Google Scholar Page
Curso RecSys(PUC)
Coursera MOOC RecSys (es)
Coursera MOOC Visualizacion (es)
Software: pyReclab
Resume(CV)
Talks' Slides (via Slideshare)
Old Web Site
Hi! I am Denis, Associate Professor at the Department of Computer Science, in the School of Engineering at PUC Chile. I am also affiliated to three centers of research excellence in Chile: i-Health, CENIA, and IMFD. I am principal researcher at the Millenium Institute for Intelligent Healthcare Engineering (iHealth), as well as associate researcher at the National Center for Artificial Intelligence (CENIA). I am also adjunct researcher at the Millennium Institute for Research on Fundamentals of Data (IMFD).
I hold a professional title of Civil Engineer in Informatics, class of 2004 from UACh, Valdivia, Chile; and a Ph.D. in Information Science from the University of Pittsburgh, USA, where I was supervised by Professor Peter Brusilovsky. I earned a Fulbright scholarship to pursue my PhD studies between 2008-2013.
My research interests are Recommender Systems, Applications of Artificial Intelligence (Medical AI, Creative AI) and Intelligent User Interfaces. My research has been presented in leading AI venues such as ICML, Neurips, ICLR, ACL and CVPR. I am currently leading the Human-centered AI and Visualization (HAIVis) research group as well as co-leading the CreativAI Lab with professor Rodrigo Cádiz. I am also Faculty member of the PUC IA Lab.
If you are interested on these topics and want to work under my supervision as Master or PhD student, send me an e-mail to denis.parra[at]uc[dot]cl.
June 2026 PhD student Pablo Messina as first author, in collaboration with KAUST, had this research article accepted at CVPR 2026, a leading conference in computer vision. This article introduces CURE, a method for adaptive curriculum learning which allow a model like MedGemma improve on Visually grounded report generation tasks. CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report Generation.
April 2026 Paper presentar at ICLR 2026 in Rio de Janeiro, Brasil. A collaboration with researchers at CENIA where we investigated potential redundancy of layers in Transformers, and we introduce a method that identifies these layers better than previous cosine similarity-based methods. Rethinking Layer Relevance in Large Language Models Beyond Cosine Similarity.
December 2025 We presented a paper at Neurips "A compressive-expressive communication framework for compositional representations" in San Diego, CA, USA. This article was co-authored with Rafael Elberg (MSc student) and Mircea Petrache, Professor at Math Departament at PUC.
September 2025 I attended de MICCAI conference in Daejeon, South Korea, were I presented two papers in the HAIC workshop.
August 2024 Attending ACL 2024 conference in Bangkok, Thailand. We are presenting one Findings paper "Extracting and Encoding: Leveraging Large Language Models and Medical Knowledge to Enhance Radiological Text Representation" [link] and two papers in the BioNLP workshop about radiological report generation led by Oscar Loch and Diego Campanini.
July 2024 Attending the ICML 2024 conference in Vienna, Austria. We are presenting the article "On the Unexpected Effectiveness of Reinforcement Learning for Sequential Recommendation"[paper][ppt], with first author MSc student Alvaro Labarca and in collaboration with colleague Rodrigo Toro.
March 2024 I am one of the three Program Chairs at the [ACM IUI conference (along Dorota Glowaka and Ofra Amir), which takes place in Greenville, SC, USA this year.
April 2023-September 2023 On sabbatical leave, I'm spending these 6 months at KU Leuven in Belgium researching on XAI and text report generation from medical images.|
Messina et al, CVPR 2026, CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report Generation [conf]
Messina et al, ACL 2024, Extracting and encoding: Leveraging large language models and medical knowledge to enhance radiological text representation [conf] Pino et al, MICCAI 2021, Clinically Correct Report Generation from Chest X-Rays Using Templates CNN-TRG [conf] Messina et al, MedNeurips 2022, Two-stage Conditional Chest X-ray Radiology Report Generation CRG [conf] |
| A Survey on Deep Learning for Automatic Generation of Radiological Reports (*for a pre-print of this article, check this ArXiv link) |
|
Slides: Evaluation of radiological report generation [pdf]
Slides: Visual, language and mutimodal models for radiological report generation [pdf] |
|
|
|
| TalkExplorer [conf][journal][demo] | SetFusion [conf][journal][video] [slides] | Moodplay [conf] [video] |
| A Survey on Interactive Recommender Systems | ||
| *for pre-prints of these articles, check my publications page | ||