Fine-tuning T5 on SCAN
↗Studied compositional generalization by fine-tuning T5-small on SCAN tasks with custom evaluation.
PyTorchTransformersT5NLPSequence Learning
Jesús Rabanal Álvarez
NLP · Linguistics & Attention · Researcher
I work at the intersection of linguistics, natural language processing, and software development. My interests include neural language models, natural cognitive processes, and practical ML systems for language data.
Currently
- MSc-level coursework in NLP, ML and Cognitive Science
- Research-oriented projects involving Linguistics, Cognitive Science, and Computational Modeling.
- Building and evaluating language models for low-resource and structured tasks.
Selected work
Multi-author Writing Style Analysis
↗Built a BERT-based approach to detect authorship changes across paragraph boundaries using ground-truth JSON annotations.
PythonBERTNLPData Analysis
Master Thesis: Ambiguous attachment through the Visual World Paradigm: An experimental study of Danish learners of Spanish
↗Studied Spanish, Danish, and Danish L2 Spanish anaphora resolution preferences using a visual-world eye-tracking paradigm.
Eye-trackingVisual World ParadigmPsycholinguisticsExperimental DesignData AnalysisAnaphora Resolution