Dynamic state representation learning is an important task in robot learning and has wide application in areas such as accelerating model free reinforcement learning, closing the simulation to reality gap, as well as reducing the motion planning complexity.
However, current dynamic state representation learning methods scale poorly on complex dynamic systems such as deformable objects, and can not directly embed well defined simulation function into the training pipeline.
In this work, we propose to progressively increase the training difficulty during learning a neural network model via a novel strategy which we call mini-batch trimming.
Mini-batch trimming makes sure that the optimizer puts its focusin the later training stages on the more difficult samples, which we identify as the ones with the highest loss in the current mini-batch.
Deep learning has made great progress in recent years, but the exploding economic and environmental costs of training neural networks are becoming unsustainable.
To address this problem, there has been a great deal of research on *algorithmically-efficient deep learning*, which seeks to reduce trainingcosts not at the hardware or implementation level, but through changes in thesemantics of the training program.