# Differentially Private Motif-Preserving Multi-modal Hashing

arXiv: https://arxiv.org/abs/2605.15460 · HTML: https://mlsanity.com/paper/2605.15460/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 38 references that are in the map.
- First posted on arXiv 14 May 2026.

## Builds on (most cited first, 13 of 13)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [Deep Learning with Differential Privacy](https://mlsanity.com/paper/1607.00133/) (2016)
- [Graph Neural Networks for Graphs with Heterophily: A Survey](https://mlsanity.com/paper/2202.07082/) (2022)
- [HashNet: Deep Learning to Hash by Continuation](https://mlsanity.com/paper/1702.00758/) (2017)
- [A Survey on Learning to Hash](https://mlsanity.com/paper/1606.00185/) (2016)
- [Stealing Links from Graph Neural Networks](https://mlsanity.com/paper/2005.02131/) (2020)
- [Higher-order organization of complex networks](https://mlsanity.com/paper/1612.08447/) (2016)
- [LinkTeller: Recovering Private Edges from Graph Neural Networks via Influence Analysis](https://mlsanity.com/paper/2108.06504/) (2021)
- [A Comprehensive Survey on Cross-modal Retrieval](https://mlsanity.com/paper/1607.06215/) (2016)
- [Deep Cross-Modal Hashing](https://mlsanity.com/paper/1602.02255/) (2016)
- [Recursive Mechanism: Towards Node Differential Privacy and Unrestricted Joins [Full Version, Draft 0.1]](https://mlsanity.com/paper/1304.4795/) (2013)
- [Discrete Latent Factor Model for Cross-Modal Hashing](https://mlsanity.com/paper/1707.08322/) (2017)

## Abstract (arXiv)

Cross-modal hashing enables efficient retrieval by encoding images and text into compact binary codes. State-of-the-art methods rely on semantic similarity graphs derived from user interactions for supervision, yet these graphs encode sensitive behavioral patterns vulnerable to link reconstruction attacks. Existing privacy-preserving approaches fail on graph-structured data: Differentially Private SGD destroys relational motifs by treating samples independently, while graph synthesis methods suffer from unbounded local sensitivity in scale-free networks, hub nodes cause single-edge modifications to alter triangle counts by O(N), necessitating prohibitive noise injection. We term this phenomenon Hubness Explosion. We propose DMP-MH, a Sanitize-then-Distill framework that decouples privacy from representation learning. Our approach first bounds sensitivity by deterministically clipping node degrees, capping the L2-sensitivity of triangle motifs independently of dataset size. A sanitized synthetic graph is then generated via Noisy Mirror Descent under (ε,δ)-Edge Differential Privacy. Finally, dual-stream hashing networks distill this topology using a holistic structural loss that enforces cross-modal alignment. Evaluated on MIRFlickr-25K and NUS-WIDE under a strict inductive protocol, DMP-MH outperforms private baselines by up to 11.4 mAP points while retaining up to 92.5% of non-private performance.