Self-supervised learning through the eyes of a child

Abstract

Within months of birth, children develop meaningful expectations about the world around them. How much of this early knowledge can be explained through generic learning mechanisms applied to sensory data, and how much of it requires more substantive innate inductive biases? Addressing this fundamental question in its full generality is currently infeasible, but we can hope to make real progress in more narrowly defined domains, such as the development of high-level visual categories, thanks to improvements in data collecting technology and recent progress in deep learning. In this paper, our goal is precisely to achieve such progress by utilizing modern self-supervised deep learning methods and a recent longitudinal, egocentric video dataset recorded from the perspective of three young children (Sullivan et al., 2020). Our results demonstrate the emergence of powerful, high-level visual representations from developmentally realistic natural videos using generic self-supervised learning objectives.

arXiv commentPublished as a conference paper at NeurIPS 2020; v3 adds a reference, fixes a typo

References in corpus (8)

• A Simple Framework for Contrastive Learning of Visual Representations
• Representation Learning with Contrastive Predictive Coding
• Improved Baselines with Momentum Contrastive Learning
• Learning deep representations by mutual information estimation and maximization
• On the importance of single directions for generalization
• Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICA
• Robustness properties of Facebook's ResNeXt WSL models
• Selectivity considered harmful: evaluating the causal impact of class selectivity in DNNs

Cited by in corpus (7)

• Self-supervised remote sensing feature learning: Learning Paradigms, Challenges, and Future Works
• Measuring and modeling the motor system with machine learning
• Partial success in closing the gap between human and machine vision
• On the surprising similarities between supervised and self-supervised models
• Enabling On-Device Large Language Model Personalization with Self-Supervised Data Selection and Synthesis
• Building Human-like Communicative Intelligence: A Grounded Perspective
• Enabling On-Device Self-Supervised Contrastive Learning With Selective Data Contrast