# SignDPO: Multi-level Direct Preference Optimisation for Skeleton-based Gloss-free Sign Language Translation

arXiv: https://arxiv.org/abs/2604.18034 · HTML: https://mlsanity.com/paper/2604.18034/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Diffusion and multimodal model alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 28 of its 39 references that are in the map.
- First posted on arXiv 20 Apr 2026.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Diffusion and multimodal model alignment](https://mlsanity.com/alignment/diffusion-and-multimodal-model-alignment/)

## Builds on (most cited first, 28 of 28)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition](https://mlsanity.com/paper/1801.07455/) (2018)
- [A General Theoretical Paradigm to Understand Learning from Human Preferences](https://mlsanity.com/paper/2310.12036/) (2023)
- [Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models](https://mlsanity.com/paper/2401.01335/) (2024)
- [Sign Language Transformers: Joint End-to-end Sign Language Recognition and Translation](https://mlsanity.com/paper/2003.13830/) (2020)
- [Aligning Modalities in Vision Large Language Models via Preference Fine-tuning](https://mlsanity.com/paper/2402.11411/) (2024)
- [Improving Sign Language Translation with Monolingual Data by Sign Back-Translation](https://mlsanity.com/paper/2105.12397/) (2021)
- [Silkie: Preference Distillation for Large Visual Language Models](https://mlsanity.com/paper/2312.10665/) (2023)
- [Two-Stream Network for Sign Language Recognition and Translation](https://mlsanity.com/paper/2211.01367/) (2022)
- [Gloss-free Sign Language Translation: Improving from Visual-Language Pretraining](https://mlsanity.com/paper/2307.14768/) (2023)
- [A Simple Multi-Modality Transfer Learning Baseline for Sign Language Translation](https://mlsanity.com/paper/2203.04287/) (2022)
- [Open-Domain Sign Language Translation Learned from Online Video](https://mlsanity.com/paper/2205.12870/) (2022)
- [SignBERT+: Hand-model-aware Self-supervised Pre-training for Sign Language Understanding](https://mlsanity.com/paper/2305.04868/) (2023)
- [Sign2GPT: Leveraging Large Language Models for Gloss-Free Sign Language Translation](https://mlsanity.com/paper/2405.04164/) (2024)
- [Uni-Sign: Toward Unified Sign Language Understanding at Scale](https://mlsanity.com/paper/2501.15187/) (2025)
- [Gloss Attention for Gloss-free Sign Language Translation](https://mlsanity.com/paper/2307.07361/) (2023)
- [Gloss-Free End-to-End Sign Language Translation](https://mlsanity.com/paper/2305.12876/) (2023)
- [Improving Gloss-free Sign Language Translation by Reducing Representation Density](https://mlsanity.com/paper/2405.14312/) (2024)
- [C2RL: Content and Context Representation Learning for Gloss-free Sign Language Translation and Retrieval](https://mlsanity.com/paper/2408.09949/) (2024)
- [Conditional Variational Autoencoder for Sign Language Translation with Cross-Modal Alignment](https://mlsanity.com/paper/2312.15645/) (2023)
- [Scaling up Multimodal Pre-training for Sign Language Understanding](https://mlsanity.com/paper/2408.08544/) (2024)
- [Focused-DPO: Enhancing Code Generation Through Focused Preference Optimization on Error-Prone Points](https://mlsanity.com/paper/2502.11475/) (2025)
- [A Token-level Contrastive Framework for Sign Language Translation](https://mlsanity.com/paper/2204.04916/) (2022)
- [RVLF: A Reinforcing Vision-Language Framework for Gloss-Free Sign Language Translation](https://mlsanity.com/paper/2512.07273/) (2025)

## Abstract (arXiv)

We present SignDPO, a novel multi-level Direct Preference Optimisation (DPO) framework designed to enhance the alignment of skeleton-based Sign Language Translation. While current skeleton-based models have made significant progress using Maximum Likelihood Estimation, they are primarily constrained by an imitation-based paradigm that lacks discriminative sensitivity to the fine-grained spatio-temporal nuances of sign language, often leading to semantic drift. To address this, SignDPO shifts the optimisation goal from simple sequence mimicry to structured preference alignment across spatial, temporal, and linguistic dimensions. Our framework involves three key designs. First, we introduce a hierarchical perturbation strategy to construct spatial and temporal non-preferred samples at both global and local granularities automatically. Second, we propose a self-guiding mechanism that leverages decoder cross-attention scores to identify and perturb semantically salient skeletal regions, forcing the model to distinguish genuine sign signals from structural distortions. Third, we establish an automated language-level preference generator by fine-tuning a dedicated perturbation model, capturing complex output-level failure modes without manual annotation. Extensive experiments on three widely adopted benchmarks, CSL-Daily, How2Sign, and OpenASL, demonstrate that SignDPO consistently outperforms state-of-the-art gloss-free methods and even rivals established gloss-based ones. Our results suggest that multi-level preference alignment is a powerful paradigm for bridging the gap between high-entropy skeletal trajectories and discrete linguistic semantics.