A Survey on Multi-Task Learning

Abstract

Multi-Task Learning (MTL) is a learning paradigm in machine learning and its aim is to leverage useful information contained in multiple related tasks to help improve the generalization performance of all the tasks. In this paper, we give a survey for MTL from the perspective of algorithmic modeling, applications and theoretical analyses. For algorithmic modeling, we give a definition of MTL and then classify different MTL algorithms into five categories, including feature learning approach, low-rank approach, task clustering approach, task relation learning approach and decomposition approach as well as discussing the characteristics of each approach. In order to improve the performance of learning tasks further, MTL can be combined with other learning paradigms including semi-supervised learning, active learning, unsupervised learning, reinforcement learning, multi-view learning and graphical models. When the number of tasks is large or the data dimensionality is high, we review online, parallel and distributed MTL models as well as dimensionality reduction and feature hashing to reveal their computational and storage advantages. Many real-world applications use MTL to boost their performance and we review representative works in this paper. Finally, we present theoretical analyses and discuss several future directions for MTL.

Accepted by IEEE Transactions on Knowledge and Data Engineering

References in corpus (8)

• IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures
• Clustered Multi-Task Learning: A Convex Formulation
• Multi-Stage Multi-Task Feature Learning
• Union Support Recovery in Multi-task Learning
• Flexible Modeling of Latent Task Structures in Multitask Learning
• Robust Online Multi-Task Learning with Correlative and Personalized Structures
• Learning to Multitask
• Learning Multiple Visual Tasks while Discovering their Structure

Cited by in corpus (183)

• A continual learning survey: Defying forgetting in classification tasks
• A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection
• Multi-Task Learning for Dense Prediction Tasks: A Survey
• Generalizing from a Few Examples: A Survey on Few-Shot Learning
• Multi-Task Learning with Deep Neural Networks: A Survey
• Precision Health Data: Requirements, Challenges and Existing Techniques for Data Security and Privacy
• Embracing Imperfect Datasets: A Review of Deep Learning Solutions for Medical Image Segmentation
• Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges
• Multi-Task Deep Neural Networks for Natural Language Understanding
• Clinical Concept Extraction: a Methodology Review
• Improving Multi-Task Deep Neural Networks via Knowledge Distillation for Natural Language Understanding
• Distributed intelligence on the Edge-to-Cloud Continuum: A systematic literature review
• Hybrid Block Successive Approximation for One-Sided Non-Convex Min-Max Problems: Algorithms and Applications
• Face Behavior a la carte: Expressions, Affect and Action Units in a Single Network
• Joint predictions of multi-modal ride-hailing demands: a deep multi-task multigraph learning-based approach
• Image to Images Translation for Multi-Task Organ Segmentation and Bone Suppression in Chest X-Ray Radiography
• Privacy-preserving Artificial Intelligence Techniques in Biomedicine
• Review: Deep Learning in Electron Microscopy
• Ontology-driven weak supervision for clinical entity classification in electronic health records
• Flexible End-to-End Dialogue System for Knowledge Grounded Conversation
• Opportunities and Challenges in Code Search Tools
• UNet++: Redesigning Skip Connections to Exploit Multiscale Features in Image Segmentation
• Cross-Domain Complementary Learning Using Pose for Multi-Person Part Segmentation
• MedGCN: Medication recommendation and lab test imputation via graph convolutional networks
• Multi-modal Deep Analysis for Multimedia
• Smart City Development with Urban Transfer Learning
• Towards Image-based Automatic Meter Reading in Unconstrained Scenarios: A Robust and Efficient Approach
• DEFT: Detection Embeddings for Tracking
• Accelerating Self-Play Learning in Go
• Multi-task, multi-label and multi-domain learning with residual convolutional networks for emotion recognition
• Learning joint segmentation of tissues and brain lesions from task-specific hetero-modal domain-shifted datasets
• Simultaneous Modeling of Multiple Complications for Risk Profiling in Diabetes Care
• Triangle-Net: Towards Robustness in Point Cloud Learning
• The description of giant dipole resonance key parameters with multitask neural networks
• FastPose: Towards Real-time Pose Estimation and Tracking via Scale-normalized Multi-task Networks
• Table2Charts: Recommending Charts by Learning Shared Table Representations
• The Multi-Agent Behavior Dataset: Mouse Dyadic Social Interactions
• Training Variational Networks with Multi-Domain Simulations: Speed-of-Sound Image Reconstruction
• AP-MTL: Attention Pruned Multi-task Learning Model for Real-time Instrument Detection and Segmentation in Robot-assisted Surgery
• Multi-Task Zipping via Layer-wise Neuron Sharing
• COVID-MTL: Multitask Learning with Shift3D and Random-weighted Loss for Automated Diagnosis and Severity Assessment of COVID-19
• ProcessTransformer: Predictive Business Process Monitoring with Transformer Network
• Learning the Pareto Front with Hypernetworks
• AutoLoss: Learning Discrete Schedules for Alternate Optimization
• Efficiently Identifying Task Groupings for Multi-Task Learning
• Modeling Prosodic Phrasing with Multi-Task Learning in Tacotron-based TTS
• Variable Selection and Task Grouping for Multi-Task Learning
• Reinforcement Learning for Slate-based Recommender Systems: A Tractable Decomposition and Practical Methodology
• Controllable Pareto Multi-Task Learning
• Multitask Learning for Blackmarket Tweet Detection
• Extending Unsupervised Neural Image Compression With Supervised Multitask Learning
• Graph Representation Learning Beyond Node and Homophily
• Reinforcement Learning Based Emotional Editing Constraint Conversation Generation
• H3DNet: 3D Object Detection Using Hybrid Geometric Primitives
• Multitask Balanced and Recalibrated Network for Medical Code Prediction
• Addressing the Cold-Start Problem in Outfit Recommendation Using Visual Preference Modelling
• A Theoretical Perspective on Differentially Private Federated Multi-task Learning
• A multi-task deep learning model for the classification of Age-related Macular Degeneration
• Fast-adapting and Privacy-preserving Federated Recommender System
• Prediction of hierarchical time series using structured regularization and its application to artificial neural networks
• Continual Learning in Neural Networks
• MLM: A Benchmark Dataset for Multitask Learning with Multiple Languages and Modalities
• Multitask Recalibrated Aggregation Network for Medical Code Prediction
• Feature Partitioning for Efficient Multi-Task Architectures
• Multi-Task Learning by Deep Collaboration and Application in Facial Landmark Detection
• Compression-Based Regularization with an Application to Multi-Task Learning
• Improving Vessel Segmentation with Multi-Task Learning and Auxiliary Data Available Only During Model Training
• A Comprehensive Evaluation of Multi-task Learning and Multi-task Pre-training on EHR Time-series Data
• Fully Decentralized Joint Learning of Personalized Models and Collaboration Graphs
• Transfer Meets Hybrid: A Synthetic Approach for Cross-Domain Collaborative Filtering with Text
• Learning to Multitask
• Multi-Objective Meta Learning
• An Improved Event-Independent Network for Polyphonic Sound Event Localization and Detection
• Multi-task learning from fixed-wing UAV images for 2D/3D city modeling
• Aggregate or Not? Exploring Where to Privatize in DNN Based Federated Learning Under Different Non-IID Scenes
• Multi-script Handwritten Digit Recognition Using Multi-task Learning
• Aggregating From Multiple Target-Shifted Sources
• Patient-independent Epileptic Seizure Prediction using Deep Learning Models
• Metadata-based Multi-Task Bandits with Bayesian Hierarchical Models
• Measuring and Harnessing Transference in Multi-Task Learning
• Review Helpfulness Prediction with Embedding-Gated CNN
• SpeechNet: A Universal Modularized Model for Speech Processing Tasks
• Transformer-based Multi-task Learning for Disaster Tweet Categorisation
• Multi-Label Transfer Learning for Multi-Relational Semantic Similarity
• Representative Task Self-selection for Flexible Clustered Lifelong Learning
• Multi-Task Variational Information Bottleneck
• Task Embedding Temporal Convolution Networks for Transfer Learning Problems in Renewable Power Time-Series Forecast
• Ranking Micro-Influencers: a Novel Multi-Task Learning and Interpretable Framework
• Weakly Supervised Multi-task Learning for Concept-based Explainability
• Weakly Supervised Open-set Domain Adaptation by Dual-domain Collaboration
• Empirical Study of Multi-Task Hourglass Model for Semantic Segmentation Task
• Multitask and Transfer Learning for Autotuning Exascale Applications
• Cartoon Face Recognition: A Benchmark Dataset
• Low Resource Multi-Task Sequence Tagging -- Revisiting Dynamic Conditional Random Fields
• Self-supervised Auxiliary Learning for Graph Neural Networks via Meta-Learning
• Meta-Learning with MAML on Trees
• Multitask Learning for Class-Imbalanced Discourse Classification
• Personalized Federated Learning of Driver Prediction Models for Autonomous Driving
• Private Multi-Task Learning: Formulation and Applications to Federated Learning
• Auxiliary Learning by Implicit Differentiation
• A Unified Multi-scale and Multi-task Learning Framework for Driver Behaviors Reasoning
• Frosting Weights for Better Continual Training
• Customizing Sequence Generation with Multi-Task Dynamical Systems
• Improving Aspect-Level Sentiment Analysis with Aspect Extraction
• Follow the bisector: a simple method for multi-objective optimization
• Multi-Task Adversarial Attack
• Investigating the Reordering Capability in CTC-based Non-Autoregressive End-to-End Speech Translation
• Hierarchical Bayesian Bandits
• SLAW: Scaled Loss Approximate Weighting for Efficient Multi-Task Learning
• A Deep Learning Framework for Lifelong Machine Learning
• Learning Multiple Dense Prediction Tasks from Partially Annotated Data
• DeepCSO: Forecasting of Combined Sewer Overflow at a Citywide Level using Multi-task Deep Learning
• Decision Making Problems with Funnel Structure: A Multi-Task Learning Approach with Application to Email Marketing Campaigns
• An Orthogonal-SGD based Learning Approach for MIMO Detection under Multiple Channel Models
• Aggregative Self-Supervised Feature Learning from a Limited Sample
• Multi-task Reinforcement Learning in Reproducing Kernel Hilbert Spaces via Cross-learning
• Machine Learning in Gamma Astronomy
• Multi-task Prediction of Patient Workload
• Two-Stream Multi-Task Network for Fashion Recognition
• Robustly Optimized and Distilled Training for Natural Language Understanding
• Nested and Balanced Entity Recognition using Multi-Task Learning
• Learning Compositional Neural Programs for Continuous Control
• Multi-Task Feature Learning for Knowledge Graph Enhanced Recommendation
• A Modularized Neural Network with Language-Specific Output Layers for Cross-lingual Voice Conversion
• A Conceptual Framework for Lifelong Learning
• Developing Multi-Task Recommendations with Long-Term Rewards via Policy Distilled Reinforcement Learning
• Using Multitask Learning to Improve 12-Lead Electrocardiogram Classification
• The Extended Dawid-Skene Model: Fusing Information from Multiple Data Schemas
• Cooperative Training and Latent Space Data Augmentation for Robust Medical Image Segmentation
• Few Shot Learning With No Labels
• Multi-Task Learning of Query Intent and Named Entities using Transfer Learning
• Spectral Algorithm for Low-rank Multitask Regression
• Predicting Material Properties Using a 3D Graph Neural Network with Invariant Local Descriptors
• Action Redundancy in Reinforcement Learning
• Efficient Multi-Domain Network Learning by Covariance Normalization
• Improving the Generalization of Meta-learning on Unseen Domains via Adversarial Shift
• Pentagon at MEDIQA 2019: Multi-task Learning for Filtering and Re-ranking Answers using Language Inference and Question Entailment
• Pretrained Language Models for Dialogue Generation with Multiple Input Sources
• Ferrograph image classification
• Multi-task Supervised Learning via Cross-learning
• Neuro-evolutionary Frameworks for Generalized Learning Agents
• Multi-task GANs for Semantic Segmentation and Depth Completion with Cycle Consistency
• Pair-view Unsupervised Graph Representation Learning
• Universal Policies for Software-Defined MDPs
• Software/Hardware Co-design for Multi-modal Multi-task Learning in Autonomous Systems
• Empirical Evaluation of Multi-task Learning in Deep Neural Networks for Natural Language Processing
• Deep Multimodality Model for Multi-task Multi-view Learning
• Context-dependent self-exciting point processes: models, methods, and risk bounds in high dimensions
• A Riemannian gossip approach to subspace learning on Grassmann manifold
• PrivNet: Safeguarding Private Attributes in Transfer Learning for Recommendation
• Multi-Task Kernel Null-Space for One-Class Classification
• On the relationship between multitask neural networks and multitask Gaussian Processes
• A Multi-Task Gradient Descent Method for Multi-Label Learning
• Constructing Multiple Tasks for Augmentation: Improving Neural Image Classification With K-means Features
• TEST_POSITIVE at W-NUT 2020 Shared Task-3: Joint Event Multi-task Learning for Slot Filling in Noisy Text
• Towards Interpretable Multi-Task Learning Using Bilevel Programming
• A Short Review on Data Modelling for Vector Fields
• Distributed Machine Learning for Predictive Analytics in Mobile Edge Computing Based IoT Environments
• Incorporating Image Gradients as Secondary Input Associated with Input Image to Improve the Performance of the CNN Model
• Heterogeneous Representation Learning: A Review
• Linear shrinkage for predicting responses in large-scale multivariate linear regression
• A Self-Supervised Auxiliary Loss for Deep RL in Partially Observable Settings
• One Network Fits All? Modular versus Monolithic Task Formulations in Neural Networks
• Attention-based multi-task learning for speech-enhancement and speaker-identification in multi-speaker dialogue scenario
• SPARTA: Speaker Profiling for ARabic TAlk
• A question-answering system for aircraft pilots' documentation
• Multi-modal, multi-task, multi-attention (M3) deep learning detection of reticular pseudodrusen: towards automated and accessible classification of age-related macular degeneration
• CvS: Classification via Segmentation For Small Datasets
• Bristle: Decentralized Federated Learning in Byzantine, Non-i.i.d. Environments
• Optimal Ensemble Construction for Multi-Study Prediction with Applications to COVID-19 Excess Mortality Estimation
• Fast Line Search for Multi-Task Learning
• Exploring Data Aggregation and Transformations to Generalize across Visual Domains
• Towards Model-informed Precision Dosing with Expert-in-the-loop Machine Learning
• Generalized Gradient Descent is a Hypergraph Functor
• Management of Resource at the Network Edge for Federated Learning
• Multitask Online Mirror Descent
• Towards All-around Knowledge Transferring: Learning From Task-irrelevant Labels
• Indoor Scene Recognition in 3D
• Metric-based Regularization and Temporal Ensemble for Multi-task Learning using Heterogeneous Unsupervised Tasks
• Personalized Attraction Enhanced Sponsored Search with Multi-task Learning
• Composite local low-rank structure in learning drug sensitivity
• Accumulating Knowledge for Lifelong Online Learning
• Affect Estimation in 3D Space Using Multi-Task Active Learning for Regression