Budget-Aware Pruning: Handling Multiple Domains with Less Parameters

TL;DR

This paper introduces a method for pruning multi-domain deep learning models to reduce parameters and computational costs, enabling efficient deployment on resource-limited devices while maintaining performance.

Contribution

It proposes a novel pruning technique that encourages shared filter usage across multiple domains, effectively reducing model size without sacrificing accuracy.

Findings

Achieves significant parameter reduction across multiple domains.

Maintains comparable classification performance with fewer filters.

Handles multiple domains with lower computational complexity than baseline.

Abstract

Deep learning has achieved state-of-the-art performance on several computer vision tasks and domains. Nevertheless, it still has a high computational cost and demands a significant amount of parameters. Such requirements hinder the use in resource-limited environments and demand both software and hardware optimization. Another limitation is that deep models are usually specialized into a single domain or task, requiring them to learn and store new parameters for each new one. Multi-Domain Learning (MDL) attempts to solve this problem by learning a single model capable of performing well in multiple domains. Nevertheless, the models are usually larger than the baseline for a single domain. This work tackles both of these problems: our objective is to prune models capable of handling multiple domains according to a user-defined budget, making them more computationally affordable while keeping a similar classification performance. We achieve this by encouraging all domains to use a similar subset of filters from the baseline model, up to the amount defined by the user's budget. Then, filters that are not used by any domain are pruned from the network. The proposed approach innovates by better adapting to resource-limited devices while being one of the few works that handles multiple domains at test time with fewer parameters and lower computational complexity than the baseline model for a single domain.