MILO: Model-Agnostic Subset Selection Framework for Efficient Model Training and Tuning

TL;DR

MILO is a model-agnostic subset selection framework that accelerates deep network training and hyperparameter tuning by pre-selecting representative data subsets, achieving significant speedups without sacrificing performance.

Contribution

MILO introduces a pre-processing subset selection method that is model-agnostic and enables faster training and tuning by decoupling subset selection from model training.

Findings

MILO trains models 3x to 10x faster.

MILO tunes hyperparameters 20x to 75x faster.

MILO maintains comparable model performance.

Abstract

Training deep networks and tuning hyperparameters on large datasets is computationally intensive. One of the primary research directions for efficient training is to reduce training costs by selecting well-generalizable subsets of training data. Compared to simple adaptive random subset selection baselines, existing intelligent subset selection approaches are not competitive due to the time-consuming subset selection step, which involves computing model-dependent gradients and feature embeddings and applies greedy maximization of submodular objectives. Our key insight is that removing the reliance on downstream model parameters enables subset selection as a pre-processing step and enables one to train multiple models at no additional cost. In this work, we propose MILO, a model-agnostic subset selection framework that decouples the subset selection from model training while enabling superior model convergence and performance by using an easy-to-hard curriculum. Our empirical results indicate that MILO can train models $3\times - 10 \times$ faster and tune hyperparameters $20\times - 75 \times$ faster than full-dataset training or tuning without compromising performance.