Deep Transfer Learning with Joint Adaptation Networks

Abstract

Deep networks have been successfully applied to learn transferable features for adapting models from a source domain to a different target domain. In this paper, we present joint adaptation networks (JAN), which learn a transfer network by aligning the joint distributions of multiple domain-specific layers across domains based on a joint maximum mean discrepancy (JMMD) criterion. Adversarial training strategy is adopted to maximize JMMD such that the distributions of the source and target domains are made more distinguishable. Learning can be performed by stochastic gradient descent with the gradients computed by back-propagation in linear-time. Experiments testify that our model yields state of the art results on standard datasets.

34th International Conference on Machine Learning

References in corpus (5)

• Natural Language Processing (almost) from Scratch
• Learning Transferable Features with Deep Adaptation Networks
• Deep Domain Confusion: Maximizing for Domain Invariance
• Domain Separation Networks
• Adversarial Discriminative Domain Adaptation

Cited by in corpus (42)

• Deep Visual Domain Adaptation: A Survey
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• Deep Transfer Learning for Person Re-identification
• Learning Disentangled Semantic Representation for Domain Adaptation
• Causally Regularized Learning with Agnostic Data Selection Bias
• Sample-to-Sample Correspondence for Unsupervised Domain Adaptation
• Label Efficient Learning of Transferable Representations across Domains and Tasks
• Support and Invertibility in Domain-Invariant Representations
• A Survey on Deep Transfer Learning
• Unsupervised Domain Adaptation of Black-Box Source Models
• Multi-target Unsupervised Domain Adaptation without Exactly Shared Categories
• Racial Faces in-the-Wild: Reducing Racial Bias by Information Maximization Adaptation Network
• RAIN: RegulArization on Input and Network for Black-Box Domain Adaptation
• Cluster Alignment with a Teacher for Unsupervised Domain Adaptation
• Importance Weighted Adversarial Nets for Partial Domain Adaptation
• Mutual Alignment Transfer Learning
• Domain-Invariant Adversarial Learning for Unsupervised Domain Adaption
• Learning Factorized Representations for Open-set Domain Adaptation
• Adversarial Alignment of Class Prediction Uncertainties for Domain Adaptation
• Domain Alignment with Triplets
• Dual Adversarial Domain Adaptation
• Deep causal representation learning for unsupervised domain adaptation
• Beyond Fine Tuning: A Modular Approach to Learning on Small Data
• Detection of Paroxysmal Atrial Fibrillation using Attention-based Bidirectional Recurrent Neural Networks
• Learning Transferable UAV for Forest Visual Perception
• Deep Domain Adaptation by Geodesic Distance Minimization
• Domain-Specific Suppression for Adaptive Object Detection
• Domain Adaptation with Randomized Expectation Maximization
• Deep Variational Transfer: Transfer Learning through Semi-supervised Deep Generative Models
• Zero-Annotation Object Detection with Web Knowledge Transfer
• Energy Confused Adversarial Metric Learning for Zero-Shot Image Retrieval and Clustering
• Hidden Covariate Shift: A Minimal Assumption For Domain Adaptation
• Unsupervised Domain Adaptation using Generative Models and Self-ensembling
• Generalizing to the Open World: Deep Visual Odometry with Online Adaptation
• Range Adaptation for 3D Object Detection in LiDAR
• Unsupervised Open Domain Recognition by Semantic Discrepancy Minimization
• A Statistical Test for Joint Distributions Equivalence
• Dynamic Adaptation on Non-Stationary Visual Domains
• Parameter Reference Loss for Unsupervised Domain Adaptation
• Deep Transfer Learning for EEG-based Brain Computer Interface
• Transferable Semantic Augmentation for Domain Adaptation
• Learning Where to Attend Like a Human Driver