Novel Approach
This research explores novel approaches across diverse fields, aiming to improve existing methods and address limitations in various machine learning and AI applications. Current efforts focus on enhancing model performance and robustness through techniques like active learning, diffusion models, and transformer architectures, often incorporating advanced concepts such as graph isomorphism networks and attention mechanisms. These advancements have significant implications for various domains, including robotics, personalized recommendations, medical image analysis, and cybersecurity, by improving accuracy, efficiency, and interpretability. The overall goal is to create more powerful, reliable, and explainable AI systems.
Papers
From Keypoints to Object Landmarks via Self-Training Correspondence: A novel approach to Unsupervised Landmark Discovery
Dimitrios Mallis, Enrique Sanchez, Matt Bell, Georgios Tzimiropoulos
A novel approach to rating transition modelling via Machine Learning and SDEs on Lie groups
Kevin Kamm, Michelle Muniz