Medusa: Universal Feature Learning via Attentional Multitasking

by   Jaime Spencer, et al.
University of Surrey

Recent approaches to multi-task learning (MTL) have focused on modelling connections between tasks at the decoder level. This leads to a tight coupling between tasks, which need retraining if a new task is inserted or removed. We argue that MTL is a stepping stone towards universal feature learning (UFL), which is the ability to learn generic features that can be applied to new tasks without retraining. We propose Medusa to realize this goal, designing task heads with dual attention mechanisms. The shared feature attention masks relevant backbone features for each task, allowing it to learn a generic representation. Meanwhile, a novel Multi-Scale Attention head allows the network to better combine per-task features from different scales when making the final prediction. We show the effectiveness of Medusa in UFL (+13.18 while maintaining MTL performance and being 25 approaches.


page 1

page 3

page 6


InvPT++: Inverted Pyramid Multi-Task Transformer for Visual Scene Understanding

Multi-task scene understanding aims to design models that can simultaneo...

Deep Multi-Task Augmented Feature Learning via Hierarchical Graph Neural Network

Deep multi-task learning attracts much attention in recent years as it a...

MTI-Net: Multi-Scale Task Interaction Networks for Multi-Task Learning

In this paper, we highlight the importance of considering task interacti...

Multi-task Learning for Human Settlement Extent Regression and Local Climate Zone Classification

Human Settlement Extent (HSE) and Local Climate Zone (LCZ) maps are both...

MTLSegFormer: Multi-task Learning with Transformers for Semantic Segmentation in Precision Agriculture

Multi-task learning has proven to be effective in improving the performa...

Representative Task Self-selection for Flexible Clustered Lifelong Learning

Consider the lifelong learning paradigm whose objective is to learn a se...

Counting with Adaptive Auxiliary Learning

This paper proposes an adaptive auxiliary task learning based approach f...

Please sign up or login with your details

Forgot password? Click here to reset