Siamese Networks for Few Shot Learning

Machine learning, and especially Deep Learning, has achieved state-of-the-art performance in a variety of applications using large databases. However, these algorithms break down when forced to make predictions with little supervised information.

This problem, also known as Few Shot Learning, can be properly addressed by Siamese Networks, which are based on a architecture that computes similarities among highly semantic features representations.

Monday 16th December 2019 (16:00 - 17:00)

The conference will be given by:

Prof. Carlos R. del Blanco

Grupo de Tratamiento de Imágenes - UPM

Where:

ETSI Telecomunicación

UNIVERSIDAD POLITÉCNICA DE MADRID

Avda. Complutense 30

A Building (Room A-123)

This Seminar is joinly organized by:

IPTC Information Processing and Telecommunications Center

MSTC Master Program: Signal Processing and Machine Learning for Big Data