Learning by coincidence: Siamese networks and common variable learning
作者：
Highlights：
• Siamese networks are proposed for unsupervised learning of a common source of variability in multimodal data.
• The common source of variability is identified using coincidence and concurrence of measurements in different modalities.
• The functionality of Siamese networks is interpreted as trying to learn equivalence classes associated with a common source of variability present in different modalities.
摘要
•Siamese networks are proposed for unsupervised learning of a common source of variability in multimodal data.•The common source of variability is identified using coincidence and concurrence of measurements in different modalities.•The functionality of Siamese networks is interpreted as trying to learn equivalence classes associated with a common source of variability present in different modalities.
论文关键词：Similarity learning,Representation learning,Unsupervised learning,Siamese networks,Common variable
论文评审过程：Received 8 March 2017, Revised 4 August 2017, Accepted 7 September 2017, Available online 8 September 2017, Version of Record 15 September 2017.
论文官网地址：https://doi.org/10.1016/j.patcog.2017.09.015
