Data Integration

Data integration aims to combine data from disparate sources, improving data quality, analysis accuracy, and model performance. Current research focuses on addressing challenges like bias in data processing, enhancing model training with diverse data types (including cross-species and multi-sensor data), and developing robust methods for handling heterogeneous data formats and schemas, often employing deep learning architectures like LSTMs and transformers, along with techniques like federated learning and ontology matching. These advancements are crucial for various applications, including activity recognition, medical image analysis, autonomous driving, and industrial AI, enabling more accurate predictions, improved decision-making, and more equitable outcomes.

Papers