Pedro Dal Bianco
Pedro Dal Bianco
Postdoctoral researcher in Deep Learning for Sign Language Translation

I'm a researcher in deep learning applied to Sign Language Translation (SLT). I did my PhD in Computer Science at III-LIDI, Facultad de Informática, UNLP, where I'm now a postdoctoral fellow working on multilingual SLT datasets and models.

My work focuses on developing, training and interpreting deep learning models for SLT, mainly pose-based ones, and on the data they need. I developed the LSA-T dataset, the first large-scale dataset for Argentinian Sign Language (LSA) video translation. This work has been published at NeurIPS and CVPR workshops, the Workshop on Sign Language Processing at IJCNLP-AACL, IBERAMIA, and in Applied Soft Computing and JCS&T.

Projects
LSA-T

LSA-T

The first continuous Argentinian Sign Language (LSA) dataset for sign language translation: about 22 hours of video from the CN Sordos YouTube channel, with the Spanish subtitles and the pose keypoints of each signer.

slt_datasets

Python library, on PyPI, that downloads and loads sign language translation datasets in several languages (RWTH-PHOENIX-2014T, LSA-T, How2Sign, LSFB-CONT, ISLTranslate, GSL) for training models in PyTorch. Used in the multilingual SLT paper at WSLP 2025.

posecraft

Python library, on PyPI, to load and transform pose keypoint sequences as PyTorch tensors: centering, normalizing distances, filling missing keypoints, resampling frames and animating the result.

Latest publications

Target-Side Paraphrase Augmentation for Sign Language Translation with Large Language ModelsWorkshop paper

Pedro Dal Bianco, Jean Paul Nunes Reinhold, Oscar Stanchi, Facundo Quiroga, Franco Ronchetti, Ulisses Brisolara Corrêa

GenSign Workshop @ CVPR 2026

A large language model (GPT-4o) writes paraphrases of the target sentences to augment sign language translation data: pre-train on the augmented corpus, then fine-tune on the original references. Evaluated on PHOENIX14T, GSL and LSA-T, with lexical metrics and an LLM-as-a-Judge semantic evaluation.

Traducción automática de lengua de señas en español utilizando Deep LearningPhD thesis

Pedro Dal Bianco

PhD thesis, Facultad de Informática, Universidad Nacional de La Plata, 2026

PhD thesis (in Spanish) on translating Argentinian Sign Language into written Spanish with little data: the LSA-T corpus, a compact gloss-free pose-based Transformer, multilingual training and LLM paraphrases of the targets, and a study of what the model's attention learns.

Multilingual Sign Language Translation with Unified Datasets and Pose-Based TransformersWorkshop paper

Pedro Dal Bianco, Oscar Stanchi, Facundo Quiroga, Franco Ronchetti

Workshop on Sign Language Processing (WSLP), IJCNLP-AACL 2025

One pose-based translation model trained jointly on German, Greek, Argentinian and Indian sign languages through a standardized data layer. Multilingual pre-training followed by a short language-specific fine-tuning beats the monolingual baselines on three of the four datasets.

All publications