Optimization as a Model for Few-Shot Learning

  • Quand ? 10/01/2017 à partir de 13:30 (America/Montreal / UTC-500)
  • Où ? Au local D4-2011 de la Faculté des sciences
  • Participants Hugo Larochelle, professeur au Département d'informatique de l'Université de Sherbrooke
  • Ajouter l'événement au calendrier iCal

RÉSUMÉ : Though deep neural networks have shown great success in the large data domain, they generally perform poorly on few-shot learning tasks, where a model has to quickly generalize after seeing very few examples from each class. The general belief is that gradient-based optimization in high capacity models requires many iterative steps over many examples to perform well. Here, we propose an LSTM-based meta-learner model to learn the exact optimization algorithm used to train another learner neural network in the few-shot regime. The parametrization of our model allows it to learn appropriate parameter updates specifically for the scenario where a set amount of updates will be made, while also learning a general initialization of the learner network that allows for quick convergence of training. We demonstrate that this meta-learning model is competitive with deep metric-learning techniques for few-shot learning.

Toutes les personnes intéressées sont cordialement invitées.