# CASSL: Curriculum Accelerated Self-Supervised Learning

arXiv: https://arxiv.org/abs/1708.01354 · HTML: https://mlsanity.com/paper/1708.01354/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Aug 2017.

## Built on by (most cited first, 6 of 6)

- [Curriculum Learning: A Survey](https://mlsanity.com/paper/2101.10382/) (2021)
- [Boost Your Human Image Generation Model via Direct Preference Optimization](https://mlsanity.com/paper/2405.20216/) (2024)
- [One After Another: Learning Incremental Skills for a Changing World](https://mlsanity.com/paper/2203.11176/) (2022)
- [Controllable Augmentations for Video Representation Learning](https://mlsanity.com/paper/2203.16632/) (2022)
- [Medical Knowledge-Guided Deep Curriculum Learning for Elbow Fracture Diagnosis from X-Ray Images](https://mlsanity.com/paper/2110.10381/) (2021)
- [Distribution Estimation to Automate Transformation Policies for Self-Supervision](https://mlsanity.com/paper/2111.12265/) (2021)

## Abstract (arXiv)

Recent self-supervised learning approaches focus on using a few thousand data points to learn policies for high-level, low-dimensional action spaces. However, scaling this framework for high-dimensional control require either scaling up the data collection efforts or using a clever sampling strategy for training. We present a novel approach - Curriculum Accelerated Self-Supervised Learning (CASSL) - to train policies that map visual information to high-level, higher- dimensional action spaces. CASSL orders the sampling of training data based on control dimensions: the learning and sampling are focused on few control parameters before other parameters. The right curriculum for learning is suggested by variance-based global sensitivity analysis of the control space. We apply our CASSL framework to learning how to grasp using an adaptive, underactuated multi-fingered gripper, a challenging system to control. Our experimental results indicate that CASSL provides significant improvement and generalization compared to baseline methods such as staged curriculum learning (8% increase) and complete end-to-end learning with random exploration (14% improvement) tested on a set of novel objects.