Gianluigi Silvestri

Title?

Abstract

In this work, a neural architecture search algorithm for multi-task learning is proposed. Given any dataset and tasks group, the method aims to find the optimal way of sharing layers among tasks in convolutional neural networks. A search space suited to multi-task learning is designed, and a novel strategy to rank different Pareto optimal solutions is developed. The core of the algo- rithm is an adaptation of a state-of-the-art neural architecture search strategy. Experimental results on the Cityscapes dataset, on the tasks of semantic seg- mentation and depth estimation, do not provide the expected results. Despite the lack of stable results, this work lays down the fundamentals to further de- velop novel multi-task neural architecture search methods.