The Siamese networks have been in recent years used for a lot of applications. Their first used in the LeCun paper was for signature verification. Since then a large number of applications have emerged, we will present some of the recent ones:
- Joint Learning of Speaker and Phonetic Similarities with Siamese Networks (https://pdfs.semanticscholar.org/4ffe/3394628a8a0ffd4cba1a77ea85e197bd4c22.pdf): They trained a multi-output Siamese network, with one output for phonetic similarity and other for the speaker similarity. They extended the work to include Triamese Network.
- Fully-Convolutional Siamese Network for Object Tracking (https://link.springer.com/chapter/10.1007/978-3-319-48881-3_56): They used convolutional Siamese network trained on ILSVRC15 dataset for object detection in Video.
- Together We stand: Siamese Networks for Similar Question Retrieval (http://www.aclweb.org/anthology/P16-1036): The paper used Siamese Network to find the semantic similarity between the current and archived questions. They also used convolutional Siamese Networks.
Besides these, Siamese networks have also been explored for face verification/recognition (https://github.com/harveyslash/Facial-Similarity-with-Siamese-Networks-in-Pytorch). They have been used for Question-Answering (https://arxiv.org/pdf/1512.05193v2.pdf).