Researcher Interests

I am a PhD student and researcher in Deep Learning, specializing in Natural Language Processing (NLP) and conversational systems. My experience spans several years working on chatbots and NLP, initially from a business perspective in developing commercial chatbots and natural language processing applications, focusing particularly on document analysis and textual entities in the legal field. Academically, my work has included research in computer vision, with an emphasis on biometric re-identification in videos and biometrics, as well as in chatbot dialogue management.

Currently, I am dedicated to advanced research in large language models, focusing on explainability from a mechanistic interpretation perspective and adapting these models to specific domains, in addition to striving to reduce their size.

News

• January 2024: I published a paper at International Journal on Natural Language Computing. Here is the link: paper
• September - December 2023: I taught classes in the Data Science and Computer Science degree programs, specifically in the Deep Learning course.
• 2023: I published a paper at NATP. Here is the link: paper.
• 2022: I published a paper at ICPRAM. Here is the link: paper.
• June 2022: I attended the Deep Learning Summer School. More information can be found here: Deep Learning Summer School.
• June 2021: I completed my Master's in Intelligent Systems and Numerical Applications in Engineering.
• June 2020: I graduated from the University of Las Palmas de Gran Canaria with a degree in Computer Science, specialized in Artificial Intelligence.

Publications

• Medina-Ramírez, M. Á., Guerra-Artal, C., & Hernández-Tejera, M. (2024). Improving Dialogue Management Through Data Optimization. Journal on Natural Language Computing
• Medina-Ramírez, M. Á., Guerra-Artal, C., & Hernández-Tejera, M. (2024). Analysis of the Impact of Dataset Quality on Task-Oriented Dialogue Management. In Proceedings of the 10th International Conference on Natural Language Processing NATP.
• Medina Ramírez, M. Á., Lorenzo Navarro, J. J., Freire Obregón, D. S., Santana, O. J., Hernández Sosa, J. D. (2022). Boosting Re-identification in the Ultra-running Scenario. In Proceedings of the 11th International Conference on Pattern Recognition Applications and Methods - ICPRAM

Experience

• Deep Learning Researcher at ULPGC - SIANI (June 2022 - Present)

Engaged in advanced research in deep learning and natural language processing. Spearheaded the development of a state-of-the-art conversational system for text generation. Specialized in working with transformer models and large language models (LLMs), focusing on modular deep learning approaches. Conducting lectures and workshops on deep learning methodologies and applications.

• Data Scientist at ULPGC - Pormar (June 2021 - June 2022)

Conducted network calculations and predictive modeling for ship movement and global maritime trade. Extensive research in data science and machine learning for maritime applications. Developed innovative visualizations for maritime traffic and predictive routing.

• Computer Vision Researcher at ULPGC - SIANI (November 2020 - May 2021)

Research and development in video-based person identification using deep learning. Applied advanced image resolution enhancement techniques. Led the creation of a biometric identification and re-identification application.

• Software Engineer at Acoran (February 2020 - November 2020)

Developed web services for document management and compliance with data protection laws. Worked on NLP algorithms for entity recognition and real-time system integration.

• Junior Software Engineer at Dr. Stetter ITQ (October 2019 - February 2020)

Involved in the development and maintenance of a RASA-based chatbot on AWS and Docker. Engaged in data visualization for a meteorological data collection application. Contributed to the development of software for intelligent farming systems, focusing on computer vision and micro-services. Performed testing and software architecture tasks for an inventory application using TypeScript and Node.js.