Ensemble Aggregation
Ensemble aggregation combines predictions from multiple machine learning models to improve accuracy and robustness. Current research focuses on understanding and mitigating prediction instability in ensembles, exploring diverse aggregation techniques like weighted averaging and Bayesian optimization, and adapting ensembles to various contexts such as streaming data, transfer learning (e.g., using manifold-based methods), and handling long clinical texts. This work is significant for enhancing the reliability and explainability of machine learning models across numerous applications, from healthcare and robotics to emotion recognition and general classification tasks.
Papers
October 17, 2024
July 3, 2024
April 24, 2024
November 8, 2023
November 2, 2023
October 1, 2023
June 2, 2023
July 4, 2022