How and when to stop the co-training process

作者：

Highlights：

• Demonstrating the usage of model's outputs for overfitting or noise detection.

• Retrieving a near-optimal co-training model without using a validation set.

• Mostly co-training results cannot be improved further after a number of iterations.

• Co-training has greater effect in transfer learning (TL) than in non-TL scenarios.

摘要

•Demonstrating the usage of model's outputs for overfitting or noise detection.•Retrieving a near-optimal co-training model without using a validation set.•Mostly co-training results cannot be improved further after a number of iterations.•Co-training has greater effect in transfer learning (TL) than in non-TL scenarios.

论文关键词：Semi-supervised,Co-training,Multi-view,Transfer learning

论文评审过程：Received 22 May 2020, Revised 14 July 2021, Accepted 30 August 2021, Available online 17 September 2021, Version of Record 11 October 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115841