\chapter{Projects, Grants and Activities}\label{app:grants}
This appendix presents a comprehensive overview of the research projects, grants, research stays, conference presentations, and other scholarly activities undertaken during the course of this PhD research. These activities have not only supported the development of the research presented in this thesis but have also provided valuable opportunities for collaboration, knowledge exchange, and professional development in the field of intelligent ML model adaptation.
\section{Projects and Research Collaborations}
My doctoral research has been supported by several significant research projects and collaborations, providing both financial support and opportunities for interdisciplinary research:
\begin{itemize} \item \textbf{BSC/Lenovo Joint Research Project (2018-2020)}: Participated as a PhD student on big data deployment, which provided the initial foundation for understanding the challenges of deploying ML models in production environments.
\item \textbf{BSC/Lenovo Joint Research Project (2020-2022)}: Continued collaboration focusing on the BSC Scanflow framework, which informed the development of monitoring approaches presented in this thesis.
\item \textbf{BSC/Lenovo Joint Research Project (2023)}: Continual Explainable Drift Detection and Adaptation for Autonomous Driving Data. Directly supporting the development of the TADIL and TADILER approaches.
\item \textbf{BSC/Lenovo Joint Research Project (2024)}: Latest collaboration on “Dual Behavior and Intelligent Agents for Continuous Learning in Production,” directly supporting the development of the CAMA and KC-Agent architectures presented in Chapter \ref{chapter5} and Chapter \ref{chapter6}. This six-month project with $50,000 funding (January-June 2024) focused on implementing and evaluating dual-process architectures for ML model improvement.
\item \textbf{Computing Resources Orchestration and Management for AI}: Regional research project funded by AGAUR (Agencia de Gestio d’Ajuts Universitaris i de Recerca) with a budget of 60,000 euros, running from January 2022 to June 2025. My participation as a PhD student under Scientific Coordinator Berral, J. has provided valuable insights into resource management for AI systems.
\item \textbf{Computaci'on de Altas Prestaciones VIII}: National research project funded by Spanish Ministry of Science (MICINN) under grant PID2019-107255GB-C22, running from 2020 to 2025. My participation in this project provided computational resources and infrastructure support for the research presented in this thesis.
\item \textbf{DALEST (Distributed Analytics and Learning in Edge-to-Supercomputing Technologies)}: National research project funded by Spanish Ministry of Science (MICINN) under grant PID2021-126248OB-I00, running from 2022 to 2025. This project supported aspects of my research on efficient ML adaptation techniques across different computing environments.
\end{itemize}
\section{Research Stays}
Research visits to other institutions have been instrumental in broadening the scope of my research and establishing international collaborations:
\begin{itemize} \item \textbf{Emory Global Diabetes Research Center (July-October 2023)}: This three-month research stay focused on designing continual learning applications for health and epidemiological research, with an emphasis on making models robust across different domains (location, nationality, noise). The visit also involved designing a multi-agent system using Large Language Models for Domain-incremental Learning for healthcare images and text, with the goal of creating an agent that interacts with experts to improve ML models interactively. This collaboration directly informed the development of the TADILER approach presented in Chapter \ref{chapter4}.
\item \textbf{Emory Global Diabetes Research Center (January-July 2024)}: Six-month research stay focused on developing a cognitive architecture for LLM-agents, particularly a decision-making procedure for cost-efficient use of limited resources. This collaboration was instrumental in refining the KC-Agent architecture presented in Chapter \ref{chapter6}, especially its mechanisms for balancing computational efficiency with adaptation effectiveness. \end{itemize}
\section{Conference Presentations}
Presenting my research at international conferences has provided valuable opportunities for feedback, discussion, and dissemination of the novel approaches developed during this PhD:
\begin{itemize} \item \textbf{CVPR Workshop: LatinX in Computer Vision Research Workshop 2023}: Presented the extended abstract “TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings” on June 18, 2023, in Vancouver, Canada. This presentation showcased the TADIL approach described in Chapter \ref{chapter3}.
\item \textbf{CVPR Workshop: LatinX in Computer Vision Research Workshop 2024}: Presented the extended abstract “Enhancing Continual Learning in Diabetic Retinopathy: Multimodal Zero-shot Clustering and Strategic Experience Replay” on June 18, 2024, in Seattle, USA. This presentation featured the TADILER approach detailed in Chapter \ref{chapter4}.
\item \textbf{International Conference on Pattern Recognition (ICPR 2024)}: Delivered an oral presentation of our paper “TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings” on December 04, 2024, in Kolkata, India. This work was selected for oral presentation from among all accepted papers and represents the core contribution detailed in Chapter \ref{chapter3}.
\item \textbf{International Conference on Computer Vision Theory and Applications (VISAPP 2025)}: Our paper “Experience Replay and Zero-shot Clustering for Continual Learning in Diabetic Retinopathy Detection” was shortlisted for the Best Student Paper Award, based on reviewer evaluations. This recognition highlights the significance and quality of the TADILER approach presented in Chapter \ref{chapter4}. It was presented on February 26, 2024, in Porto, Portugal.
\item \textbf{24th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2025)}: Will present our paper “Feature Engineering for Agents: An Adaptive Cognitive Architecture for Interpretable ML Monitoring” on Thursday, May 22, 2025, in Detroit, Michigan, USA. AAMAS is recognized as the premier and most influential conference for research on AI agents and multiagent systems. This presentation will showcase the CAMA architecture detailed in Chapter \ref{chapter5}, contributing to the field’s understanding of interpretable monitoring in intelligent agent systems.
\end{itemize}
\section{Invited Talks and Poster Presentations}
In addition to conference presentations, I have had opportunities to share my research through invited talks and poster sessions:
\begin{itemize} \item \textbf{ContinualAI Invited Speaker (2023)}: Delivered an invited talk titled “TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings” for the ContinualAI community’s YouTube channel, focusing on task-agnostic continual learning for the open world. This presentation reached a specialized audience interested in CL approaches. URL: \url{https://youtu.be/LMtnVBFfgPc?si=G0z95aFMP3L5XsUu}
\item \textbf{Xarxa RDI-IA Annual Conference (2024)}: Presented a poster on “TADIL: Task-Agnostic Domain-Incremental Learning through Task-ID Inference using Transformer Nearest-Centroid Embeddings” at this conference organized by the Agencia de Gestio d’Ajuts Universitaris de Recerca in Barcelona, Spain, on May 22, 2024.
\item \textbf{Joint EurAI Advanced Course on AI, TAILOR Summer School (2022)}: Presented my PhD research lines and latest work in Barcelona, Spain, providing an overview of the approaches being developed for intelligent ML model adaptation. \end{itemize}
\section{Media Contributions}
My research has also reached broader audiences through media contributions:
\begin{itemize} \item \textbf{Lenovo Press Article (2023)}: Authored an article titled “Autonomous Vehicles Use New AI Algorithm to Learn from Changes in the Environment,” published on April 14, 2023. This article highlighted the applications of the TADIL approach to autonomous driving scenarios. URL: \url{https://lenovopress.lenovo.com/lp1716-autonomous-vehicles-use-new-ai-algorithm} \end{itemize}
\section{Grants and Awards}
Throughout my PhD, I have been fortunate to receive several grants and awards that have supported my research and conference attendance:
\begin{itemize} \item \textbf{VISIGRAPP 2025 Best Student Paper Award Finalist}: Our paper on experience replay and zero-shot clustering for continual learning in diabetic retinopathy detection was shortlisted for this prestigious award from the Institute for Systems and Technologies of Information, Control and Communication.
\item \textbf{CVPR 2024 Grants}: \begin{itemize} \item Stipend grant award from Accel AI Institute \item Stipend grant award from IEEE Computer Society \item Registration grant award from IEEE Computer Society \end{itemize}
\item \textbf{CVPR 2023 Grant}: Registration grant award from Accel AI Institute \end{itemize}
\section{Attendance at Courses, Seminars, and Conferences}
Participation in various courses, seminars, and conferences has provided valuable opportunities for continuous learning and staying current with developments in the field:
\begin{itemize} \item \textbf{International Conference on Computer Vision Theory and Applications (VISAPP 2025)}: Attended this specialized conference on computer vision applications from February 24-26, 2025, at the Vila Gale hotel in Porto, Portugal.
\item \textbf{International Conference on Pattern Recognition (ICPR 2024)}: Attended this prestigious conference on pattern recognition from December 1-5, 2024, at the Biswa Bangla Convention Centre in Kolkata, India.
\item \textbf{IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2024)}: Attended this premier computer vision conference from June 17-21, 2024, at the Seattle Convention Center in Seattle, USA.
\item \textbf{IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2023)}: Attended this premier conference on computer vision from June 18-22, 2023, at the Vancouver Convention Center in Vancouver, Canada.
\item \textbf{ContinualAI Unconference (2023)}: Participated in this online event focusing on continual learning approaches and challenges.
\item \textbf{PyData Global (2022)}: Attended this online conference on data science, machine learning, and AI.
\item \textbf{AI \& Big Data Congress (2022)}: Participated in this congress in Barcelona, Spain, focusing on advances in AI and big data technologies.
\item \textbf{CIDAI-Masterclass on AI Observability (2022)}: Attended this masterclass in Barcelona, Spain, focusing on AI observability and AI stores.
\item \textbf{Joint EurAI Advanced Course on AI, TAILOR Summer School (2022)}: Participated in this advanced course in Barcelona, Spain, where I also presented my PhD research lines. \end{itemize}
These diverse activities have significantly contributed to the development of the research presented in this thesis, providing opportunities for feedback, collaboration, and professional growth. The international nature of these activities has ensured exposure to diverse perspectives and approaches, enriching the research outcomes and their potential applications.