Object detection models are being widely used in many applications, such as autonomous driving, construction management, and cancer detection. Evaluating the performance of the object detection model is more complicated than other computer vision models such as image classification models. Most of the images have several objects to be detected,...
Agent-based models (ABM) are widely used in network data analysis, and due to their simple structures and sophisticated outcomes, they serve as good tools in understanding the dynamics in networks. In this thesis, we develop an agent-based dynamic network model, and show that it can replicate the expected degree distribution...
Reachability analysis enables the safety assurance of control systems despite uncertain initial conditions and control inputs, and can be an important component to run on-board an autonomous system. This thesis explores the characteristics of reachability analysis with different algorithmic configurations and runtime parameters running onboard the F1Tenth 1/10th-scale autonomous racing...
Humans are remarkably efficient in learning by interacting with other people and observing their behavior. Children learn by watching their parents’ actions and mimic their behavior. When they are not sure about their parents demonstration, they communicate with them, ask questions, and learn from their feedback. On the other hand,...
In open set recognition, a classifier must label instances of known classes while detecting instances of unknown classes not encountered during training. To detect unknown classes while still generalizing to new instances of existing classes, this thesis introduces a dataset augmentation technique called counterfactual image generation. This approach, based on...
Many methods have been explored in the literature of multi-label learning, ranging from simple problem transformation to more complex method that capture correlation among labels. However, mostly all existing works do not address the challenge with incomplete label data. The goal of this project is to extend the work of...
Semi-supervised clustering aims to improve clustering performance by considering user supervision in the form of pairwise constraints. In this paper, we study the active learning problem of selecting pairwise must-link and cannot-link constraints for semisupervised clustering. We consider active learning in an iterative manner where in each iteration queries are...
In an increasingly computation-driven world, algorithms and mathematical models significantly impact decision making across various fields. To foster trust and understanding, it is crucial to provide users with clear and concise explanations of the reasoning behind the results produced by computational tools, especially when recommendations appear counterintuitive. Legal frameworks in...
The types and rates of label noise in real-world data sets present a challenge to machine learning projects. In this thesis, we propose a novel approach to address this issue. Our method combines a noise modelling technique for correcting label noise across the entire data set with a robust loss...
Learning latent space representations of high-dimensional world states has been at the core of recent rapid growth in reinforcement learning(RL). At the same time, RL algo- rithms have suffered from ignored uncertainties in the predicted estimates of model-free or model-based methods. In our work, we investigate both of these aspects...