1. Awesome Transfer Learning Papers

Let’s read some awesome transfer learning / domain adaptation papers.
这里收录了迁移学习各个研究领域的最新文章。

• 1. Awesome Transfer Learning Papers
• 1.1. General Transfer Learning (普通迁移学习)
• 1.1.1. Theory (理论)
• 1.1.2. Others (其他)

• 1.2. Domain Adaptation (领域自适应)
• 1.2.1. Traditional Methods (传统迁移方法)
• 1.2.2. Deep / Adversarial Methods (深度/对抗迁移方法)

• 1.3. Domain Generalization
• 1.4. Multi-source Transfer Learning (多源迁移学习)
• 1.5. Heterogeneous Transfer Learning (异构迁移学习)
• 1.6. Online Transfer Learning (在线迁移学习)
• 1.7. Zero-shot / Few-shot Learning
• 1.7.1. Zero-shot Learning based on Data Synthesis (基于样本生成的零样本学习)

• 1.8. Deep Transfer Learning (深度迁移学习)
• 1.8.1. Non-Adversarial Transfer Learning (非对抗深度迁移)
• 1.8.2. Deep Adversarial Transfer Learning (对抗迁移学习)

• 1.9. Multi-task Learning (多任务学习)
• 1.10. Transfer Reinforcement Learning (强化迁移学习)
• 1.11. Transfer Metric Learning (迁移度量学习)
• 1.12. Transitive Transfer Learning (传递迁移学习)
• 1.13. Lifelong Learning (终身迁移学习)
• 1.14. Negative Transfer (负迁移)
• 1.15. Transfer Learning Applications (应用)

1.1. General Transfer Learning (普通迁移学习)

1.1.1. Theory (理论)

• 20200220 Butterfly: One-step Approach towards Wildly Unsupervised Domain Adaptation

• Noisy domain adaptation

• 用于噪声环境中的domain adaptation的方法

• 20210127 A Unified Joint Maximum Mean Discrepancy for Domain Adaptation

• 一个理论上更一般化的MMD差异用于领域自适应

• A more general MMD for domain adaptation

• 20200813 A Boundary Based Out-of-Distribution Classifier for Generalized Zero-Shot Learning

• OOD classifier for generalized zero-shot learning

• 20200813 ICML-20 On Learning Language-Invariant Representations for Universal Machine Translation

• Theory for universal machine translation

• 对统一机器翻译模型进行了理论论证

• 20200702 ICML-20 Few-shot domain adaptation by causal mechanism transfer

• The first work on causal transfer learning

• 日本理论组大佬Sugiyama的工作，causal transfer learning

• 20191008 CVPR-19 Characterizing and Avoiding Negative Transfer

• Characterizing and avoid negative transfer

• 形式化并提出如何避免负迁移

• 20190301 ALT-19 A Generalized Neyman-Pearson Criterion for Optimal Domain Adaptation

• A new criterion for domain adaptation

• 提出一种新的可以强化domain adaptation表现的度量

• 20181219 arXiv PAC Learning Guarantees Under Covariate Shift

• PAC learning theory for covariate shift

• Covariate shift问题的PAC学习理论

• 20181206 arXiv Transferring Knowledge across Learning Processes

• Transfer learning across learning processes

• 学习过程中的知识迁移

• 20181128 arXiv Theoretical Guarantees of Transfer Learning

• Some theoretical analysis of transfer learning

• 一些关于迁移学习的理论分析

• 20181117 arXiv Theoretical Perspective of Deep Domain Adaptation

• Providing some theory analysis on deep domain adaptation

• 对deep domain adaptaiton做出了一些理论上的分析

• 20181106 workshop GENERALIZATION BOUNDS FOR DOMAIN ADAPTATION VIA DOMAIN TRANSFORMATIONS

• Analyze some generalization bound for domain adaptation

• 对domain adaptation进行了一些理论上的分析

1.1.2. Others (其他)

• 20210426 Distill on the Go: Online knowledge distillation in self-supervised learning

• Online knowledge distillation in self-supervised learning

• 自监督学习中的在线知识蒸馏

• 20210319 Cross-domain Activity Recognition via Substructural Optimal Transport | 知乎文章 | 微信公众号

• Using sub-structures for domain adaptation

• 采用子结构进行domain adaptation，比传统方法快5倍

• 20210202 ICLR-21 Rethinking Soft Labels for Knowledge Distillation: A Bias-Variance Tradeoff Perspective

• Rethink soft labels for KD in a bias-variance tradeoff perspective

• 从偏差-方差的角度重新思考蒸馏中的软标签

• 20210106 Style Normalization and Restitution for DomainGeneralization and Adaptation

• Style normalization and restitution for DA and DG

• 风格归一化用于DA和DG任务

• 20210104 A Closer Look at Few-Shot Crosslingual Transfer: Variance, Benchmarks and Baselines

• A closer look at few-shot crosslingual transfer

• 20201222 AAAI-21 DecAug: Out-of-Distribution Generalization via Decomposed Feature Representation and Semantic Augmentation

• OOD generalization

• 用特征分解和语义增强做OOD泛化

• 20201208 TIP Effective Label Propagation for Discriminative Semi-Supervised Domain Adaptation

• 用label propagation做半监督domain adaptation

• 20201203 Pre-Trained Image Processing Transformer

• 用transformer做low-level的图像任务

• 20201203 Unpaired Image-to-Image Translation via Latent Energy Transport

• 用能量模型做图像翻译

• 20201016 Deep Ensembles for Low-Data Transfer Learning

• Deep ensemble models for transfer learning

• 20201016 Energy-based Out-of-distribution Detection

• Energy-based OOD

• 20200927 Privacy-preserving Transfer Learning via Secure Maximum Mean Discrepancy

• 加密情况下的MMD用于迁移学习

• 20200914 A First Step Towards Distribution Invariant Regression Metrics

• 分布无关的回归评价

• 20200821 ECCV-20 Towards Recognizing Unseen Categories in Unseen Domains

• Recognizing unseen classes in unseen domains

• 对未知领域识别未知类

• 20200821 Self-Supervised Learning Across Domains

• 跨领域自监督学习

• 20200813 ECCV-20 Learning to Cluster under Domain Shift

• Learning to cluster under domain shift

• 在domain shift的情况下进行聚类

• 20200706 Learn Faster and Forget Slower via Fast and Stable Task Adaptation

• 20200706 In Search of Lost Domain Generalization

• 20200706 [ICML-20] Continuously Indexed Domain Adaptation

• 20200706 Interactive Knowledge Distillation

• 20200706 Domain Adaptation without Source Data

• 20200629 [ICML-20] Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation
  
• Source-free adaptation
  
• 在adaptation过程中不访问source data

• 20200629 Transfer learning via L1 regulaziation

• Using L1 regularizationg for transfer learning

• 20200629 [ICML-20] [Graph Optimal Transport for Cross-Domain Alignment])(https://arxiv.org/abs/2006.14744)

• Graph OT for cross-domain alignment

• 20200615 Rethinking Pre-training and Self-training

• 20200615 Double Double Descent: On Generalization Errors in Transfer Learning between Linear Regression Tasks

• 20200412 ICML-19 Towards understanding knowledge distillation

• Some theoretical and empirical understanding to knowledge distllation

• 对知识蒸馏的一些理论和实验的分析

• 20200210 WACVW-20 Impact of ImageNet Model Selection on Domain Adaptation

• A good experiment paper to indicate the power of representations

• 一篇很好的实验paper，揭示了深度特征+传统方法的有效性

• 20191202 AAAI-20 Towards Oracle Knowledge Distillation with Neural Architecture Search

• Using NAS for knowledge Distillation

• 用NAS帮助知识蒸馏

• 20191202 AAAI-20 Stable Learning via Sample Reweighting

• Theoretical sample reweigting

• 理论和方法，用于sample reweight

• 20191202 arXiv Domain-invariant Stereo Matching Networks

• Domain-invariant stereo matching networks

• 领域不变的匹配网络

• 20191202 arXiv Learning Generalizable Representations via Diverse Supervision

• Diverse supervision helps to learn generalizable representations

• 20191202 arXiv Domain-Aware Dynamic Networks

• Edge devices adaptative computing

• 边缘计算上的自适应计算

• 20191029 Adversarial Feature Alignment: Avoid Catastrophic Forgetting in Incremental Task Lifelong Learning

• Avoid catastrophic forgeeting in incremental task lifelong learning

• 在终身学习中避免灾难遗忘

• 20191029 Reducing Domain Gap via Style-Agnostic Networks

• Use style-agnostic networks to avoid domain gap

• 通过风格无关的网络来避免领域的gap

• 20191015 arXiv The Visual Task Adaptation Benchmark

• A new large benchmark for visual adaptation tasks by Google

• Google提出的一个巨大的视觉迁移任务数据集

• 20191011 arXiv Estimating Transfer Entropy via Copula Entropy

• Evaluate the transfer entopy via copula entropy

• 评估迁移熵

• 20191011 arXiv Learning to Remember from a Multi-Task Teacher

• Dealing with the catastrophic forgetting during sequential learning

• 在序列学习时处理灾难遗忘

• 20191008 arXiv DIVA: Domain Invariant Variational Autoencoders

• Domain invariant variational autoencoders

• 领域不变的变分自编码器

• 20190821 arXiv Transfer Learning-Based Label Proportions Method with Data of Uncertainty

• Transfer learning with source and target having uncertainty

• 当source和target都有不确定label时进行迁移

• 20190806 KDD-19 Relation Extraction via Domain-aware Transfer Learning

• Relation extraction using transfer learning for knowledge base construction

• 利用迁移学习进行关系抽取

• 20190703 arXiv Inferred successor maps for better transfer learning

• Inferred successor maps for better transfer learning

• 20190626 arXiv Transfer of Machine Learning Fairness across Domains

• Transfer of machine learning fairness across domains

• 机器学习的公平性的迁移

• 20190531 IJCAI-19 Adversarial Imitation Learning from Incomplete Demonstrations

• Adversarial imitation learning from imcomplete demonstrations

• 基于不完整实例的对抗模仿学习

• 20190517 arXiv Budget-Aware Adapters for Multi-Domain Learning

• Budget-Aware Adapters for Multi-Domain Learning

• 20190409 arXiv Improving Image Classification Robustness through Selective CNN-Filters Fine-Tuning

• Improving Image Classification Robustness through Selective CNN-Filters Fine-Tuning

• 通过可选择的CNN滤波器进行图像分类的fine-tuning

• 20190401 arXiv Distilling Task-Specific Knowledge from BERT into Simple Neural Networks

• Distill knowledge from BERT to simple neural networks

• 从BERT模型中迁移知识到简单网络中

• 20190305 arXiv Let’s Transfer Transformations of Shared Semantic Representations

• Transfer transformations from shared semantic representations

• 从共享的语义表示中进行特征迁移

• 20190301 SysML-19 FixyNN: Efficient Hardware for Mobile Computer Vision via Transfer Learning

• An efficient hardware for mobile computer vision applications using transfer learning

• 提出一个高效的用于移动计算机视觉应用的硬件

• 20190118 arXiv Domain Adaptation for Structured Output via Discriminative Patch Representations

• Domain adaptation for structured output

• Domain adaptation用于结构化输出

• 20190111 arXiv Low-Cost Transfer Learning of Face Tasks

• Infer what task transfers better and how to transfer

• 探索对于一个预训练好的网络来说哪个任务适合迁移、如何迁移

• 20190111 arXiv Transfer Representation Learning with TSK Fuzzy System

• Transfer learning with fuzzy system

• 基于模糊系统的迁移学习

• 20190102 arXiv An introduction to domain adaptation and transfer learning

• Another introduction to transfer learning

• 另一个迁移学习和domain adaptation综述

• 20181217 arXiv When Semi-Supervised Learning Meets Transfer Learning: Training Strategies, Models and Datasets

• Combining semi-supervised learning and transfer learning

• 将半监督方法应用于迁移学习

• 20181127 arXiv Privacy-preserving Transfer Learning for Knowledge Sharing

• First work on privacy preserving in transfer learning

• 第一篇探讨迁移学习中隐私保护的文章(第四范式、杨强)

• 20181121 arXiv An Efficient Transfer Learning Technique by Using Final Fully-Connected Layer Output Features of Deep Networks

• Using final fc layer to perform transfer learning

• 使用最后一层全连接层进行迁移学习

• 20181121 arXiv Not just a matter of semantics: the relationship between visual similarity and semantic similarity

• Interpreting relationships between visual similarity and semantic similarity

• 解释了视觉相似性和语义相似性的不同

• 20181008 arXiv Unsupervised Learning via Meta-Learning

• Meta-learning for unsupervised learning

• 用于无监督学习的元学习

• 20101008 arXiv Concept-drifting Data Streams are Time Series; The Case for Continuous Adaptation

• Continuous adaptation for time series data

• 对时间序列进行连续adaptation

• 20180925 arXiv DT-LET: Deep Transfer Learning by Exploring where to Transfer

• Explore the suitable layers to transfer

• 探索深度网络中效果表现好的对应的迁移层

• 20180919 JMLR Invariant Models for Causal Transfer Learning

• Invariant models for causal transfer learning

• 针对causal transfer learning提出不变模型

• 20180912 arXiv Transfer Learning with Neural AutoML

• Applying transfer learning into autoML search

• 将迁移学习思想应用于automl

• 20190904 arXiv On the Minimal Supervision for Training Any Binary Classifier from Only Unlabeled Data

• Train binary classifiers from only unlabeled data

• 仅从无标记数据训练二分类器

• 20180904 arXiv Learning Data-adaptive Nonparametric Kernels

• Learn a kernel that can do adaptation

• 学习一个可以自适应的kernel

• 20180901 arXiv Distance Based Source Domain Selection for Sentiment Classification

• Propose a new domain selection method by combining existing distance functions

• 提出一种混合已有多种距离公式的源领域选择方法

• 20180901 KBS Transfer subspace learning via low-rank and discriminative reconstruction matrix

• Transfer subspace learning via low-rank and discriminative reconstruction matrix

• 通过低秩和重构进行迁移学习

• 20180825 arXiv Transfer Learning for Estimating Causal Effects using Neural Networks

• Using transfer learning for casual effect learning

• 用迁移学习进行因果推理

• 20180724 ICPKR-18 Knowledge-based Transfer Learning Explanation

• Explain transfer learning things with some knowledge-based theory

• 用一些基于knowledge的方法解释迁移学习

• 20180628 arXiv 提出Office数据集的实验室又放出一个数据集用于close set、open set、以及object detection的迁移学习：Syn2Real: A New Benchmark forSynthetic-to-Real Visual Domain Adaptation

• 20180605 arXiv 解决federated learning中的数据不同分布的问题：Federated Learning with Non-IID Data

• 20180604 arXiv 在Open set domain adaptation中，用共享和私有部分重建进行问题的解决：Learning Factorized Representations for Open-set Domain Adaptation

• 20180403 arXiv 选择最优的子类生成方便迁移的特征：Class Subset Selection for Transfer Learning using Submodularity

• 20180326 ICMLA-17 在类别不平衡情况下比较了一些迁移学习和传统方法的性能，并做出一些结论：Comparing Transfer Learning and Traditional Learning Under Domain Class Imbalance

1.2. Domain Adaptation (领域自适应)

Including domain adaptation and partial domain adaptation.

1.2.1. Traditional Methods (传统迁移方法)

• 20200615 Deep Transfer Learning with Ridge Regression

• 20200324 Domain Adaptation by Class Centroid Matching and Local Manifold Self-Learning

• Domain adaptation by class centroid matching and local manifold self-learning

• 集合了聚类、中心匹配，及自学习的DA

• 20191204 arXiv Transferability versus Discriminability: Joint Probability Distribution Adaptation (JPDA)

• Joint adaptation with different weights

• 不同权重的联合概率适配

• 20191125 AAAI-20 Unsupervised Domain Adaptation via Structured Prediction Based Selective Pseudo-Labeling

• DA with selective pseudo label

• 结构化和选择性的伪标签用于DA

• 20190703 arXiv Domain Adaptation via Low-Rank Basis Approximation

• Domain adaptation with low-rank basis approximation

• 低秩分解进行domain adaptation

• 20190508 IJCNN-19 Unsupervised Domain Adaptation using Graph Transduction Games

• Domain adaptation using graph transduction games

• 用图转换博弈进行domain adaptation

• 20190403 ICME-19 Easy Transfer Learning By Exploiting Intra-domain Structures Code

• An easy transfer learning approach with good performance

• 一个非常简单但效果很好的迁移方法

• 20180724 ACMMM-18 Visual Domain Adaptation with Manifold Embedded Distribution Alignment

• The state-of-the-art results of domain adaptation, better than most traditional and deep methods

• 目前效果最好的非深度迁移学习方法，领先绝大多数最近的方法
• Code: MEDA

• 20180912 arXiv Unsupervised Domain Adaptation Based on Source-guided Discrepancy

• Using source domain information to help domain adaptation

• 使用源领域数据辅助目标领域进行domain adaptation

• 20181219 arXiv Domain Adaptation on Graphs by Learning Graph Topologies: Theoretical Analysis and an Algorithm

• Domain adaptation on graphs

• 在图上的领域自适应

• 20181121 arXiv Deep Discriminative Learning for Unsupervised Domain Adaptation

• Deep discriminative learning for domain adaptation

• 同时进行源域和目标域上的分类判别

• 20181114 arXiv Multiple Subspace Alignment Improves Domain Adaptation

• Project domains into multiple subsapce to do domain adaptation

• 将domain映射到多个subsapce上然后进行adaptation

• 20180912 ICIP-18 Structural Domain Adaptation with Latent Graph Alignment

• Using graph alignment for domain adaptation

• 使用图对齐方式进行domain adaptation

• 20180912 IEEE Access Unsupervised Domain Adaptation by Mapped Correlation Alignment

• Mapped correlation alignment for domain adaptation

• 用映射的关联对齐进行domain adaptation

• 20180912 ICALIP-18 Domain Adaptation for Gaussian Process Classification

• Domain Adaptation for Gaussian Process Classification

• 在高斯过程分类中进行domain adaptation

• 20180701 arXiv 对domain adaptation问题，基于optimal transport提出一种新的特征选择方法：Feature Selection for Unsupervised Domain Adaptation using Optimal Transport

• 20180510 IEEE Trans. Cybernetics 提出一个通用的迁移学习框架，对不同的domain进行不同的特征变换：Transfer Independently Together: A Generalized Framework for Domain Adaptation

• 20180403 TIP-18 一篇传统方法做domain adaptation的文章，比很多深度方法结果都好：An Embarrassingly Simple Approach to Visual Domain Adaptation

• 20180326 ICMLA-17 利用subsapce alignment进行迁移学习：Transfer Learning for Large Scale Data Using Subspace Alignment

• 20180228 arXiv 一篇通过标签一致性和MMD准则进行domain adaptation的文章: Discriminative Label Consistent Domain Adaptation

• 20180226 AAAI-18 清华龙明盛组最新工作：Unsupervised Domain Adaptation with Distribution Matching Machines

• 20180110 arXiv 一篇比较新的传统方法做domain adaptation的文章 Close Yet Discriminative Domain Adaptation

• 20180105 arXiv 最优的贝叶斯迁移学习 Optimal Bayesian Transfer Learning

• 20171201 ICCV-17 When Unsupervised Domain Adaptation Meets Tensor Representations

• 第一篇将Tensor与domain adaptation结合的文章。代码

• 我的解读

• 201711 ICCV-17 Open set domain adaptation。

• 当source和target只共享某一些类别时，怎么处理？这个文章获得了ICCV 2017的Marr Prize Honorable Mention，值得好好研究。

• 我的解读

• 201710 Domain Adaptation in Computer Vision Applications 里面收录了若干篇domain adaptation的文章，可以大概看看。

• 学习迁移(Learning to Transfer, L2T)

• 迁移学习领域的新方向：与在线、增量学习结合

• 我的解读

• 201707 Mutual Alignment Transfer Learning

• 201708 Learning Invariant Riemannian Geometric Representations Using Deep Nets

• 20170812 ICML-18 Learning To Transfer，将迁移学习和增量学习的思想结合起来，为迁移学习的发展开辟了一个崭新的研究方向。我的解读

• NIPS-17 JDOT: Joint distribution optimal transportation for domain adaptation

• AAAI-16 Return of Frustratingly Easy Domain Adaptation

• JMLR-16 Distribution-Matching Embedding for Visual Domain Adaptation

• CoRR abs/1610.04420 (2016) Theoretical Analysis of Domain Adaptation with Optimal Transport

• CVPR-14 Transfer Joint Matching for Unsupervised Domain Adaptation

• ICCV-13 Transfer Feature Learning with Joint Distribution Adaptation

• 迁移成分分析方法(Transfer component analysis, TCA)

• Domain adaptation via tranfer component analysis

• 发表在IEEE Trans. Neural Network期刊上（现改名为IEEE trans. Neural Network and Learning System），前作会议文章发在AAAI-09上
• 我的解读

• 联合分布适配方法（joint distribution adaptation，JDA）

• Transfer Feature Learning with Joint Distribution Adaptation

• 发表在2013年的ICCV上
• 我的解读

• 测地线流式核方法(Geodesic flow kernel, GFK)

• Geodesic flow kernel for unsupervised domain adaptation

• 发表在CVPR-12上
• 我的解读

• 领域不变性迁移核学习(Transfer Kernel Learning, TKL)

• Domain invariant transfer kernel learning

• 发表在IEEE Trans. Knowledge and Data Engineering期刊上

1.2.2. Deep / Adversarial Methods (深度/对抗迁移方法)

• 20210329 ICLR-21 Tent: Fully Test-Time Adaptation by Entropy Minimization

• Test time adaptation by entropy minimization

• 测试时通过熵最小化进行adaptation

• 20210329 Adversarial Branch Architecture Search for Unsupervised Domain Adaptation

• NAS for domain adaptation

• 用神经网络结构搜索做领域自适应

• 20210312 Discrepancy-Based Active Learning for Domain Adaptation

• Discrepancy and active learning for DA

• 基于主动学习的DA

• 20210312 Unbalanced minibatch Optimal Transport; applications to Domain Adaptation

• Unbalanced minibatch OT for DA

• 非均衡的OT用于DA问题

• 20210127 Hierarchical Domain Invariant Variational Auto-Encoding with weak domain supervision

• 利用VAE和解耦去做domain generalization

• Using VAE and disentanglement for domain generalization

• 20201214 WWW-20 Domain Adaptation with Category Attention Network for Deep Sentiment Analysis

• Unify pivots and non-pivots, and provide interpretability for domain adaptation in sentiment analysis

• 统一pivots和non-pivots，并提供可解释性进行DA情感分析

• 20201208 NIPS-20 Heuristic Domain Adaptation

• 启发式domain adaptation

• 20200804 ECCV-20 spotlight Side-Tuning: A Baseline for Network Adaptation via Additive Side Networks

• 将现有的finetune机制进行扩展

• Extending finetune mechanism
• 20200804 ACMMM-20 Adversarial Bipartite Graph Learning for Video Domain Adaptation
• Video domain adaptation
• 视频的领域自适应

• 20200804 MICCAI-20 Whole MILC: generalizing learned dynamics across tasks, datasets, and populations
• Generalizing across tasks, datasets, populations
• 在任务、数据集、人群之间做泛化

• 20200724 Learning to Match Distributions for Domain Adaptation

• 自动深度迁移学习

• Automatic domain adaptation

• 20200529 TNNLS Deep Subdomain Adaptation Network for Image Classification

• A fine-grained adaptation method with LMMD, which is very simple and effective

• 一种细粒度自适应的方法，使用LMMD进行对齐，该方法非常简单有效

• 20200420 arXiv One-vs-Rest Network-based Deep Probability Model for Open Set Recognition

• One-vs-rest deep model for open set recognition

• 用于开放集的识别的深度网络

• 20200414 ICLR-20 Gradient as features for deep representation learning

• Gradients as features for deep representation learning on pretrained models

• 在预训练模型基础上，将梯度作为额外的feature，提高学习表现

• 20200414 ICLR-20 Domain adaptive multi-branch networks

• A domain adaptation framework using a multi-branch cascade structure

• 一个用了多层级联、多分支结构的DA框架

• 20200405 CVPR-20 Towards Discriminability and Diversity: Batch Nuclear-norm Maximization under Label Insufficient Situations

• A simple regularization-based adaptation method

• 一个非常简单的基于能量最小化的adaptation方法

• 20200210 AAAI-20 Bi-Directional Generation for Unsupervised Domain Adaptation

• Bidirectional GANs for domain adaptation

• 双向的GAN用来做DA

• 20191202 PR-19 Correlation-aware Adversarial Domain Adaptation and Generalization

• CORAL and adversarial for adaptation and generalization

• 基于CORAL和对抗网络的DA和DG

• 20191201 BMVC-19 Domain Adaptation for Object Detection via Style Consistency

• Use style consistency for domain adaptation

• 通过结构一致性来进行domain adaptation

• 20191124 AAAI-20 Knowledge Graph Transfer Network for Few-Shot Recognition

• GNN for semantic transfer for few-shot learning

• 用GNN进行类别的语义迁移用于few-shot learning

• 20191124 arXiv Improving Unsupervised Domain Adaptation with Variational Information Bottleneck

• Information bottleneck for unsupervised da

• 用了信息瓶颈来进行DA

• 20191124 AAAI-20 (AdaFilter: Adaptive Filter Fine-tuning for Deep Transfer Learning)(https://arxiv.org/abs/1911.09659)

• Adaptively determine which layer to transfer or finetune

• 自适应地决定迁移哪个层或微调哪个层

• 20191113 arXiv Knowledge Distillation for Incremental Learning in Semantic Segmentation

• Knowledge distillation for incremental learning in semantic segmentation

• 在语义分割问题中针对增量学习进行知识蒸馏

• 20191111 NIPS-19 PointDAN: A Multi-Scale 3D Domain Adaption Network for Point Cloud Representation

• Multi-scale 3D DA network for point cloud representation

• 20191111 CCIA-19 Feature discriminativity estimation in CNNs for transfer learning

• Feature discriminativity estimation in CNN for TL

• 20191012 ICCV-19 Drop to Adapt: Learning Discriminative Features for Unsupervised Domain Adaptation

• Drop to Adapt: Learning Discriminative Features for Unsupervised Domain Adaptation

• 直接適應：學習非監督域自適應的判別功能

• 20191015 arXiv Deep Kernel Transfer in Gaussian Processes for Few-shot Learning

• Deep kernel transfer learing in Gaussian process

• 高斯过程中的深度迁移学习

• 20191008 EMNLP-19 workshop Domain Differential Adaptation for Neural Machine Translation

• Embrace the difference between domains for adaptation

• 拥抱domain的不同，并利用这些不同帮助adaptation

• 20191008 BMVC-19 Multi-Weight Partial Domain Adaptation

• Class and sample weight contribution for partial domain adaptation

• 同时考虑类别和样本比重用于部分迁移学习

• 20190813 ICCV-19 oral UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation

• A unified framework for domain adaptation

• 一个统一的用于domain adaptation的框架

• 20190809 arXiv Multi-Purposing Domain Adaptation Discriminators for Pseudo Labeling Confidence

• Improve pseudo label confidence using multi-purposing DA

• 用多目标DA提高伪标签准确率

• 20190809 arXiv Semi-supervised representation learning via dual autoencoders for domain adaptation

• Semi-supervised learning via autoencoders

• 半监督autoencoder用于DA

• 20190809 arXiv Mind2Mind : transfer learning for GANs

• Transfer learning using GANs

• 用GAN进行迁移学习

• 20190809 arXiv Self-supervised Domain Adaptation for Computer Vision Tasks

• Self-supervised DA

• 自监督DA

• 20190809 arXiv Hidden Covariate Shift: A Minimal Assumption For Domain Adaptation

• Hidden covariate shift

• 一种新的DA假设

• 20190809 PR-19 Cross-domain Network Representations

• Cross-domain network representation learning

• 跨领域网络表达学习

• 20190809 ICCV-19 Larger Norm More Transferable: An Adaptive Feature Norm Approach for Unsupervised Domain Adaptation

• Adaptive Feature Norm Approach for Unsupervised Domain Adaptation

• 自适应的特征归一化用于DA

• 20190731 MICCAI-19 Unsupervised Domain Adaptation via Disentangled Representations: Application to Cross-Modality Liver Segmentation

• Disentangled representations for unsupervised domain adaptation

• 基于解耦表征的无监督领域自适应

• 20190719 arXiv Agile Domain Adaptation

• Domain adaptation by considering the difficulty in classification

• 通过考虑不同样本分离的难度进行domain adaptation

• 20190718 arXiv Measuring the Transferability of Adversarial Examples

• Measure the transferability of adversarial examples

• 度量对抗样本的可迁移性

• 20190604 IJCAI-19 DANE: Domain Adaptive Network Embedding

• Transfered network embeddings for different networks

• 不同网络表达的迁移

• 20190604 arXiv Learning to Transfer: Unsupervised Meta Domain Translation

• Unsupervised meta domain translation

• 无监督领域翻译

• 20190530 arXiv Learning Bregman Divergences

• Learning Bregman divergence

• 学习Bregman差异

• 20190530 arXiv Adversarial Domain Adaptation Being Aware of Class Relationships

• Using class relationship for adversarial domain adaptation

• 利用类别关系进行对抗的domain adaptaition

• 20190530 arXiv Cross-Domain Transferability of Adversarial Perturbations

• Cross-Domain Transferability of Adversarial Perturbations

• 20190525 PAMI-19 Learning More Universal Representations for Transfer-Learning

• Learning more universal representations for transfer learning

• 对迁移学习设计2种方式学到更通用的表达

• 20190517 ICML-19 Learning What and Where to Transfer

• Learning what and where to transfer in deep networks

• 学习深度网络从何处迁移

• 20190517 ICML-19 Zero-Shot Voice Style Transfer with Only Autoencoder Loss

• Zero-shot voice style transfer with only autoencoder loss

• 零次声音迁移

• 20190515 CVPR-19 Diversify and Match: A Domain Adaptive Representation Learning Paradigm for Object Detection

• Domain adaptation for object detection

• 领域自适应用于物体检测

• 20190507 NAACL-HLT 19 Transfer of Adversarial Robustness Between Perturbation Types

• Transfer of Adversarial Robustness Between Perturbation Types

• 20190416 arXiv ACE: Adapting to Changing Environments for Semantic Segmentation

• Propose a new method that can adapt to new environments

• 提出一种可以适配不同环境的方法

• 20190416 arXiv Incremental multi-domain learning with network latent tensor factorization

• Incremental multi-domain learning with network latent tensor factorization

• 网络隐性tensor分解应用于多领域增量学习

• 20190415 PAKDD-19 Parameter Transfer Unit for Deep Neural Networks

• Propose a parameter transfer unit for DNN

• 对深度网络提出参数迁移单元

• 20190412 PAMI-19 Beyond Sharing Weights for Deep Domain Adaptation

• Domain adaptation by not sharing weights

• 通过不共享权重来进行domain adaptation

• 20190405 IJCNN-19 Accelerating Deep Unsupervised Domain Adaptation with Transfer Channel Pruning

• The first work to accelerate transfer learning

• 第一个加速迁移学习的工作

• 20190102 WSDM-19 Learning to Selectively Transfer: Reinforced Transfer Learning for Deep Text Matching

• Reinforced transfer learning for deep text matching

• 迁移学习进行深度文本匹配

• 20190102 arXiv DART: Domain-Adversarial Residual-Transfer Networks for Unsupervised Cross-Domain Image Classification

• Adversarial + residual for domain adaptation

• 对抗+残差进行domain adaptation

• 20181220 arXiv TWINs: Two Weighted Inconsistency-reduced Networks for Partial Domain Adaptation

• Two weighted inconsistency-reduced networks for partial domain adaptation

• 两个权重网络用于部分domain adaptation

• 20181127 arXiv Learning Grouped Convolution for Efficient Domain Adaptation

• Group convolution for domain adaptation

• 群体卷积进行domain adaptation

• 20181121 arXiv Unsupervised Domain Adaptation: An Adaptive Feature Norm Approach

• A nonparametric method for domain adaptation

• 一种无参数的domain adaptation方法

• 20181121 arXiv Domain Adaptive Transfer Learning with Specialist Models

• Sample reweighting methods for domain adaptative

• 样本权重更新法进行domain adaptation

• 20180926 ICLR-18 Self-ensembling for visual domain adaptation

• Self-ensembling for domain adaptation

• 将self-ensembling应用于da

• 20180620 CVPR-18 用迁移学习进行fine tune：Large Scale Fine-Grained Categorization and Domain-Specific Transfer Learning

• 20180321 CVPR-18 构建了一个迁移学习算法，用于解决跨数据集之间的person-reidenfication: Unsupervised Cross-dataset Person Re-identification by Transfer Learning of Spatial-Temporal Patterns

• 20180315 ICLR-17 一篇综合进行two-sample stest的文章：Revisiting Classifier Two-Sample Tests

• 20171214 arXiv Investigating the Impact of Data Volume and Domain Similarity on Transfer Learning Applications

• 在实验中探索了数据量多少，和相似度这两个因素对迁移学习效果的影响

• NIPS-17 Learning Multiple Tasks with Multilinear Relationship Networks

• 深度适配网络（Deep Adaptation Network, DAN）

• 发表在ICML-15上：learning transferable features with deep adaptation networks

• 我的解读

1.3. Domain Generalization

• 2021 Domain generalization: IJCAI-21 Generalizing to Unseen Domains: A Survey on Domain Generalization | 知乎文章 | 微信公众号

• First survey on domain generalization

• 第一篇对Domain generalization (领域泛化)的综述

• 20180701 arXiv 做迁移时，只用source数据，不用target数据训练：Generalizing to Unseen Domains via Adversarial Data Augmentation

• 201711 ICLR-18 GENERALIZING ACROSS DOMAINS VIA CROSS-GRADIENT TRAINING

• 不同于以往的工作，本文运用贝叶斯网络建模label和domain的依赖关系，抓住training、inference 两个过程，有效引入domain perturbation来实现domain adaptation。

• ICLR-18 generalizing across domains via cross-gradient training

• 20181106 PRCV-18 Domain Attention Model for Domain Generalization in Object Detection

• Adding attention for domain generalization

• 在domain generalization中加入了attention机制

• 20181225 WACV-19 Multi-component Image Translation for Deep Domain Generalization

• Using GAN generated images for domain generalization

• 用GAN生成的图像进行domain generalization

• 20180724 arXiv Domain Generalization via Conditional Invariant Representation

• Using Conditional Invariant Representation for domain generalization

• 生成条件不变的特征表达，用于domain generalization问题

• 20181212 arXiv Beyond Domain Adaptation: Unseen Domain Encapsulation via Universal Non-volume Preserving Models

• Domain generalization method

• 一种针对于unseen domain的学习方法

• 20171210 AAAI-18 Learning to Generalize: Meta-Learning for Domain Generalization

• 将Meta-Learning与domain generalization结合的文章，可以联系到近期较为流行的few-shot learning进行下一步思考。

1.4. Multi-source Transfer Learning (多源迁移学习)

• 20200427 TriGAN: Image-to-Image Translation for Multi-Source Domain Adaptation

• A cycle-gan style multi-source DA

• 类似于cyclegan的多源领域适应

• 20190902 AAAI-19 Aligning Domain-Specific Distribution and Classifier for Cross-Domain Classification from Multiple Sources

• Multi-source domain adaptation using both features and classifier adaptation

• 利用特征和分类器同时适配进行多源迁移，效果很好

• 20181212 AIKP Multi-source Transfer Learning

• Multi-source transfer

• 20181207 arXiv Moment Matching for Multi-Source Domain Adaptation

• Moment matching and propose a new large dataset for domain adaptation

• 提出一种moment matching的网络，并且提出一种新的domain adaptation数据集，很大

• CoRR abs/1711.09020 (2017) StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation

• 20180524 arXiv 探索了Multi-source迁移学习的一些理论：Algorithms and Theory for Multiple-Source Adaptation

• 20181117 AAAI-19 Robust Optimization over Multiple Domains

• Optimization on multi domains

• 针对多个domain建模并优化

• 20180912 arXiv Multi-target Unsupervised Domain Adaptation without Exactly Shared Categories

• Multi-target domain adaptation

• 多目标的domain adaptation

• 20180316 arXiv 用optimal transport解决domain adaptation中类别不平衡的问题：Optimal Transport for Multi-source Domain Adaptation under Target Shift

1.5. Heterogeneous Transfer Learning (异构迁移学习)

• 20190717 AAAI Heterogeneous Transfer Learning via Deep Matrix Completion with Adversarial Kernel Embedding

• Transfer Learning via Deep Matrix Completion with Adversarial Kernel Embedding

• 异构迁移学习中用对抗核嵌入的深度矩阵

• 20190829 ACMMM-19 Heterogeneous Domain Adaptation via Soft Transfer Network

• Soft-mmd loss in heterogeneous domain adaptation

• 异构迁移学习中用soft-mmd loss

• 20181113 ACML-18 Unsupervised Heterogeneous Domain Adaptation with Sparse Feature Transformation

• Heterogeneous domain adaptation

• 异构domain adaptation

• 20180901 TKDE A General Domain Specific Feature Transfer Framework for Hybrid Domain Adaptation

• Hybrid DA: special case in Heterogeneous DA

• 提出一种新的混合DA问题和方法

• 20180606 arXiv 一篇最近的对非对称情况下的异构迁移学习综述：Asymmetric Heterogeneous Transfer Learning: A Survey

• 20180403 Neural Processing Letters-18 异构迁移学习：Label Space Embedding of Manifold Alignment for Domain Adaption

• 20180105 arXiv 异构迁移学习 Heterogeneous transfer learning

1.6. Online Transfer Learning (在线迁移学习)

• 20180326 考虑主动获取label的budget情况下的在线迁移学习：Online domain adaptation by exploiting labeled features and pro-active learning

• 20180128 第一篇在线迁移学习的文章，发表在ICML-10上，系统性地定义了在线迁移学习的任务，给出了进行在线同构和异构迁移学习的两种学习模式。Online Transfer Learning

• 扩充的期刊文章发在2014年的AIJ上：Online Transfer Learning

• 我的解读
• 文章代码：OTL

• 20180126 两篇在线迁移学习：

• Online transfer learning by leveraging multiple source domains

• Online Heterogeneous Transfer by Hedge Ensemble of Offline and Online Decisions

• 20180126 TKDE-17 同时有多个同构和异构源域时的在线迁移学习：Online Transfer Learning with Multiple Homogeneous or Heterogeneous Sources

• KIS-17 Online transfer learning by leveraging multiple source domains 提出一种综合衡量多个源域进行在线迁移学习的方法。文章的related work是很不错的survey。

• CIKM-13 OMS-TL: A Framework of Online Multiple Source Transfer Learning 第一次在mulitple source上做online transfer，也是用的分类器集成。

• ICLR-17 ONLINE BAYESIAN TRANSFER LEARNING FOR SEQUENTIAL DATA MODELING 用贝叶斯的方法学习在线的HMM迁移学习模型，并应用于行为识别、睡眠监测，以及未来流量分析。

• KDD-14 Scalable Hands-Free Transfer Learning for Online Advertising 提出一种无参数的SGD方法，预测广告量

• TNNLS-17 Online Feature Transformation Learning for Cross-Domain Object Category Recognition 在线feature transformation方法

• ICPR-12 Online Transfer Boosting for Object Tracking 在线transfer 样本

• TKDE-14 Online Feature Selection and Its Applications 在线特征选择

• AAAI-15 Online Transfer Learning in Reinforcement Learning Domains 应用于强化学习的在线迁移学习

• AAAI-15 Online Boosting Algorithms for Anytime Transfer and Multitask Learning 一种通用的在线迁移学习方法，可以适配在现有方法的后面

• IJSR-13 Knowledge Transfer Using Cost Sensitive Online Learning Classification 探索在线迁移方法，用样本cost

1.7. Zero-shot / Few-shot Learning

• 20210426 Few-shot Continual Learning: a Brain-inspired Approach

• Few-shot continual learning

• 小样本持续学习

• 20201203 How to fine-tune deep neural networks in few-shot learning?

• 对few-shot任务如何fine-tune深度网络？

• 20201116 Filter Pre-Pruning for Improved Fine-tuning of Quantized Deep Neural Networks

• 量子神经网络中的finetune

• 20200608 ICML-20 Few-Shot Learning as Domain Adaptation: Algorithm and Analysis

• Using domain adaptation to solve the few-shot learning

• 20200408 ICLR-20 A Baseline for Few-Shot Image Classification
  
• A simple finetune+entropy minimization approach with strong baseline
  
• 一个微调+最小化熵的小样本学习方法，结果很强

• 20200405 ICCV-19 Variational few-shot learning

• Variational few-shot learning

• 变分小样本学习

• 20200405 ICLR-20 A baseline for few-shot image classification

• A simple but powerful baseline for few-shot image classification

• 一个简单但是很有效的few-shot baseline

• 20200324 IEEE TNNLS Few-Shot Learning with Geometric Constraints

• Few-shot learning with geometric constraints

• 用了一些几何约束进行小样本学习

• 20190813 arXiv Domain-Specific Embedding Network for Zero-Shot Recognition

• Domain-specific embedding network for zero-shot learning

• 领域自适应的zero-shot learning

• 20190401 TIp-19 Few-Shot Deep Adversarial Learning for Video-based Person Re-identification

• Few-shot deep adversarial learning

• Few-shot对抗学习

• 20190305 arXiv Zero-Shot Task Transfer

• Zero-shot task transfer

• Zero-shot任务迁移学习

• 20190221 arXiv Adaptive Cross-Modal Few-Shot Learning

• Adaptive cross-modal few-shot learning

• 跨模态的few-shot

• 20180612 CVPR-18 泛化的Zero-shot learning：Generalized Zero-Shot Learning via Synthesized Examples

• 20181106 arXiv Zero-Shot Transfer VQA Dataset

• English: A dataset for zero-shot VQA transfer

• 中文：一个针对zero-shot VQA的迁移学习数据集

• 20171222 NIPS 2017 用adversarial网络，当target中有很少量的label时如何进行domain adaptation：Few-Shot Adversarial Domain Adaptation

• 20181225 arXiv Learning Compositional Representations for Few-Shot Recognition

• Few-shot recognition

• 20181127 WACV-19 Self Paced Adversarial Training for Multimodal Few-shot Learning

• Multimodal training for single modal testing

• 用多模态数据针对单一模态进行迁移

• 20180728 arXiv Meta-learning autoencoders for few-shot prediction

• Using meta-learning for few-shot transfer learning

• 用元学习进行迁移学习

• 20171216 arXiv Zero-Shot Deep Domain Adaptation

• 当target domain的数据不可用时，如何用相关domain的数据进行辅助学习？

1.7.1. Zero-shot Learning based on Data Synthesis (基于样本生成的零样本学习)

详细介绍

• 20191204 arXiv MetAdapt: Meta-Learned Task-Adaptive Architecture for Few-Shot Classification

• Task adaptive structure for few-shot learning

• 目标自适应的结构用于小样本学习

• 20190409 ICLR-19 A Closer Look at Few-shot Classification

• Give some important conclusions on few-shot classification

• 在few-shot上给了一些有用的结论

• 20190401 IJCNN-19 Zero-shot Image Recognition Using Relational Matching, Adaptation and Calibration

• Zero-shot image recognition

• 零次学习的图像识别

• 20190301 NeurIPS-18 workshp One-Shot Federated Learning

• One-shot federated learning

• 20171022 ICCVW-17 Zero-shot learning posed as a missing data problem

• 算法首先学习 semantic embeddings 的结构性知识，利用学习到的知识和已知类的 image features 合成未知类的 image features。再利用无标记的未知类数据对合成数据进行修正。 算法假设未知类数据呈混合高斯分布，用 GMM-EM 算法进行无监督修正。

• 20180516 arXiv-18 A Large-scale Attribute Dataset for Zero-shot Learning

• 传统 ZSL 数据集（如 AwA, CUB）存在规模小，属性标注不丰富等问题。本文提出一个新的属性数据集 LAD 用于测试零样本学习算法。新数据集包含 230 类， 78,017 张图片，标注了 359 种属性。基于此数据集举办了 AI Challenger 零样本学习竞赛。 110+ 支来自海内外的参赛队伍提交了成绩。

• 20180710 ICML-18 MSplit LBI: Realizing Feature Selection and Dense Estimation Simultaneously in Few-shot and Zero-shot Learning

• 针对 L1 （欠拟合） 和 L2 （无特征选择、有偏） 正则项存在的问题，提出 MSplit LBI 用于同时实现特征选择和密集估计。在 Few-shot Learning 和 Zero-shot Learning 两个问题上进行了实验。实验表明 MSplit LBI 由优于 L1 和 L2。针对 ZSL 进行了特征可视化实验。

• 20190108 WACV-19 Zero-shot Learning via Recurrent Knowledge Transfer

• 基于样本合成的零样本学习算法通常将 semantic embeddings 的知识迁移到 image features 以实现 ZSL。然而，这种 training 和 testing space 的不一致，会导致这种迁移失效。因此，本文提出 Space Shift Problem，并针对此问题，提出一种（在 image feature space 和 semantic embedding space 之间）递归传递知识的解决方案。

1.8. Deep Transfer Learning (深度迁移学习)

1.8.1. Non-Adversarial Transfer Learning (非对抗深度迁移)

• 20210420 arXiv On Universal Black-Box Domain Adaptation

• Universal black-box domain adaptation

• 黑盒情况下的universal domain adaptation

• 20210319 Learning Invariant Representations across Domains and Tasks

• Automatically learn to match distributions

• 自动适配分布的任务适配网络

• 20191222 arXiv Dreaming to Distill: Data-free Knowledge Transfer via DeepInversion

• Generate data without priors for transfer learning based on deep dream

• 只用网络架构不用原来数据，生成新数据用于迁移

• 20191222 AAAI-20 Improved Knowledge Distillation via Teacher Assistant

• Teacher assistant helps knowledge distillation

• 20191204 AAAI-20 Online Knowledge Distillation with Diverse Peers

• Online Knowledge Distillation with Diverse Peers

• 20191201 arXiv A Unified Framework for Lifelong Learning in Deep Neural Networks

• A unified framework for life-long learing in DNN

• 20191201 arXiv ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring

• Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoring

• 20191119 ICDM-19 Towards Making Deep Transfer Learning Never Hurt

• Towards making deep transfer learning never hurt

• 通过正则避免负迁移

• 20191119 NIPS-19 workshop Collaborative Unsupervised Domain Adaptation for Medical Image Diagnosis

• Ensemble DA using noise labels

• 在ensemble中出现noise label时如何处理

• 20191119 NIPS-19 Transferable Normalization: Towards Improving Transferability of Deep Neural Networks

• Transfer normalization

• 20191029 KBS Semi-supervised representation learning via dual autoencoders for domain adaptation

• Semi-supervised domain adaptation with autoencoders

• 用自动编码器进行半监督的DA

• 20190929 NeurIPS-19 Deep Model Transferability from Attribution Maps

• Using attribution map for network similarity

• 与cvpr18的taskmony类似，这次用了属性图的方式探索网络的相似性

• 20190926 arXiv Learning a Domain-Invariant Embedding for Unsupervised Domain Adaptation Using Class-Conditioned Distribution Alignment

• Use class-conditional DA for domain adaptation

• 使用类条件对齐进行domain adaptation

• 20190926 arXiv A Deep Learning-Based Approach for Measuring the Domain Similarity of Persian Texts

• Deep learning based domain similarity learning

• 利用深度学习进行领域相似度的学习

• 20190926 arXiv Transfer Learning across Languages from Someone Else’s NMT Model

• Transfer learning across languages from NMT pretrained model

• 利用预训练好的NMT模型进行迁移学习

• 20190926 arXiv FEED: Feature-level Ensemble for Knowledge Distillation

• Feature-level knowledge distillation

• 特征层面的知识蒸馏

• 20190926 ICCV-19 Self-similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-identification

• A simple approach for domain adaptation

• 一个很简单的DA方法

• 20190910 BMVC-19 Curriculum based Dropout Discriminator for Domain Adaptation

• Curriculum dropout for domain adaptation

• 基于课程学习的dropout用于DA

• 20190909 IJCAI-FML-19 FedHealth: A Federated Transfer Learning Framework for Wearable Healthcare

• The first work on federated transfer learning for wearable healthcare

• 第一个将联邦迁移学习用于可穿戴健康监护的工作

• 20190909 PAMI Inferring Latent Domains for Unsupervised Deep Domain Adaptation

• Inferring latent domains for unsupervised deep domain

• 在深度迁移学习中推断隐含领域

• 20190729 ICCV workshop Multi-level Domain Adaptive learning for Cross-Domain Detection

• Multi-level domain adaptation for cross-domain Detection

• 多层次的domain adaptation

• 20190626 IJCAI-19 Bayesian Uncertainty Matching for Unsupervised Domain Adaptation

• Bayesian uncertainty matching for da

• 贝叶斯网络用于da

• 20190419 CVPR-19 DDLSTM: Dual-Domain LSTM for Cross-Dataset Action Recognition

• Dual-Domain LSTM for Cross-Dataset Action Recognition

• 跨数据集的动作识别

• 20190109 InfSc Robust Unsupervised Domain Adaptation for Neural Networks via Moment Alignment

• Extension of Central Moment Discrepancy (ICLR-17) approach

• 20181212 ICONIP-18 Domain Adaptation via Identical Distribution Across Models and Tasks

• Transfer from large net to small net

• 从大网络迁移到小网络

• 20181212 AIKP Deep Domain Adaptation

• Low-rank + deep nn for domain adaptation

• Low-rank用于深度迁移

• 20181211 arXiv Deep Variational Transfer: Transfer Learning through Semi-supervised Deep Generative Models

• Transfer learning with deep generative model

• 通过深度生成模型进行迁移学习

• 20181207 arXiv Feature Matters: A Stage-by-Stage Approach for Knowledge Transfer

• Feature transfer in student-teacher networks

• 在学生-教师网络中进行特征迁移

• 20181128 arXiv Low-resolution Face Recognition in the Wild via Selective Knowledge Distillation

• Knowledge distilation for low-resolution face recognition

• 将知识蒸馏应用于低分辨率的人脸识别

• 20181128 arXiv One Shot Domain Adaptation for Person Re-Identification

• One shot learning for REID

• One shot for再识别

• 20181123 arXiv SpotTune: Transfer Learning through Adaptive Fine-tuning

• Very interesting work: how exactly determine the finetune process?

• 很有意思的工作：如何决定finetune的策略？

• 20181121 arXiv Integrating domain knowledge: using hierarchies to improve deep classifiers

• Using hierarchies to help deep learning

• 借助于层次关系来帮助深度网络进行学习

• 20181117 arXiv AdapterNet - learning input transformation for domain adaptation

• Learning input transformation for domain adaptation

• 对domain adaptation任务学习输入的自适应

• 20181115 AAAI-19 Exploiting Local Feature Patterns for Unsupervised Domain Adaptation

• Local domain alignment for domain adaptation

• 局部领域自适应

• 20181115 NIPS-18 Co-regularized Alignment for Unsupervised Domain Adaptation

• The idea is the same with the above one…

• 仍然是局部对齐。。。

• 20181113 NIPS-18 Generalized Zero-Shot Learning with Deep Calibration Network

• Deep calibration network for zero-shot learning

• 提出deep calibration network进行zero-shot learning

• 20181110 AAAI-19 Knowledge Transfer via Distillation of Activation Boundaries Formed by Hidden Neurons

• Transfer learning for bounding neuron activation boundaries

• 使用迁移学习进行神经元激活边界判定

• 20181109 PAMI-18 Transferable Representation Learning with Deep Adaptation Networks

• Journal version of DAN paper

• DAN的Journal版本

• 20181108 arXiv Deep feature transfer between localization and segmentation tasks

• Feature transfer between localization and segmentation

• 在定位与分割任务间进行迁移

• 20181107 BigData-18 Transfer learning for time series classification

• First work on deep transfer learning for time series classification

• 第一个将深度迁移学习用于时间序列分类

• 20181106 PRCV-18 Deep Local Descriptors with Domain Adaptation

• Adding MMD layers to conv and fc layers

• 在卷积和全连接层都加入MMD

• 20181106 LNCS-18 LSTN: Latent Subspace Transfer Network for Unsupervised Domain Adaptation

• Combine subspace learning and neural network for DA

• 将子空间表示和深度网络结合起来用于DA

• 20181105 SIGGRAPI-18 Unsupervised representation learning using convolutional and stacked auto-encoders: a domain and cross-domain feature space analysis

• Representation learning for cross-domains

• 跨领域的特征学习

• 20181105 arXiv Progressive Memory Banks for Incremental Domain Adaptation

• Progressive memory bank in RNN for incremental DA

• 针对增量的domain adaptation，进行记忆单元的RNN学习

• 20180909 arXiv A Survey on Deep Transfer Learning

• A survey on deep transfer learning

• 深度迁移学习的survey

• 20180901 arXiv Joint Domain Alignment and Discriminative Feature Learning for Unsupervised Deep Domain Adaptation

• deep domain adaptation + intra-class / inter-class distance

• 深度domain adaptation再加上类内类间距离学习

• 20180819 arXiv Conceptual Domain Adaptation Using Deep Learning

• A search framework for deep transfer learning

• 提出一个可以搜索的framework进行迁移学习

• 20180731 ECCV-18 DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation

• Deep + Joint distribution adaptation + optimal transport

• 深度 + 联合分布适配 + optimal transport

• 20180731 ICLR-18 Few Shot Learning with Simplex

• Represent deep learning using the simplex

• 用单纯性表征深度学习

• 20180724 AIAI-18 Improving Deep Models of Person Re-identification for Cross-Dataset Usage

• apply deep models to cross-dataset RE-ID

• 将深度迁移学习应用于跨数据集的Re-ID

• 20180724 ECCV-18 Zero-Shot Deep Domain Adaptation

• Perform zero-shot domain adaptation when there is no target domain data available

• 当目标领域的数据不可用时如何进行domain adaptation :

• 20180724 ICCSE-18 Deep Transfer Learning for Cross-domain Activity Recognition

• Provide source domain selection and activity recognition for cross-domain activity recognition

• 提出了跨领域行为识别中的深度方法模型，以及相关的领域选择方法

• 20180530 arXiv 用于深度网络的鲁棒性domain adaptation方法：Robust Unsupervised Domain Adaptation for Neural Networks via Moment Alignment

• 20180522 arXiv 用CNN进行跨领域的属性学习：Cross-domain attribute representation based on convolutional neural network

• 20180428 CVPR-18 相互协同学习：Deep Mutual Learning

• 20180428 ICLR-18 自集成学习用于domain adaptation：Self-ensembling for visual domain adaptation

• 20180428 IJCAI-18 将knowledge distilation用于transfer learning，然后进行视频分类：Better and Faster: Knowledge Transfer from Multiple Self-supervised Learning Tasks via Graph Distillation for Video Classification

• 20180426 arXiv 深度学习中的参数如何进行迁移？（杨强团队）：Parameter Transfer Unit for Deep Neural Networks

• 20180425 CVPR-18(oral) 对不同的视觉任务进行建模，从而可以进行深层次的transfer：Taskonomy: Disentangling Task Transfer Learning

• 20180410 ICLR-17 第一篇用可变RNN进行多维时间序列迁移的文章：Variational Recurrent Adversarial Deep Domain Adaptation

• 20180403 arXiv 本地和云端CNN迁移融合的图片分类：Hierarchical Transfer Convolutional Neural Networks for Image Classification

• 20180402 CVPR-18 渐进式domain adaptation：Cross-Domain Weakly-Supervised Object Detection through Progressive Domain Adaptation

• 20180329 arXiv 基于attention机制的多任务学习：End-to-End Multi-Task Learning with Attention

• 20180326 arXiv 将迁移学习用于Faster R-CNN对象识别中：Domain Adaptive Faster R-CNN for Object Detection in the Wild

• 20180326 Pattern Recognition-17 多标签迁移学习方法应用于脸部属性分类：Multi-label Learning Based Deep Transfer Neural Network for Facial Attribute Classification

• 20180326 类似于ResNet的思想，在传统layer的ReLU之前加一个additive layer进行domain adaptation，思想简洁，效果非常好：Layer-wise domain correction for unsupervised domain adaptation

• 20180326 Pattern Recognition-17 基于Batch normalization提出了AdaBN，很简单：Adaptive Batch Normalization for practical domain adaptation

• 20180309 arXiv 利用已有网络的先验知识来加速目标网络的训练：Transfer Automatic Machine Learning

• 2018 ICLR-18 最小熵领域对齐方法 Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation code

• 2018 arXiv 最新发表在arXiv上的深度domain adaptation的综述：Deep Visual Domain Adaptation: A Survey

• ICLR-17 Central Moment Discrepancy (CMD) for Domain-Invariant Representation Learning

• ICCV-17 AutoDIAL: Automatic DomaIn Alignment Layers

• ICCV-17 CCSA: Unified Deep Supervised Domain Adaptation and Generalization

• ICML-17 JAN: Deep Transfer Learning with Joint Adaptation Networks

• 2017 Google: Learning Transferable Architectures for Scalable Image Recognition

• NIPS-16 RTN: Unsupervised Domain Adaptation with Residual Transfer Networks

• CoRR abs/1603.04779 (2016) AdaBN: Revisiting batch normalization for practical domain adaptation

• JMLR-16 DANN: Domain-adversarial training of neural networks

• 20171226 NIPS 2016 把传统工作搬到深度网络中的范例：不是只学习domain之间的共同feature，还学习每个domain specific的feature。这篇文章写得非常清楚，通俗易懂！ Domain Separation Networks | 代码

• 20171222 ICCV 2017 对于target中只有很少量的标记数据，用深度网络结合孪生网络的思想进行泛化：Unified Deep Supervised Domain Adaptation and Generalization | 代码和数据

• 20171126 NIPS-17 Label Efficient Learning of Transferable Representations acrosss Domains and Tasks

• 李飞飞小组发在NIPS 2017的文章。针对不同的domain、不同的feature、不同的label space，统一学习一个深度网络进行表征。

• 201711 一个很好的深度学习+迁移学习的实践教程，有代码有数据，可以直接上手：基于深度学习和迁移学习的识花实践

• ECCV-16 Deep CORAL: Correlation Alignment for Deep Domain Adaptation

• ECCV-16 DRCN: Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation

• ICML-15 DAN: Learning Transferable Features with Deep Adaptation Networks

• ICML-15 GRL: Unsupervised Domain Adaptation by Backpropagation

• ICCV-15 Simultaneous Deep Transfer Across Domains and Tasks

• NIPS-14 How transferable are features in deep neural networks?

• CoRR abs/1412.3474 (2014) Deep Domain Confusion(DDC): Maximizing for Domain Invariance

• ICML-14 著名的DeCAF特征：DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition

• Simultaneous Deep Transfer Across Domains and Tasks

• 发表在ICCV-15上，在传统深度迁移方法上又加了新东西

• 我的解读

• 深度适配网络（Deep Adaptation Network, DAN）

• 发表在ICML-15上：learning transferable features with deep adaptation networks

• 我的解读

• 深度联合适配网络（Joint Adaptation Network, JAN）

• Deep Transfer Learning with Joint Adaptation Networks

• 延续了之前的DAN工作，这次考虑联合适配

1.8.2. Deep Adversarial Transfer Learning (对抗迁移学习)

• 20191214 arXiv Learning Domain Adaptive Features with Unlabeled Domain Bridges

• Learning domain adaptive features with unlabeled CycleGAN

• 20191214 AAAI-20 Adversarial Domain Adaptation with Domain Mixup

• Domain adaptation with data mixup

• 20190916 arXiv Compound Domain Adaptation in an Open World

• Domain adaptation using the target domain knowledge

• 使用目标域的知识来进行domain adaptation

• 20101008 ICCV-19 Enhancing Adversarial Example Transferability with an Intermediate Level Attack

• Enhancing adversarial examples transerability

• 增强对抗样本的可迁移性

• 20190531 arXiv Adaptive Deep Kernel Learning

• Adaptive deep kernel learning

• 自适应深度核学习

• 20190531 arXiv Multi-task Learning in Deep Gaussian Processes with Multi-kernel Layers

• Multi-task learning in deep Gaussian process

• 深度高斯过程中的多任务学习

• 20190531 arXiv Image Alignment in Unseen Domains via Domain Deep Generalization

• Deep domain generalization for image alignment

• 深度领域泛化用于图像对齐

• 20190408 arXiv DeceptionNet: Network-Driven Domain Randomization

• Using only source data for domain randomization

• 仅利用源域数据进行domain randomization

• 20190220 arXiv Fully-Featured Attribute Transfer

• Fully-featured image attribute transfer

• 图像特征迁移

• 20190220 arXiv Unsupervised Domain Adaptation using Deep Networks with Cross-Grafted Stacks

• Domain adaptation using deep learning with cross-grafted stacks

• 用跨领域嫁接栈进行domain adaptation

• 20181217 arXiv DLOW: Domain Flow for Adaptation and Generalization

• Domain flow for adaptation and generalization

• 域流方法应用于领域自适应和扩展

• 20181212 arXiv Learning Transferable Adversarial Examples via Ghost Networks

• Use ghost networks to learn transferrable adversarial examples

• 使用ghost网络来学习可迁移的对抗样本

• 20181211 arXiv Adversarial Transfer Learning

• A survey on adversarial domain adaptation

• 一个关于对抗迁移的综述，特别用在domain adaptation上

• 20181205 arXiv Unsupervised Domain Adaptation using Generative Models and Self-ensembling

• UDA using CycleGAN

• 基于CycleGAN的domain adaptation

• 20181205 arXiv VADRA: Visual Adversarial Domain Randomization and Augmentation

• Domain randomization and augmentation

• Domain randomization和增强

• 20181130 arXiv Identity Preserving Generative Adversarial Network for Cross-Domain Person Re-identification

• Cross-domain reID

• 跨领域的行人再识别

• 20181129 AAAI-19 Exploiting Coarse-to-Fine Task Transfer for Aspect-level Sentiment Classification

• Aspect-level sentiment classification

• 迁移学习用于情感分类

• 20181128 arXiv Geometry-Consistent Generative Adversarial Networks for One-Sided Unsupervised Domain Mapping

• CycleGAN for domain adaptation

• CycleGAN用于domain adaptation

• 20181127 arXiv Distorting Neural Representations to Generate Highly Transferable Adversarial Examples

• Generate transferrable examples to fool networks

• 生成一些可迁移的对抗样本来迷惑神经网络，在各个网络上都表现好

• 20181123 arXiv Progressive Feature Alignment for Unsupervised Domain Adaptation
  
• Progressively selecting confident pseudo labeled samples for transfer
  
• 渐进式选择置信度高的伪标记进行迁移

• 20181113 NIPS-18 Conditional Adversarial Domain Adaptation

• Using conditional GAN for domain adaptation

• 用conditional GAN进行domain adaptation

• 20181107 NIPS-18 Invariant Representations without Adversarial Training

• Get invariant representations without adversarial training

• 不进行对抗训练获得不变特征表达

• 20181105 arXiv Efficient Multi-Domain Dictionary Learning with GANs

• Dictionary learning for multi-domains using GAN

• 用GAN进行多个domain的字典学习

• 20181012 arXiv Domain Confusion with Self Ensembling for Unsupervised Adaptation

• Domain confusion and self-ensembling for DA

• 用于Domain adaptation的confusion和self-ensembling方法

• 20180912 arXiv Improving Adversarial Discriminative Domain Adaptation

• Improve ADDA using source domain labels

• 提高ADDA方法的精度，使用source domain的label

• 20180731 ECCV-18 Dist-GAN: An Improved GAN using Distance Constraints

• Embed an autoencoder in GAN to improve its stability in training and propose two distances

• 将autoencoder集成到GAN中，提出相应的两种距离进行度量，提高了GAN的稳定性
• Code: Tensorflow

• 20180724 arXiv Generalization Bounds for Unsupervised Cross-Domain Mapping with WGANs

• Provide a generalization bound for unsupervised WGAN in transfer learning

• 对迁移学习中无监督的WGAN进行了一些理论上的分析

• 20180724 ECCV-18 Unsupervised Image-to-Image Translation with Stacked Cycle-Consistent Adversarial Networks

• Using stacked CycleGAN to perform image-to-image translation

• 用stacked cycleGAN进行image-to-image的翻译

• 20180628 ICML-18 Pixel-level和feature-level的domain adaptation：CyCADA: Cycle-Consistent Adversarial Domain Adaptation

• 20180619 CVPR-18 将optimal transport加入adversarial中进行domain adaptation：Re-weighted Adversarial Adaptation Network for Unsupervised Domain Adaptation

• 20180616 CVPR-18 用GAN进行domain adaptation：Generate To Adapt: Aligning Domains using Generative Adversarial Networks

• 20180612 ICML-18 利用多个数据集辅助，从而提高目标领域的学习能力：RadialGAN: Leveraging multiple datasets to improve target-specific predictive models using Generative Adversarial Networks

• 20180612 ICML-18 利用GAN进行多个domain的联合分布优化：JointGAN: Multi-Domain Joint Distribution Learning with Generative Adversarial Nets

• 20180605 arXiv NAM: Non-Adversarial Unsupervised Domain Mapping

• 20180508 arXiv 利用GAN，从有限数据中生成另一个domain的数据：Transferring GANs: generating images from limited data

• 20180501 arXiv Open set domain adaptation的对抗网络版本：Open Set Domain Adaptation by Backpropagation

• 20180427 arXiv 提出了adversarial residual的概念，进行深度对抗迁移：Unsupervised Domain Adaptation with Adversarial Residual Transform Networks

• 20180424 CVPR-18 用GAN和迁移学习进行图像增强：Adversarial Feature Augmentation for Unsupervised Domain Adaptation

• 20180413 arXiv 一种思想非常简单的深度迁移方法，仅考虑进行domain之间的类别概率矫正就能取得非常好的效果：Simple Domain Adaptation with Class Prediction Uncertainty Alignment

• 20180413 arXiv Mingming Gong提出的用因果生成网络进行深度迁移：Causal Generative Domain Adaptation Networks

• 20180410 CVPR-18(oral) 用两个分类器进行对抗迁移：Maximum Classifier Discrepancy for Unsupervised Domain Adaptation 代码

• 20180403 CVPR-18 将样本权重应用于对抗partial transfer中：Importance Weighted Adversarial Nets for Partial Domain Adaptation

• 20180326 MLSP-17 把domain separation network和对抗结合起来，提出了一个对抗网络进行迁移学习：Adversarial domain separation and adaptation

• 20180326 ICIP-17 类似于domain separation network，加入了对抗判别训练，同时优化分类、判别、相似度三个loss：Semi-supervised domain adaptation via convolutional neural network

• 20180312 arXiv 来自Google Brain团队的Wasserstein Auto-Encoders 代码

• 20180226 CVPR-18 当源域的类别大于目标域的类别时，如何进行迁移学习？Partial Transfer Learning with Selective Adversarial Networks

• 20180116 ICLR-18 用对偶的形式替代对抗训练中原始问题的表达，从而进行分布对齐 Stable Distribution Alignment using the Dual of the Adversarial Distance

• 20180111 arXiv 在GAN中用原始问题的对偶问题替换max问题，使得梯度更好收敛 Stable Distribution Alignment Using the Dual of the Adversarial Distance

• 20180110 AAAI-18 将Wasserstein GAN用到domain adaptaiton中 Wasserstein Distance Guided Representation Learning for Domain Adaptation

• 20171218 arXiv Partial Transfer Learning with Selective Adversarial Networks

• 假设target domain中的class是包含在source domain中，然后进行选择性的对抗学习

• 201707 发表在CVPR-17上，目前最好的对抗迁移学习文章：Adversarial Representation Learning For Domain Adaptation

• AAAI-18 Multi-Adversarial Domain Adaptation

• ICCV-17 CycleGAN: Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks

• ICCV-17 DualGAN: DualGAN: Unsupervised Dual Learning for Image-to-Image Translation

• CVPR-17 Asymmetric Tri-training for Unsupervised Domain Adaptation

• ICML-17 DiscoGAN: Learning to Discover Cross-Domain Relations with Generative Adversarial Networks

1.9. Multi-task Learning (多任务学习)

• 20200927 Knowledge Distillation for Multi-task Learning

• 针对多任务学习的知识蒸馏

• 20200914 ECML-PKDD-20 Towards Interpretable Multi-Task Learning Using Bilevel Programming

• 用bilevel programming解释多任务学习

• 20191202 arXiv AdaShare: Learning What To Share For Efficient Deep Multi-Task Learning

• Learning what to share for multi-task learning

• 对多任务学习如何share

• 20191125 AAAI-20 Adaptive Activation Network and Functional Regularization for Efficient and Flexible Deep Multi-Task Learning

• Adaptive activation network for deep multi-task learning

• 自适应的激活网络用于深度多任务学习

• 20191015 arXiv Gumbel-Matrix Routing for Flexible Multi-task Learning

• Effective method for flexible multi-task learning

• 一种很有效的方法用于多任务学习

• 20190718 arXiv Task Selection Policies for Multitask Learning

• Task selection in multitask learning

• 在多任务学习中的任务选择机制

• 20190509 FG-19 Multi-task human analysis in still images: 2D/3D pose, depth map, and multi-part segmentation

• Multi-task human analysis in still images

• 多任务人体静止图像分析

• 20190409 NAACL-19 AutoSeM: Automatic Task Selection and Mixing in Multi-Task Learning

• Automatic Task Selection and Mixing in Multi-Task Learning

• 多任务学习中自动任务选择和混淆

• 20190409 TNNLS-19 Heterogeneous Multi-task Metric Learning across Multiple Domains

• Heterogeneous Multi-task Metric Learning across Multiple Domains

• 在多个领域之间进行异构多任务度量学习

• 20190409 NeurIPS-18 Synthesized Policies for Transfer and Adaptation across Tasks and Environments

• Transfer across tasks and environments

• 通过任务和环境之间进行迁移

• 20190408 ICMR-19 Learning Task Relatedness in Multi-Task Learning for Images in Context

• Using task relatedness in multi-task learning

• 在多任务学习中学习任务之间的相关性

• 20190408 CVPR-19 End-to-End Multi-Task Learning with Attention

• End-to-End Multi-Task Learning with Attention

• 基于attention的端到端的多任务学习

• 20190401 arXiv Many Task Learning with Task Routing

• From multi-task leanring to many-task learning

• 许多任务同时学习

• 20190324 arXiv A Principled Approach for Learning Task Similarity in Multitask Learning

• Provide some theoretical analysis of the similarity learning in multi-task learning

• 为多任务学习中的相似度学习提供了一些理论分析

• 20181128 arXiv A Framework of Transfer Learning in Object Detection for Embedded Systems

• A Framework of Transfer Learning in Object Detection for Embedded Systems

• 一个用于嵌入式系统的迁移学习框架

• 20181012 NIPS-18 Multi-Task Learning as Multi-Objective Optimization

• Solve the multi-task learning as a multi-objective optimization problem

• 将多任务问题看成一个多目标优化问题进行求解

• 20181008 PSB-19 The Effectiveness of Multitask Learning for Phenotyping with Electronic Health Records Data

• Evaluate the effectiveness of multitask learning for phenotyping

• 评估多任务学习对于表型的作用

• 20180828 arXiv Self-Paced Multi-Task Clustering

• Multi-task clustering

• 多任务聚类

• 20180622 arXiv 探索了多任务迁移学习中的不确定性：Uncertainty in Multitask Transfer Learning

• 20180524 arXiv 杨强团队、与之前的learning to learning类似，这里提供了一个从经验中学习的learning to multitask框架：Learning to Multitask

1.10. Transfer Reinforcement Learning (强化迁移学习)

• 20191214 arXiv Does Knowledge Transfer Always Help to Learn a Better Policy?

• Transfer learning in reinforcement learning

• 20191212 AAAI-20 Transfer value iteration networks

• Transferred value iteration networks

• 20190821 arXiv Transfer in Deep Reinforcement Learning using Knowledge Graphs

• Use knowledge graph to transfer in reinforcement learning

• 用知识图谱进行强化迁移

• 20190320 arXiv Learning to Augment Synthetic Images for Sim2Real Policy Transfer

• Augment synthetic images for sim to real policy transfer

• 学习对于策略迁移如何合成图像

• 20190305 arXiv [Sim-to-Real Transfer for Biped Locomotion]

• Transfer learning for robot locomotion

• 用迁移学习进行机器人定位

• 20190220 arXiv DIViS: Domain Invariant Visual Servoing for Collision-Free Goal Reaching

• Transfer learning for robot reinforcement learning

• 迁移学习用于机器人的强化学习目标搜寻

• 20181212 NeurIPS-18 workshop Efficient transfer learning and online adaptation with latent variable models for continuous control

• Reinforcement transfer learning with latent models

• 隐变量模型用于迁移强化学习的控制

• 20181128 arXiv Hardware Conditioned Policies for Multi-Robot Transfer Learning

• Hardware Conditioned Policies for Multi-Robot Transfer Learning

• 多个机器人之间的迁移学习

• 20180926 arXiv Target Transfer Q-Learning and Its Convergence Analysis

• Analyze the risk of transfer q-learning

• 提供了在Q learning的任务迁移中一些理论分析

• 20180926 arXiv Domain Adaptation in Robot Fault Diagnostic Systems

• Apply domain adaptation in robot fault diagnostic system

• 将domain adaptation应用于机器人故障检测系统

• 20180912 arXiv VPE: Variational Policy Embedding for Transfer Reinforcement Learning

• Policy transfer in reinforcement learning

• 增强学习中的策略迁移

• 20180909 arXiv Transferring Deep Reinforcement Learning with Adversarial Objective and Augmentation

• deep + adversarial + reinforcement learning transfer

• 深度对抗迁移学习用于强化学习

• 20180530 ICML-18 强化迁移学习：Importance Weighted Transfer of Samples in Reinforcement Learning

• 20180524 arXiv 用深度强化学习的方法学习domain adaptation中的采样策略：Learning Sampling Policies for Domain Adaptation

• 20180516 arXiv 探索了强化学习中的任务迁移：Adversarial Task Transfer from Preference

• 20180413 NIPS-17 基于后继特征迁移的强化学习：Successor Features for Transfer in Reinforcement Learning

• 20180404 IEEE TETCI-18 用迁移学习来玩星际争霸游戏：StarCraft Micromanagement with Reinforcement Learning and Curriculum Transfer Learning

1.11. Transfer Metric Learning (迁移度量学习)

• 20190515 TNNLS-19 A Distributed Approach towards Discriminative Distance Metric Learning

• Discriminative distance metric learning

• 分布式度量学习

• 20190409 TNNLS-19 Heterogeneous Multi-task Metric Learning across Multiple Domains

• Heterogeneous Multi-task Metric Learning across Multiple Domains

• 在多个领域之间进行异构多任务度量学习

• 20190409 PAMI-19 Transferring Knowledge Fragments for Learning Distance Metric from A Heterogeneous Domain

• Heterogeneous transfer metric learning by transferring fragments

• 通过迁移知识片段来进行异构迁移度量学习

• 20190409 arXiv Decomposition-Based Transfer Distance Metric Learning for Image Classification

• Transfer metric learning based on decomposition

• 基于特征向量分解的迁移度量学习

• 20181012 arXiv Transfer Metric Learning: Algorithms, Applications and Outlooks

• A survey on transfer metric learning

• 一篇迁移度量学习的综述

• 20180622 arXiv 基于深度迁移学习的度量学习：DEFRAG: Deep Euclidean Feature Representations through Adaptation on the Grassmann Manifold

• 20181117 arXiv Distance Measure Machines

• Machines that measures distances

• 衡量距离的算法

• 20180605 KDD-10 迁移度量学习：Transfer metric learning by learning task relationships

• 20180606 arXiv 将流形和统计信息联合起来构成一个domain adaptation框架：A Unified Framework for Domain Adaptation using Metric Learning on Manifolds

• 20180605 CVPR-15 深度度量迁移学习：Deep metric transfer learning

• 20180425 arXiv 探索各个层对于迁移任务的作用，方便以后的迁移。比较有意思：CactusNets: Layer Applicability as a Metric for Transfer Learning

1.12. Transitive Transfer Learning (传递迁移学习)

• 传递迁移学习的第一篇文章，来自杨强团队，发表在KDD-15上：Transitive Transfer Learning

• AAAI-17 杨强团队最新的传递迁移学习：Distant Domain Transfer Learning

• 20180819 LNCS-2018 Distant Domain Adaptation for Text Classification

• Propose a selected algorithm for distant domain text classification

• 提出一个用于远域的文本分类方法

1.13. Lifelong Learning (终身迁移学习)

• 20190912 NeurIPS-19 Meta-Learning with Implicit Gradients

• Meta-learning with implicit gradients

• 隐式梯度的元学习

• 20180323 arXiv 终身迁移学习与增量学习结合：Incremental Learning-to-Learn with Statistical Guarantees

• 20180111 arXiv 一种新的终身学习框架，与L2T的思路有一些类似 Lifelong Learning for Sentiment Classification

1.14. Negative Transfer (负迁移)

• 20181128 arXiv Characterizing and Avoiding Negative Transfer
• Analyzing and formalizing negative transfer, then propose a new method
• 分析并形式化负迁移，进而提出自己的方法

1.15. Transfer Learning Applications (应用)

See HERE for a full list of transfer learning applications.