Confidence Measures for Deep Learning in Domain Adaptation

• Published in: Applied sciences Vol. 9; no. 11; p. 2192
• Main Authors: Bonechi, Simone, Andreini, Paolo, Bianchini, Monica, Pai, Akshay, Scarselli, Franco
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.06.2019
• Subjects: Adaptation; Annotations; Artificial intelligence; confidence measures; Datasets; Deep learning; Generative Adversarial Networks; Image analysis; Image processing; Information processing; International conferences; Learning algorithms; Machine learning; Medical imaging; Network reliability; Neural networks; uncertainty estimation; unsupervised domain adaptation; Voice recognition; Denmark; Sydney New South Wales Australia; Canada; Australia; United States > US
• ISSN: 2076-3417; 2076-3417
• DOI: 10.3390/app9112192

In recent years, Deep Neural Networks (DNNs) have led to impressive results in a wide variety of machine learning tasks, typically relying on the existence of a huge amount of supervised data. However, in many applications (e.g., bio–medical image analysis), gathering large sets of labeled data can be very difficult and costly. Unsupervised domain adaptation exploits data from a source domain, where annotations are available, to train a model able to generalize also to a target domain, where labels are unavailable. Recent research has shown that Generative Adversarial Networks (GANs) can be successfully employed for domain adaptation, although deciding when to stop learning is a major concern for GANs. In this work, we propose some confidence measures that can be used to early stop the GAN training, also showing how such measures can be employed to predict the reliability of the network output. The effectiveness of the proposed approach has been tested in two domain adaptation tasks, with very promising results.