Artificial Intelligence and Machine Learning
The Artificial Intelligence (AI) and Machine Learning (ML) research pillar at the FLAG lab, focuses in the development of cutting edge fundamental reserach approaches for the advancement of this important domain. Furthermore, we are interested in the intersection of AI and Machine Learning with other disciplines in the domain of computer science such as recommender systems, smart cities, and the Internet of Things, to mention a few. These approaches are applied in a variety of domains including technology enchanced learning, software verification, and programming languages.
Natural Language Processing
Development of NLP and Large Language Models (LLMs) at FLAG lab is divided into different topics of interest.
-
Working on memory models to improve the agent reasoning across sessions, as well as the incorporation of new knowledge that can be effectively exploited by agents.
-
Developing explainability techniques to help undesrtand the internal reasoning process of agents and the reasons why certain decisions are taking.
-
Evaluation of LLMs and their role for software development.
-
Generation of datasets and analysis of low resource languages.
Reinforcement Learning
Our research in Reinforcement Learning (RL) is split into three main themes. (1) The development of new RL algorithms and techniques to improve learning and explore interaction between agents. In particular, we are interested in accelerating/improvong agent learning in environments to fulfill complex tasks.(2) Exploring the use of RL for the development of adaptive systems at a programming language level, and the proposal of programming language abstractions to manage/hide the complexity of RL algorithms. (3) The evolution and adaptations of Rl agents through their life with life-long learning techniques. (4) Creating software development and analysis tools for RL programs, with emphasis on software quality and debugging to improve the experience of RL developers.
Our work is applied to robotic environments and industrial environments in federated and distributed systems.
Most of our work on RL is developed in collaboration with the Science Foundation Ireland center for research training in AI, at Trinity College Dublin.
Machine Learning
Studying the use and development of supervised and unsupervised ML algorithms for both Deep and regular Neural Networks.
Get into the code
To check out our current work you can go to our dedicated github repositories
Nicolás Cardozo - Haydemar Nuñez - Ruben Manrique - Manuel Mosquera - Nicolas De la Hoz - Luis Gerardo Manrique -