ScreenerNet: Learning Self-Paced Curriculum for Deep Neural Networks

Abstract

We propose to learn a curriculum or a syllabus for supervised learning and deep reinforcement learning with deep neural networks by an attachable deep neural network, called ScreenerNet. Specifically, we learn a weight for each sample by jointly training the ScreenerNet and the main network in an end-to-end self-paced fashion. The ScreenerNet neither has sampling bias nor requires to remember the past learning history. We show the networks augmented with the ScreenerNet achieve early convergence with better accuracy than the state-of-the-art curricular learning methods in extensive experiments using three popular vision datasets such as MNIST, CIFAR10 and Pascal VOC2012, and a Cart-pole task using Deep Q-learning. Moreover, the ScreenerNet can extend other curriculum learning methods such as Prioritized Experience Replay (PER) for further accuracy improvement.

arXiv commentcurricular learning, deep learning, deep q-learning

References in corpus (8)

• Adam: A Method for Stochastic Optimization
• Playing Atari with Deep Reinforcement Learning
• Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks
• Focal Loss for Dense Object Detection
• Hindsight Experience Replay
• Online Batch Selection for Faster Training of Neural Networks
• Intrinsic Motivation and Automatic Curricula via Asymmetric Self-Play
• Adaptive Feeding: Achieving Fast and Accurate Detections by Adaptively Combining Object Detectors

Cited by in corpus (6)

• Curriculum Learning for Reinforcement Learning Domains: A Framework and Survey
• Optimizing Millions of Hyperparameters by Implicit Differentiation
• A Survey on Curriculum Learning
• AutoAssist: A Framework to Accelerate Training of Deep Neural Networks
• Self-paced Data Augmentation for Training Neural Networks
• Learning Error-Driven Curriculum for Crowd Counting