Marcelo Mendoza
Marcelo Mendoza is a computer scientist whose AI research brings together natural language processing, information retrieval and social-network analysis. He earned degrees in engineering and computing in Chile, held academic and research posts at the University of Valparaíso, Yahoo! Research and UTFSM, and has served as president of the Chilean Association for Pattern Recognition since 2015. His work includes social-bot detection, a BERT-style approach to discourse, the CLNews misinformation dataset and language-model methods for extracting key facts from news. In 2025, he was named to the World's Top 2% Scientists list, and as of October 2026, he is an associate professor at Pontificia Universidad Católica de Chile and a principal investigator at CENIA.
- Now at
- Chilean Association for Pattern Recognition, Pontifical Catholic University of Chile and National Center for Artificial Intelligence
- Badges
Timeline
-
1997–2007
Engineering and a computing PhD
At UTFSM, he earns a licentiate in engineering in 1997 and a master's in informatics engineering in 2001. At the University of Chile's Department of Computer Science, he earns a doctorate in computing in 2007.
-
2002–2014
Valparaíso and Yahoo Research
From 2002 to 2009, he is a full-time associate professor at the University of Valparaíso. From 2009 to 2010, he is a postdoctoral researcher at Yahoo! Research in the University of Chile's Department of Computer Science. He also works part-time as a Yahoo Labs researcher from 2009 to 2014.
-
2010–2022
Associate professor at UTFSM
From 2010 to 2022, he is a full-time associate professor in the Department of Informatics at UTFSM.
-
Since 2015
President of ACHIRP
He serves as president of the Chilean Association for Pattern Recognition from 2015.
-
2020–2022
Bots, BERT and CLNews
His research spans text mining, information retrieval and data mining in social networks. In 2020, he coauthors “Bots in Social and Interaction Networks: Detection and Impact Estimation”; in 2021, he coauthors a BERT-style approach that uses predictive coding for discourse representations and improves performance in six of 11 tasks. In 2022, he coauthors CLNews, a dataset for disinformation analysis of Chile's social outbreak.
-
April 2025
Extracting 5W1H from news
At The ACM Web Conference in April 2025, he presents “Imitating Human Reasoning to Extract 5W1H in News.” The paper introduces extractive and question-level reasoning methods that guide language models through news extraction.
-
2025
The World's Top 2% Scientists
CENIA includes him in the 2025 edition of the World's Top 2% Scientists list for the sustained impact of his work.
-
As of October 2026
UC professorship and Spanish medical AI
As of October 2026, he is an associate professor in the Department of Computer Science at Pontificia Universidad Católica de Chile. In March 2026, he serves as a principal investigator at CENIA and co-investigator on a CENIA and UC project selected for an NVIDIA academic grant to develop a Spanish-language medical AI model.
Career
- Since 2026National Center for Artificial IntelligencePrincipal Investigator
- Since 2015Chilean Association for Pattern RecognitionPresident
- NowPontifical Catholic University of ChileAssociate Professor
- 2010–2022Federico Santa María Technical UniversityAssociate Professor
- 2009–2014YahooPostdoctoral Researcher, Researcher
- 2002–2009University of ValparaísoAssociate Professor
Education
- 2007University of ChileDoctorate in Computing
- 2001Federico Santa María Technical UniversityMaster's in Informatics Engineering
- 1997Federico Santa María Technical UniversityLicentiate in Engineering
Sources and credits
Photo: We have no published photo of Marcelo Mendoza yet. Send us a free-licensed photo.
- Pagina DCC - Marcelo Mendoza R.
- PDI - Resultado de Búsqueda
- Marcelo Mendoza – Millennium Institute Foundational Research on Data
- Augmenting BERT-style Models with Predictive Coding to Improve Discourse-level Representations
- PUBLICATIONS – Millennium Institute Foundational Research on Data
- Investigador de Cenia presenta en The ACM Web Conference 2025 método de IA que mejora extracción de información en noticias
- Investigadores de CENIA Pablo Barceló y Marcelo Mendoza destacan en ranking mundial del 2% de los científicos más influyentes - CENIA
- Investigadores CENIA y UC obtienen apoyo de NVIDIA para el primer modelo de IA médica especializado en español - CENIA
- Marcelo Mendoza - Licenciatura en Ingeniería en Ciencia de la Computación
