Thy Neighbor

"Thy Neighbor" research explores how leveraging information from nearby data points (neighbors) improves various machine learning tasks. Current efforts focus on developing algorithms and model architectures that effectively utilize this neighborhood information, including graph neural networks, k-nearest neighbor methods, and contrastive learning approaches, across diverse applications like federated learning, image classification, and graph-based anomaly detection. This research significantly impacts the field by enhancing model accuracy, efficiency, and robustness, particularly in scenarios with limited data, noisy observations, or privacy constraints, leading to improvements in various applications.