cv
Contact Information
| Name | Eliseo Bao |
| eliseo.bao@udc.es | |
| Website | https://eliseobao.github.io |
Research Interests
- NLP for Mental Health — detecting and explaining depression-related symptoms in social media using clinical frameworks such as the DSM-5 and BDI-II.
- Explainable AI — developing interpretable methods and clinically grounded explanations for mental-health assessment.
- Conversational AI — building conversational agents that support clinicians in reviewing and assessing mental-health information.
- Persona Simulation and Evaluation — evaluating how faithfully LLM-based personas simulate real users for mental-health research.
- Multilingual NLP — adapting language models to underrepresented languages, including Galician.
Education
-
2024 - present Galicia, Spain
Ph.D.
University of A Coruña
Computer Science
-
2024 Galicia, Spain
M.Sc.
University of A Coruña
Computer Science
- Extraordinary Master’s Award, 2023/2024 cohort.
- Master’s thesis awarded highest honors (Matrícula de Honor).
-
2022 Galicia, Spain
B.Sc.
University of A Coruña
Computer Science
- Bachelor’s thesis awarded highest honors (Matrícula de Honor), 9.5/10.
-
2018 Galicia, Spain
Tech.
IES San Clemente
Application Development
Experience
-
2024 - present Predoctoral Fellow
Universidade da Coruña
- Xunta de Galicia Predoctoral Grant ED481A-2024-079.
-
2025 - 2025 Sheffield, UK
Visiting Researcher
Natural Language Processing Group, University of Sheffield
- Hosted by Dr. Xi Wang; work on LLM persona simulation led to the Eval4Sim paper (CIKM 2026).
-
2023 - 2024 Research Assistant
CITIC Centro de Investigacion TIC - Universidade da Coruña
- Working at the Information Retrieval Lab (IRLab) within the project PLEC2021-007662 Big-eRisk Early Prediction of Personal Risks on Massive Data funded by MCIN/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR.
-
2022 - 2023 Internship
ITG Instituto Tecnológico de Galicia
- Contributed to the CEL.IA group, developing NLP solutions, with a focus on the ASR research line.
-
2018 - 2018 Apprenticeship
NATICA, SISTEMAS DE INFORMACION SL
Publications
-
2026 Eval4Sim: An Evaluation Framework for Persona Simulation
Proc. of the 35th ACM Int'l Conference on Information and Knowledge Management (CIKM 2026), Rome, Italy. Forthcoming
Introduces a three-dimensional framework (adherence, consistency, naturalness) for evaluating how well LLM personas simulate human conversation.
-
2026 Improving clinical reliability of LLM reasoning for depression assessment via structured generation and GRPO
Science Progress, 109(3)
Shows GRPO-trained LLMs can explain depression-screening decisions through DSM-5-aligned reasoning without sacrificing classification accuracy.
-
2026 Learning Evidence of Depression Symptoms via Prompt Induction
Proc. of the 49th International ACM SIGIR Conference (SIGIR 2026), Melbourne, Australia
Proposes Symptom Induction, compressing labeled examples into interpretable guidelines that improve LLM classification of depressive symptoms.
-
2025 How does depression talk on social media? Modeling depression language with relevance-based statistical language models
Online Social Networks and Media, 50, 100339
Builds relevance-based statistical language models of depression-related language to rank users by risk across Reddit and Twitter.
-
2025 ReDSM5: A Reddit Dataset for DSM-5 Depression Detection
Proc. of the 34th ACM Int'l Conference on Information and Knowledge Management (CIKM 2025), Seoul, Republic of Korea
A 1,484-post Reddit dataset with sentence-level, clinician-annotated DSM-5 depressive-symptom labels and rationales.
-
2025 Towards Explainable and Safe Systems for Health Data
SIGIR 2025 Doctoral Consortium, Padua, Italy
Doctoral consortium paper outlining a research plan for transparent, explainable, and safe health information systems.
-
2025 MindWell: A Conversational Agent for Professional Depression Screening on Social Media
Advances in Information Retrieval — ECIR 2025, LNCS, Part V, pp. 47-52, Lucca, Italy
Demonstrates an open-source RAG chatbot that helps clinicians review social media history for BDI-II depressive symptoms.
-
2024 Explainable depression symptom detection in social media
Health Information Science and Systems, 12(1), 47
Transformer-based models that both detect and explain depressive symptom markers in social media text.
-
2024 Adapting Large Language Models for Underrepresented Languages
VII Congreso XoveTIC, Universidade da Coruña — CICAS-Sngular Award
Introduces a Galician-adapted Llama 3.1 model and instruction dataset, outperforming the base model and another Galician model.
-
2024 MindWell: an open-source chatbot to assist in the detection and monitoring of depressive disorders
M.Sc. Thesis, Universidade da Coruña
Designed and evaluated the MindWell chatbot with a professional psychologist.
-
2022 Ranking of Reddit users using Relevance Models for depressive disorders
B.Sc. Thesis, Universidade da Coruña
Built depression-risk vocabularies and ranking methods from Reddit using relevance-based language models.
Honors, Awards and Scholarships
-
2025 Extraordinary Master's Award
Universidade da Coruña
M.Sc. in Computer Engineering, 2023/2024 cohort.
-
2024 CICAS-Sngular Award
VII XoveTIC Congress
Awarded for the best work in line with best practices for Open Science, Open Data, Open Access and Open Source. Certification · Publication
-
2023 B.Sc. thesis category award
1st edition of the "Tell me about your research" competition, University of A Coruña
-
2022 -
2021 Departmental Collaboration Scholarship
Universidade da Coruña
Department of Computer Science and Information Technologies, November 2020–May 2021.
Talks
- 28 October 2026 — “Artificial Intelligence and Mental Health: What Can Social Media Tell Us?”, with Anxo Pérez Vila, 12th Healthy October programme, Faculty of Computer Science, Universidade da Coruña.
- 19 May 2026 — “Explainable AI for Depression Symptom Detection on Social Media,” invited talk, International University of La Rioja (UNIR).
- 13 February 2026 — “Explainable depression symptom detection in social media,” 2nd Symposium on Mental Health and Artificial Intelligence, Université de Caen Normandie, France.
Program Committees & Reviewing
- 2027 (upcoming) — Program Committee, ECIR 2027 (Resource Papers track).
- 2027 (upcoming) — Reviewer, WWW 2027 (full papers).
- 2026 — Program Committee, CIKM 2026 (Resources track; Short Papers track).
- 2026 — Program Committee, CLEF 2026 (eRisk Lab).
- 2026 — Reviewer, SIGIR 2026 (full papers).
- 2026 — Reviewer, Journal of Intelligent & Fuzzy Systems.
- 2026 — Reviewer, Language Resources and Evaluation (Springer).
- 2026 — Reviewer, PLOS Digital Health.
- 2025 — Program Committee, CIKM 2025 (Resource Papers track).
- 2025 — Program Committee, CLEF 2025 (eRisk: Early Risk Prediction on the Internet).
- 2025 — Reviewer, SIGIR 2025 (subreviewer / external reviewer).
- 2025 — Reviewer, Digital Health (SAGE).
- 2025 — Reviewer, Health Information Science and Systems (Springer).
- 2025 — Reviewer, IEEE Transactions on Affective Computing.
- 2024 — Reviewer, Health Information Science and Systems (Springer).
Supervision and Mentoring
- 2026 — Co-supervised 2 bachelor’s theses and 1 master’s thesis.
- 21–22 October 2026 — Mentor, Datathon Justicia, Challenge 2: Childhood and Youth.
- 2025 — Co-supervised 1 bachelor’s thesis and 2 master’s theses.
- 2024 — Co-supervised 1 bachelor’s thesis.
- 2020–2021 — Student mentor, Student Mentoring Programme (PAT), Faculty of Computer Science, Universidade da Coruña, both semesters.
Outreach
- 25 September 2026 — Outreach participant, G-Night / European Researchers’ Night, Universidade da Coruña.
- 2026 — Participation in the DAFIC Research Fair, Faculty of Computer Science, Universidade da Coruña.
Academic Training
- 2025 — 16th European Summer School on Information Retrieval (ESSIR 2025), University of Wolverhampton, UK.