Knowledge-Adaptation Priors

TL;DR

K-priors are a novel approach that significantly reduces retraining costs in machine learning by enabling quick adaptation to new tasks using a combination of weight and function-space priors, matching full retraining performance with minimal data.

Contribution

This paper introduces K-priors, a new method combining weight and function-space priors to facilitate rapid and accurate model adaptation with minimal data, generalizing many existing strategies.

Findings

Achieves adaptation performance similar to full retraining

Requires training on only a few past examples

Can recover the exact retrained model with simple gradient methods

Abstract

Humans and animals have a natural ability to quickly adapt to their surroundings, but machine-learning models, when subjected to changes, often require a complete retraining from scratch. We present Knowledge-adaptation priors (K-priors) to reduce the cost of retraining by enabling quick and accurate adaptation for a wide-variety of tasks and models. This is made possible by a combination of weight and function-space priors to reconstruct the gradients of the past, which recovers and generalizes many existing, but seemingly-unrelated, adaptation strategies. Training with simple first-order gradient methods can often recover the exact retrained model to an arbitrary accuracy by choosing a sufficiently large memory of the past data. Empirical results show that adaptation with K-priors achieves performance similar to full retraining, but only requires training on a handful of past examples.