Transfer Learning

Transfer Learning: Algorithms and Applications presents an in-depth discussion on practices for transfer learning, exploring emerging fields that includes a theoretical analysis of various algorithms and problems that lay a solid foundation for future advances in the field.

In the era of Big Data, machine learning methods are widely used in natural language processing, computer vision, speech, and in signal processing communities. However, the current standard machine learning techniques, such as supervised classifiers, tend to fail when the data distribution and/or structure changes over training and test settings.

Current techniques addressing machine learning problems can only address a few isolated tasks at one time. Transfer learning, adapted from how humans learn, models the distribution and structure difference between training and test settings.

• Introduces transfer learning with a systematic approach, discussing theory and providing applications, including but not limited to, image classification, natural language techniques, medicine, and web search ranking techniques
• Provides a state-of-the-art overview of the most recent developments in transfer learning, including unsupervised, supervised, and semi-supervised transfer learning, multitask learning, domain similarity estimation, and the applications of transfer learning
• Presents relevant algorithms with detailed discussions, including background, derivation, and comparisons
• Discusses extensive experimental results using real application datasets to demonstrate the performance of various algorithms