# Regularizing CNN Transfer Learning with Randomised Regression

## TL;DR

This paper introduces Pseudo-task Regularization (PtR), a novel CNN regularization method for transfer learning that improves accuracy without extra data or domain knowledge by using pseudo-regression targets during fine-tuning.

## Contribution

It proposes a new regularization approach that does not rely on concrete tasks or additional annotations, dynamically balancing regularization strength during training.

## Key findings

- PtR achieves equal or better accuracy than state-of-the-art methods.
- The method does not require extra data or domain expertise.
- Dynamic regularization maintains effective training balance.

## Abstract

This paper is about regularizing deep convolutional networks (CNNs) based on an adaptive framework for transfer learning with limited training data in the target domain. Recent advances of CNN regularization in this context are commonly due to the use of additional regularization objectives. They guide the training away from the target task using some forms of concrete tasks. Unlike those related approaches, we suggest that an objective without a concrete goal can still serve well as a regularized. In particular, we demonstrate Pseudo-task Regularization (PtR) which dynamically regularizes a network by simply attempting to regress image representations to pseudo-regression targets during fine-tuning. That is, a CNN is efficiently regularized without additional resources of data or prior domain expertise. In sum, the proposed PtR provides: a) an alternative for network regularization without dependence on the design of concrete regularization objectives or extra annotations; b) a dynamically adjusted and maintained strength of regularization effect by balancing the gradient norms between objectives on-line. Through numerous experiments, surprisingly, the improvements on classification accuracy by PtR are shown greater or on a par to the recent state-of-the-art methods.