# Semantic Hardness Is Not Visual Hardness: Sign-Aware Hard Negative Mining for Sign Language Retrieval

arXiv: https://arxiv.org/abs/2607.09263 · HTML: https://mlsanity.com/paper/2607.09263/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 32 references that are in the map.
- First posted on arXiv 10 Jul 2026.

## Builds on (most cited first, 22 of 22)

- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [FastText.zip: Compressing text classification models](https://mlsanity.com/paper/1612.03651/) (2016)
- [Sign Language Transformers: Joint End-to-end Sign Language Recognition and Translation](https://mlsanity.com/paper/2003.13830/) (2020)
- [Advances in Pre-Training Distributed Word Representations](https://mlsanity.com/paper/1712.09405/) (2017)
- [Teaching CLIP to Count to Ten](https://mlsanity.com/paper/2302.12066/) (2023)
- [MS-ASL: A Large-Scale Data Set and Benchmark for Understanding American Sign Language](https://mlsanity.com/paper/1812.01053/) (2018)
- [BSL-1K: Scaling up co-articulated sign language recognition using mouthing cues](https://mlsanity.com/paper/2007.12131/) (2020)
- [Gloss-free Sign Language Translation: Improving from Visual-Language Pretraining](https://mlsanity.com/paper/2307.14768/) (2023)
- [Skeleton Aware Multi-modal Sign Language Recognition](https://mlsanity.com/paper/2103.08833/) (2021)
- [LLMs are Good Sign Language Translators](https://mlsanity.com/paper/2404.00925/) (2024)
- [Verbs in Action: Improving verb understanding in video-language models](https://mlsanity.com/paper/2304.06708/) (2023)
- [SignBERT: Pre-Training of Hand-Model-Aware Representation for Sign Language Recognition](https://mlsanity.com/paper/2110.05382/) (2021)
- [TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives](https://mlsanity.com/paper/2411.02545/) (2024)
- [Contrasting Intra-Modal and Ranking Cross-Modal Hard Negatives to Enhance Visio-Linguistic Compositional Understanding](https://mlsanity.com/paper/2306.08832/) (2023)
- [BEST: BERT Pre-Training for Sign Language Recognition with Coupling Tokenization](https://mlsanity.com/paper/2302.05075/) (2023)
- [Read and Attend: Temporal Localisation in Sign Language Videos](https://mlsanity.com/paper/2103.16481/) (2021)
- [CiCo: Domain-Aware Sign Language Retrieval via Cross-Lingual Contrastive Learning](https://mlsanity.com/paper/2303.12793/) (2023)
- [Sign Language Production: A Review](https://mlsanity.com/paper/2103.15910/) (2021)
- [Sign Language Video Retrieval with Free-Form Textual Queries](https://mlsanity.com/paper/2201.02495/) (2022)
- [Uncertainty-aware sign language video retrieval with probability distribution modeling](https://mlsanity.com/paper/2405.19689/) (2024)
- [SEDS: Semantically Enhanced Dual-Stream Encoder for Sign Language Retrieval](https://mlsanity.com/paper/2407.16394/) (2024)

## Abstract (arXiv)

Sign Language Retrieval (SLRet) enables efficient access to sign language content but remains fragile in fine-grained scenarios where visually similar signs must be distinguished. We show that this limitation does not stem from model capacity, but from ineffective hard negative supervision. Specifically, we formulate fine-grained retrieval failures as a negative distribution mismatch: semantically distinct yet visually confusable signs are rarely treated as hard negatives, while existing text-based mining strategies fail to capture such visual ambiguity. To address this issue, we propose Sign-Aware Hard Negative Mining (SAN), which constructs hard negatives based on visual confusability in the sign embedding space rather than linguistic similarity. Experiments on PHOENIX-2014T demonstrate that SAN substantially improves fine-grained retrieval performance while preserving coarse-grained accuracy, highlighting the importance of aligning negative supervision with visual ambiguity in sign language retrieval.