Tailoring: encoding inductive biases by optimizing unsupervised objectives at prediction time

TL;DR

This paper introduces tailoring and meta-tailoring methods that optimize neural networks at prediction time using unsupervised objectives, improving generalization by customizing models for each input based on inductive biases.

Contribution

It proposes a novel approach to encode inductive biases through input-specific fine-tuning and meta-learning, addressing generalization gaps of auxiliary losses.

Findings

Tailoring improves prediction accuracy on diverse tasks.

Meta-tailoring enhances model robustness and generalization.

Methods are validated empirically across multiple examples.

Abstract

From CNNs to attention mechanisms, encoding inductive biases into neural networks has been a fruitful source of improvement in machine learning. Adding auxiliary losses to the main objective function is a general way of encoding biases that can help networks learn better representations. However, since auxiliary losses are minimized only on training data, they suffer from the same generalization gap as regular task losses. Moreover, by adding a term to the loss function, the model optimizes a different objective than the one we care about. In this work we address both problems: first, we take inspiration from \textit{transductive learning} and note that after receiving an input but before making a prediction, we can fine-tune our networks on any unsupervised loss. We call this process {\em tailoring}, because we customize the model to each input to ensure our prediction satisfies the inductive bias. Second, we formulate {\em meta-tailoring}, a nested optimization similar to that in meta-learning, and train our models to perform well on the task objective after adapting them using an unsupervised loss. The advantages of tailoring and meta-tailoring are discussed theoretically and demonstrated empirically on a diverse set of examples.