Iñigo Alonso

Computational Linguistics Researcher

14 years of combined experience in academia and industry, specializing in Table Understanding and Visually-Situated Language, building large-scale multimodal applications.

University of Edinburgh
GitHub Scholar Bluesky

Work Experience

Research Associate
Jan. 2025 - Present
EdinburghNLP, University of Edinburgh

Conducting research on multimodal approaches to visually-situated text tasks, including table, document, and chart understanding.

  • Multimodal Table Understanding
  • Visually-Situated Language
  • Optical Context Compression
Research Assistant
Jan. 2023 - Dec. 2024
HiTZ Center, University of the Basque Country UPV/EHU

Contributed to Luminous (European project for multimodal dialog models in mixed reality) and Antidote (medical question answering through RAG).

  • Cross-modal Entity Linking Research
  • Fine-tuned LLMs applied to medical question answering
  • Retrieval-Augmented Generation (RAG)
Machine Learning Engineer
May 2018 - Sep. 2021
Sherpa.ai

Fine-tuned and deployed Deep Neural Network-based models for text summarisation and sentence slot filling for commercial applications.

  • Deep Learning applied to Natural Language Processing
  • Abstractive summarisation and Data-to-Text
  • Developing and deploying production ready ML models on AWS infrastructure
Research Assistant
Jan. 2017 - Sep. 2017
Deustotech Computing

Applied data science research methods to business challenges including anomaly detection and customer segmentation.

  • Machine Learning applied to retail customer segmentation
  • Assistance on Deep Learning research projects
Research Intern
Feb. 2012 - Jan. 2015
Deustotech Computing

Worked on research projects in Information Security, IoT, and Machine Learning, developing foundational data science and research skills.


Publications

TABLET: A Large-Scale Dataset for Robust Visual Table Understanding
Iñigo Alonso, Imanol Miranda, Eneko Agirre, Mirella Lapata
ICLR 2025
In review…
MATE: A Cross-modal Entity Linking Benchmark for Vision Language Models
Iñigo Alonso, Ander Salaberria, Gorka Azkune, Oier Lopez de Lacalle
ACL 2025
Accepted at Findings
PixT3: Pixel-based Table-To-Text Generation
Iñigo Alonso, Eneko Agirre, Mirella Lapata
ACL 2024
Accepted at Main Conference
MedExpQA: Multilingual Benchmarking of Large Language Models for Medical Question Answering
Iñigo Alonso, Maite Oronoz, Rodrigo Agerri
Artificial Intelligence in Medicine
Accepted • 2024
Automatic Logical Forms improve fidelity in Table-to-Text generation
Iñigo Alonso, Eneko Agirre
Expert Systems with Applications
Accepted • 2023