Universidade da Coruña · CITIC · A Coruña, Spain
We study how people look for information, and how to help them find it.
We are the Information Retrieval Lab at the Universidade da Coruña. Our work starts with search: how to find the right information, how to rank it, and how to tell whether a system got it right. These days we apply it mostly to health: noticing early signs of depression in what people write online, answering health questions with sources you can check, and asking AI systems to explain how they reached an answer.
What we work on
All research linesEach line lists the people working on it, its projects and its recent papers.
-
Early risk detection and digital mental health
Models, test collections and LLM reasoning to spot early signs of depression, ADHD and other risks in what people write online.
“early signs of depression in social media” -
Social network analysis: hate speech, bias and online harms
What people post on social networks: hate speech, conspiracy and partisan narratives, and the biases of the models that read them.
“hate speech detection in Galician and Spanish” -
Search evaluation and LLMs as assessors
Pooling, statistical testing and LLM-assisted judging that keep search evaluation reliable at lower cost.
“can LLMs replace human relevance judgments” -
Search experience, RAG and health information access
Query suggestion, keyphrase search and retrieval-augmented generation that stays faithful to the evidence, applied to health.
“trustworthy answers to health questions” -
Knowledge representation, explainable reasoning and clinical AI
Logic for AI that can explain its conclusions, and interpretable models that help hospital pharmacists choose drug doses after a transplant.
“answer set programming with temporal logic” -
Bio-inspired computing and evolutionary learning
Evolutionary algorithms and genetic programming that search huge spaces for solutions people can read, from protein structures to business failure prediction.
“genetic programming interpretable prediction models”
Why we do what we do
-
Mental health
People often write about how they feel long before they ask for help.
We build tools that notice early signs of depression and other difficulties in what people write online, so that support can arrive sooner, and we do it with care for their privacy.
How we work on it -
Health information
A health question deserves an answer you can check.
Search engines and chatbots now answer questions directly. We work on systems that show where each part of an answer comes from, so you can judge it for yourself.
How we work on it -
Online conversations
Conversations online can turn hostile, and not everyone gets hit the same.
We study how hate speech and misleading stories spread, and we build ways to detect them that treat every community fairly.
How we work on it -
AI you can follow
When a machine helps make a decision, it should be able to explain it.
From adjusting a medicine dose after a transplant to judging which search results are useful, we work on AI whose reasoning people can follow and question.
How we work on it
News
All news-
The ROI of small examples, accepted at EACL 2027
Jorge Gabín, Anxo Pérez and Javier Parapar study in-context learning under a fixed token budget. Many short examples beat a few long ones, and filling the budget with concise demonstrations outperforms longer demonstrations and random sampling. The paper will be presented in Athens in March 2027.
-
ESSIR 2027 comes to A Coruña
The 18th European Summer School on Information Retrieval will be held at the Faculty of Computer Science, 12 to 16 July 2027, organised by the IRLab.
-
Masterclass for psychologists on AI and early risk detection
Anxo Pérez taught "AI in Psychology and Mental Health, from the prompt to early risk detection" in the series run by the Galician and Portuguese psychology associations, online for up to 500 professionals.
-
Talk on AI and mental health for Healthy October
Eliseo Bao and Anxo Pérez gave the talk "Artificial Intelligence and Mental Health: What Can Social Media Tell Us?" at the Faculty of Computer Science, part of the 12th Healthy October programme at the Universidade da Coruña.
Recent papers
All publications- Conference
- Conference
- Conference
- Conference
- Journal
Strong Equivalence in Answer Set Programming with Constraints
Theory and Practice of Logic Programming · 2026
The team
EveryoneLecturers and researchers in computer science and mathematics, PhD students and research staff.
- Álvaro Barreiro
- Pedro Cabalar
- José Santos
- Javier Parapar
- Concepción Vidal
- Felicidad Aguado
- Gilberto Pérez
- Anxo Pérez
- David Otero
- Brais Muñiz
- Óscar Iglesias
- M. Eduardo Ares
- Jorge Gabín
- Paloma Piot
- Alfonso Landín
- Eliseo Bao
- Laura Hermo
- Juan Luis Filgueiras
- Juan Manuel Rey
- Diego Martínez
- Ovidio Manteiga
- Manuel Couto
- Pablo Fernández
- Andrés Lires
- Patricia Martín Rodilla
- David E. Losada
- César Piñeiro
- Paula López Otero
Find us
Lab 4.2 in the Facultade de Informática, Campus de Elviña, A Coruña. Most members are also part of CITIC, the research centre next door.
Open map