Self-Guided Masked Autoencoders for Domain-Agnostic Self-Supervised Learning

TL;DR

This paper introduces Self-guided Masked Autoencoders (SMA), a domain-agnostic self-supervised learning method that learns masks without domain-specific assumptions, achieving state-of-the-art results across diverse scientific data modalities.

Contribution

SMA is the first fully domain-agnostic masked modeling approach that learns masks via attention, removing the need for domain-specific augmentations or assumptions.

Findings

Achieves state-of-the-art performance on protein biology benchmarks

Effective in chemical property prediction tasks

Performs well in particle physics data analysis

Abstract

Self-supervised learning excels in learning representations from large amounts of unlabeled data, demonstrating success across multiple data modalities. Yet, extending self-supervised learning to new modalities is non-trivial because the specifics of existing methods are tailored to each domain, such as domain-specific augmentations which reflect the invariances in the target task. While masked modeling is promising as a domain-agnostic framework for self-supervised learning because it does not rely on input augmentations, its mask sampling procedure remains domain-specific. We present Self-guided Masked Autoencoders (SMA), a fully domain-agnostic masked modeling method. SMA trains an attention based model using a masked modeling objective, by learning masks to sample without any domain-specific assumptions. We evaluate SMA on three self-supervised learning benchmarks in protein biology, chemical property prediction, and particle physics. We find SMA is capable of learning representations without domain-specific knowledge and achieves state-of-the-art performance on these three benchmarks.