# Incremental multi-domain learning with network latent tensor factorization

## TL;DR

This paper introduces a tensor factorization approach for incremental multi-domain learning that efficiently adapts models to new tasks without forgetting, using fewer parameters and leveraging correlations across layers.

## Contribution

The proposed method uses low-rank Tucker tensor structure to enable incremental learning across domains, reducing parameters and improving performance over previous methods.

## Key findings

- Achieves 7.5x parameter reduction on Visual Decathlon datasets
- Maintains competitive accuracy and Decathlon scores
- Leverages tensor structure for better layer correlation modeling

## Abstract

The prominence of deep learning, large amount of annotated data and increasingly powerful hardware made it possible to reach remarkable performance for supervised classification tasks, in many cases saturating the training sets. However the resulting models are specialized to a single very specific task and domain. Adapting the learned classification to new domains is a hard problem due to at least three reasons: (1) the new domains and the tasks might be drastically different; (2) there might be very limited amount of annotated data on the new domain and (3) full training of a new model for each new task is prohibitive in terms of computation and memory, due to the sheer number of parameters of deep CNNs. In this paper, we present a method to learn new-domains and tasks incrementally, building on prior knowledge from already learned tasks and without catastrophic forgetting. We do so by jointly parametrizing weights across layers using low-rank Tucker structure. The core is task agnostic while a set of task specific factors are learnt on each new domain. We show that leveraging tensor structure enables better performance than simply using matrix operations. Joint tensor modelling also naturally leverages correlations across different layers. Compared with previous methods which have focused on adapting each layer separately, our approach results in more compact representations for each new task/domain. We apply the proposed method to the 10 datasets of the Visual Decathlon Challenge and show that our method offers on average about 7.5x reduction in number of parameters and competitive performance in terms of both classification accuracy and Decathlon score.