Embedder: Harnessing Innate Abilities of Multimodal LLMs via Building Zero-Shot Discriminative Embedding Model](https://mlsanity.com/paper/2508.00955/) (2025)
- [From Mapping to Composing: A Two-Stage Framework for Zero-shot Composed Image Retrieval](https://mlsanity.com/paper/2504.17990/) (2025)

## Abstract (arXiv)

Cross-modal retrieval (CMR) typically involves learning common representations to directly measure similarities between multimodal samples. Most existing CMR methods commonly assume multimodal samples in pairs and employ joint training to learn common representations, limiting the flexibility of CMR. Although some methods adopt independent training strategies for each modality to improve flexibility in CMR, they utilize the randomly initialized orthogonal matrices to guide representation learning, which is suboptimal since they assume inter-class samples are independent of each other, limiting the potential of semantic alignments between sample representations and ground-truth labels. To address these issues, we propose a novel method termed Deep Reversible Consistency Learning (DRCL) for cross-modal retrieval. DRCL includes two core modules, ie Selective Prior Learning (SPL) and Reversible Semantic Consistency learning (RSC). More specifically, SPL first learns a transformation weight matrix on each modality and selects the best one based on the quality score as the Prior, which greatly avoids blind selection of priors learned from low-quality modalities. Then, RSC employs a Modality-invariant Representation Recasting mechanism (MRR) to recast the potential modality-invariant representations from sample semantic labels by the generalized inverse matrix of the prior. Since labels are devoid of modal-specific information, we utilize the recast features to guide the representation learning, thus maintaining semantic consistency to the fullest extent possible. In addition, a feature augmentation mechanism (FA) is introduced in RSC to encourage the model to learn over a wider data distribution for diversity. Finally, extensive experiments conducted on five widely used datasets and comparisons with 15 state-of-the-art baselines demonstrate the effectiveness and superiority of our DRCL.