Learning Multiple Tasks with Multilinear Relationship Networks

Abstract

Deep networks trained on large-scale data can learn transferable features to promote learning multiple tasks. Since deep features eventually transition from general to specific along deep networks, a fundamental problem of multi-task learning is how to exploit the task relatedness underlying parameter tensors and improve feature transferability in the multiple task-specific layers. This paper presents Multilinear Relationship Networks (MRN) that discover the task relationships based on novel tensor normal priors over parameter tensors of multiple task-specific layers in deep convolutional networks. By jointly learning transferable features and multilinear relationships of tasks and features, MRN is able to alleviate the dilemma of negative-transfer in the feature layers and under-transfer in the classifier layer. Experiments show that MRN yields state-of-the-art results on three multi-task learning datasets.

NIPS 2017

Cited by in corpus (49)

• Multi-Task Learning for Dense Prediction Tasks: A Survey
• A Survey on Multi-Task Learning
• Deep Reinforcement Learning: An Overview
• Multi-Task Learning with Deep Neural Networks: A Survey
• Joint predictions of multi-modal ride-hailing demands: a deep multi-task multigraph learning-based approach
• MEDIC: A Multi-Task Learning Dataset for Disaster Image Classification
• Domain Generalization via Optimal Transport with Metric Similarity Learning
• Reinforcement Learning and Deep Learning based Lateral Control for Autonomous Driving
• Regularizing Deep Multi-Task Networks using Orthogonal Gradients
• Joint Learning of Salient Object Detection, Depth Estimation and Contour Extraction
• Pseudo-task Augmentation: From Deep Multitask Learning to Intratask Sharing---and Back
• Controllable Pareto Multi-Task Learning
• Where is the Model Looking At?--Concentrate and Explain the Network Attention
• Cross-dataset Training for Class Increasing Object Detection
• Empirical Analysis of Multi-Task Learning for Reducing Model Bias in Toxic Comment Detection
• Geometry Uncertainty Projection Network for Monocular 3D Object Detection
• Auxiliary Task Reweighting for Minimum-data Learning
• A Unified Model with Structured Output for Fashion Images Classification
• MLM: A Benchmark Dataset for Multitask Learning with Multiple Languages and Modalities
• Branched Multi-Task Networks: Deciding What Layers To Share
• Modeling the Sequential Dependence among Audience Multi-step Conversions with Multi-task Learning in Targeted Display Advertising
• Spatio-Temporal Momentum: Jointly Learning Time-Series and Cross-Sectional Strategies
• A Simple General Approach to Balance Task Difficulty in Multi-Task Learning
• Auxiliary Learning for Deep Multi-task Learning
• Weakly Supervised Open-set Domain Adaptation by Dual-domain Collaboration
• Empirical Study of Multi-Task Hourglass Model for Semantic Segmentation Task
• CrowdTransfer: Enabling Crowd Knowledge Transfer in AIoT Community
• STG-MTL: Scalable Task Grouping for Multi-Task Learning Using Data Map
• NIPS - Not Even Wrong? A Systematic Review of Empirically Complete Demonstrations of Algorithmic Effectiveness in the Machine Learning and Artificial Intelligence Literature
• Multi-Task Variational Information Bottleneck
• Deep Multi-Task Learning via Generalized Tensor Trace Norm
• Cartoon Face Recognition: A Benchmark Dataset
• Byzantine Resilient Distributed Multi-Task Learning
• Cross-Task Consistency Learning Framework for Multi-Task Learning
• Modular Universal Reparameterization: Deep Multi-task Learning Across Diverse Domains
• Multi-task Learning by Leveraging the Semantic Information
• Tasks Structure Regularization in Multi-Task Learning for Improving Facial Attribute Prediction
• TPG-DNN: A Method for User Intent Prediction Based on Total Probability Formula and GRU Loss with Multi-task Learning
• Multi-Task Feature Learning for Knowledge Graph Enhanced Recommendation
• Adaptive Activation Network and Functional Regularization for Efficient and Flexible Deep Multi-Task Learning
• Network Clustering for Multi-task Learning
• Dropping Networks for Transfer Learning
• Multiple Classification with Split Learning
• Multi-Task Neural Processes
• Fast Line Search for Multi-Task Learning
• Task-Assisted Domain Adaptation with Anchor Tasks
• Deep Virtual Networks for Memory Efficient Inference of Multiple Tasks
• Constructing Multiple Tasks for Augmentation: Improving Neural Image Classification With K-means Features
• Disentangling Transfer and Interference in Multi-Domain Learning