Exploring Pose-based Sign Language Translation: Ablation Studies and Attention Insights

• URL: http://arxiv.org/abs/2507.01532v1
• Date: Wed, 02 Jul 2025 09:36:26 GMT
• Title: Exploring Pose-based Sign Language Translation: Ablation Studies and Attention Insights
• Authors: Tomas Zelezny, Jakub Straka, Vaclav Javorek, Ondrej Valach, Marek Hruz, Ivan Gruber,
• Abstract summary: Sign Language Translation (SLT) has evolved significantly, moving from isolated recognition approaches to complex, continuous gloss-free translation systems.<n>This paper explores the impact of pose-based data preprocessing techniques on SLT performance.<n>We employ a transformer-based architecture, adapting a modified T5 encoder-decoder model to process pose representations.
• Score: 0.5277756703318045
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Sign Language Translation (SLT) has evolved significantly, moving from isolated recognition approaches to complex, continuous gloss-free translation systems. This paper explores the impact of pose-based data preprocessing techniques - normalization, interpolation, and augmentation - on SLT performance. We employ a transformer-based architecture, adapting a modified T5 encoder-decoder model to process pose representations. Through extensive ablation studies on YouTubeASL and How2Sign datasets, we analyze how different preprocessing strategies affect translation accuracy. Our results demonstrate that appropriate normalization, interpolation, and augmentation techniques can significantly improve model robustness and generalization abilities. Additionally, we provide a deep analysis of the model's attentions and reveal interesting behavior suggesting that adding a dedicated register token can improve overall model performance. We publish our code on our GitHub repository, including the preprocessed YouTubeASL data.

Related papers

• Dynamic Chunking for End-to-End Hierarchical Sequence Modeling [17.277753030570263]
  We introduce techniques that enable a dynamic chunking mechanism which automatically learns content- and context- dependent segmentation strategies.<n> incorporating this into an explicit hierarchical network (H-Net) allows replacing the (implicitly hierarchical) tokenization-LM-detokenization pipeline with a single model learned fully end-to-end.<n>Iterating the hierarchy to multiple stages further increases its performance by modeling multiple levels of abstraction.<n>H-Nets pretrained on English show significantly increased character-level robustness, and qualitatively learn meaningful data-dependent chunking strategies without anys or explicit supervision.
  arXiv  Detail & Related papers  (2025-07-10T17:39:37Z)
• Contextually Guided Transformers via Low-Rank Adaptation [14.702057924366345]
  Large Language Models (LLMs) based on Transformers excel at text processing, but their reliance on prompts for specialized behavior introduces computational overhead.<n>We propose a modification to a Transformer architecture that eliminates the need for explicit prompts by learning to encode context into the model's weights.
  arXiv  Detail & Related papers  (2025-06-06T01:34:39Z)
• SignAttention: On the Interpretability of Transformer Models for Sign Language Translation [2.079808290618441]
  This paper presents the first comprehensive interpretability analysis of a Transformer-based Sign Language Translation model. We examine the attention mechanisms within the model to understand how it processes and aligns visual input with sequential glosses. This work contributes to a deeper understanding of SLT models, paving the way for the development of more transparent and reliable translation systems.
  arXiv  Detail & Related papers  (2024-10-18T14:38:37Z)
• Fine-Grained Verifiers: Preference Modeling as Next-token Prediction in Vision-Language Alignment [57.0121616203175]
  We propose FiSAO, a novel self-alignment method that utilizes the model's own visual encoder as a fine-grained verifier to improve vision-language alignment.<n>By leveraging token-level feedback from the vision encoder, FiSAO significantly improves vision-language alignment, even surpassing traditional preference tuning methods that require additional data.
  arXiv  Detail & Related papers  (2024-10-18T03:34:32Z)
• DETAIL: Task DEmonsTration Attribution for Interpretable In-context Learning [75.68193159293425]
  In-context learning (ICL) allows transformer-based language models to learn a specific task with a few "task demonstrations" without updating their parameters.<n>We propose an influence function-based attribution technique, DETAIL, that addresses the specific characteristics of ICL.<n>We experimentally prove the wide applicability of DETAIL by showing our attribution scores obtained on white-box models are transferable to black-box models in improving model performance.
  arXiv  Detail & Related papers  (2024-05-22T15:52:52Z)
• End-to-End Lip Reading in Romanian with Cross-Lingual Domain Adaptation and Lateral Inhibition [2.839471733237535]
  We analyze several architectures and optimizations on the underrepresented, short-scale Romanian language dataset called Wild LRRo. We obtain state-of-the-art results using our proposed method, namely cross-lingual domain adaptation and unlabeled videos. We also assess the performance of adding a layer inspired by the neural inhibition mechanism.
  arXiv  Detail & Related papers  (2023-10-07T15:36:58Z)
• Interpretable Sentence Representation with Variational Autoencoders and Attention [0.685316573653194]
  We develop methods to enhance the interpretability of recent representation learning techniques in natural language processing (NLP) We leverage Variational Autoencoders (VAEs) due to their efficiency in relating observations to latent generative factors. We build two models with inductive bias to separate information in latent representations into understandable concepts without annotated data.
  arXiv  Detail & Related papers  (2023-05-04T13:16:15Z)
• Multilingual Extraction and Categorization of Lexical Collocations with Graph-aware Transformers [86.64972552583941]
  We put forward a sequence tagging BERT-based model enhanced with a graph-aware transformer architecture, which we evaluate on the task of collocation recognition in context. Our results suggest that explicitly encoding syntactic dependencies in the model architecture is helpful, and provide insights on differences in collocation typification in English, Spanish and French.
  arXiv  Detail & Related papers  (2022-05-23T16:47:37Z)
• Modeling Intensification for Sign Language Generation: A Computational Approach [13.57903290481737]
  End-to-end sign language generation models do not accurately represent the prosody in sign language. We aim to improve the prosody in generated sign languages by modeling intensification in a data-driven manner. We find that our efforts in intensification modeling yield better results when evaluated with automatic metrics.
  arXiv  Detail & Related papers  (2022-03-18T01:13:21Z)
• Examining Scaling and Transfer of Language Model Architectures for Machine Translation [51.69212730675345]
  Language models (LMs) process sequences in a single stack of layers, and encoder-decoder models (EncDec) utilize separate layer stacks for input and output processing. In machine translation, EncDec has long been the favoured approach, but with few studies investigating the performance of LMs.
  arXiv  Detail & Related papers  (2022-02-01T16:20:15Z)
• Unsupervised Paraphrasing with Pretrained Language Models [85.03373221588707]
  We propose a training pipeline that enables pre-trained language models to generate high-quality paraphrases in an unsupervised setting. Our recipe consists of task-adaptation, self-supervision, and a novel decoding algorithm named Dynamic Blocking. We show with automatic and human evaluations that our approach achieves state-of-the-art performance on both the Quora Question Pair and the ParaNMT datasets.
  arXiv  Detail & Related papers  (2020-10-24T11:55:28Z)
• Learning Source Phrase Representations for Neural Machine Translation [65.94387047871648]
  We propose an attentive phrase representation generation mechanism which is able to generate phrase representations from corresponding token representations. In our experiments, we obtain significant improvements on the WMT 14 English-German and English-French tasks on top of the strong Transformer baseline.
  arXiv  Detail & Related papers  (2020-06-25T13:43:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.