Curriculum Graph Co-Teaching for Multi-Target Domain Adaptation (2021-04-01T00:00:00.000000Z)

TL;DR

This paper develops a co-teaching strategy with the dual classifier head that is assisted by curriculum learning to obtain more reliable pseudo-labels and proposes Domain-aware Curriculum Learning (DCL), a sequential adaptation strategy that first adapts on the easier target domains, followed by the harder ones.

Abstract

In this paper we address multi-target domain adaptation (MTDA), where given one labeled source dataset and multiple unlabeled target datasets that differ in data distributions, the task is to learn a robust predictor for all the target domains. We identify two key aspects that can help to alleviate multiple domain-shifts in the MTDA: feature aggregation and curriculum learning. To this end, we propose Curriculum Graph Co-Teaching (CGCT) that uses a dual classifier head, with one of them being a graph convolutional network (GCN) which aggregates features from similar samples across the domains. To prevent the classifiers from over-fitting on its own noisy pseudo-labels we develop a co-teaching strategy with the dual classifier head that is assisted by curriculum learning to obtain more reliable pseudo-labels. Furthermore, when the domain labels are available, we propose Domain-aware Curriculum Learning (DCL), a sequential adaptation strategy that first adapts on the easier target domains, followed by the harder ones. We experimentally demonstrate the effectiveness of our proposed frameworks on several benchmarks and advance the state-of-the-art in the MTDA by large margins (e.g. +5.6% on the DomainNet).

Authors

N. Sebe

21 papers

E. Ricci

11 papers

Zhun Zhong

8 papers

TL;DR

Abstract

Authors

References58 items

Generative Adversarial Networks

The Bottleneck

Heterogeneous Graph Attention Network for Unsupervised Multiple-Target Domain Adaptation

Overcoming Negative Transfer: A Survey

Unsupervised Multi-Target Domain Adaptation Through Knowledge Distillation

Curriculum Manager for Source Selection in Multi-Source Domain Adaptation

Progressive Graph Learning for Open-Set Domain Adaptation

TriGAN: image-to-image translation for multi-source domain adaptation

Minimum Class Confusion for Versatile Domain Adaptation

Unsupervised Domain Adaptation Using Full-Feature Whitening and Colouring

Open Compound Domain Adaptation

Curriculum based Dropout Discriminator for Domain Adaptation

Transferable Curriculum for Weakly-Supervised Domain Adaptation

GCAN: Graph Convolutional Adversarial Network for Unsupervised Domain Adaptation

Blending-Target Domain Adaptation by Adversarial Meta-Adaptation Networks

Domain-Specific Batch Normalization for Unsupervised Domain Adaptation

Domain Agnostic Learning with Disentangled Representations

Unsupervised Domain Adaptation Using Feature-Whitening and Consensus Loss

Moment Matching for Multi-Source Domain Adaptation

Progressive Feature Alignment for Unsupervised Domain Adaptation

Unsupervised Multi-Target Domain Adaptation: An Information Theoretic Approach

Re-weighted Adversarial Adaptation Network for Unsupervised Domain Adaptation

Boosting Domain Adaptation by Discovering Latent Domains

Co-teaching: Robust training of deep neural networks with extremely noisy labels

Deep Cocktail Network: Multi-source Unsupervised Domain Adaptation with Category Shift

A DIRT-T Approach to Unsupervised Domain Adaptation

Maximum Classifier Discrepancy for Unsupervised Domain Adaptation

CyCADA: Cycle-Consistent Adversarial Domain Adaptation

Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation

Deeper, Broader and Artier Domain Generalization

Partial Transfer Learning with Selective Adversarial Networks

Deep Hashing Network for Unsupervised Domain Adaptation

Self-ensembling for visual domain adaptation

Conditional Adversarial Domain Adaptation

From Source to Target and Back: Symmetric Bi-Directional Adaptive GAN

AutoDIAL: Automatic Domain Alignment Layers

Generate to Adapt: Aligning Domains Using Generative Adversarial Networks

Just DIAL: DomaIn Alignment Layers for Unsupervised Domain Adaptation

Adversarial Discriminative Domain Adaptation

Synthetic to Real Adaptation with Generative Correlation Alignment Networks

Semi-Supervised Classification with Graph Convolutional Networks

Deep CORAL: Correlation Alignment for Deep Domain Adaptation

Coupled Generative Adversarial Networks

Deep Transfer Learning with Joint Adaptation Networks

Revisiting Batch Normalization For Practical Domain Adaptation

Deep Residual Learning for Image Recognition

Domain-Adversarial Training of Neural Networks

Deep Transfer Network: Unsupervised Domain Adaptation

Learning Transferable Features with Deep Adaptation Networks

Deep Domain Confusion: Maximizing for Domain Invariance

Unbiased look at dataset bias

Adapting Visual Category Models to New Domains

Curriculum learning

A new model for learning in graph domains

The Elements of Statistical Learning

A Comprehensive Survey on Graph Neural Networks

Reading Digits in Natural Images with Unsupervised Feature Learning

Gradient-based learning applied to document recognition

Field of Study

Journal Information

Name

Page

Venue Information

Name

Type

URL

Alternate Names