weixin_44563093的博客

原创 【论文总结】 Multi-source Domain Adaptation (持续更新)

Paper notes for Multi-source Domain Adaptation

原创 [论文笔记] Maximum Classifier Discrepancy for Unsupervised Domain Adaptation

This paper introduces a new approach that attempts to align distributions of source and target by utilizing the task-specific decision boundaries

原创 [论文笔记] Classes Matter: A Fine-grained Adversarial Approach to Cross-domain Semantic Segmentation

This paper adopts a fine-grained domain discriminator that not only plays as a domain distinguisher but also differentiates domains at class level

原创 [论文笔记] WiFi-Based Cross-Domain Gesture Recognition via Modified Prototypical Networks

This paper proposes a WiFi-based cross-domain gesture recognition system (WiGr) which has a domain-transferable mapping to construct an embedding space where the representations of samples from the same class are clustered

转载 [论文笔记] A DIRT-T approach to unsupervised domain adaptation

In this paper, we address these issues through the lensof the cluster assumption, i.e., decision boundaries should not cross high-density data regions.

原创 [论文笔记] Super Low Resolution RF Powered Accelerometers for Alerting on Hospitalized Patient Bed Exits

This paper investigates a batteryless sensing modality with low cost wirelessly powered Radio Frequency Identification (RFID) technology with the potential for convenient integration into clothing, such as hospital gowns

转载 [论文笔记] Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks

原文链接：https://arxiv.org/abs/1703.10593Contributionspix2pix的image-to-image translation需要图像对，而获得这样的训练数据不仅困难而且昂贵。而目前不需图像对的方法通常是domain-specific的，或者需要假定一些先验，通用性不够，并且效果也不是太好。本文提出了cycle consistency loss（循环一致性损失），使得通用的 unpaired image-to-image translation...

原创 [论文笔记/综述类] A Survey of Unsupervised Deep Domain Adaptation (持续更新)

原文链接：https://arxiv.org/abs/1812.028491. IntroductionThis paperprovides anoverview of generative adversarial networks (GANs) to provide background for the increasingly widespread use of adversarial techniques in domain adaptation. investigate the v..

原创 [论文笔记] Domain Adaptation 论文整理(持续更新)

原创 [论文笔记] Adversarial Discriminative Domain Adaptation

原文链接：https://arxiv.org/abs/1702.05464（CVPR2017）Motivation在做区分的任务或者域偏移较大任务上之前的方法比较不那么令人满意。例子是：CoGANs。这种方法只在源域和目标域非常相似的情况下(如 MNIST 和 USPS )显示出优势，在本文实验中，很难使它收敛到更大的分布偏移。 在能够处理大偏差任务的方法上，大多采用了固定权重的方法来进行训练（主要就是 target domain 的分类那边，基本是源和目标共用一个分类器）。 其实很多方法都没.

原创 [DL常见问题]Issue: CUDA error 59: Device-side assert triggered

Problem: Run into a CUDA error during trainingSolutions:This error occurs due to the following two reasons:Inconsistency between the number of labels/classes and the number of output units The input of the loss function may be incorrectIn my case .

原创 [论文笔记] Learning Gestures From WiFi: A Siamese Recurrent Convolutional Architecture

原文链接：Learning Gestures From WiFi: A Siamese Recurrent Convolutional Architecture | IEEE Journals & Magazine | IEEE XploreMethodology [原文图片，侵删]System Design 上的改进1）文章利用OpenWrt, 一款 lightweight Linux OS 作为IoT platform 的OS来收集WiFi CSI的数据， 相较于传统的Intel 5

原创 [论文笔记] Consensus Adversarial Domain Adaptation

原文链接：https://ojs.aaai.org/index.php/AAAI/article/download/4552/4430文章提出了 Consensus Adversarial Domain Adaptation (CADA), a novel unsupervised ADA scheme that gives freedom to both target encoder and source encoder.ObjectiveTo improve the generalizati

转载 [论文笔记] Few-Shot Adversarial Domain Adaptation

原文链接：https://proceedings.neurips.cc/paper/2017/file/21c5bba1dd6aed9ab48c2b34c1a0adde-Paper.pdfgithub链接：GitHub - HX-idiot/FADA-Pytorch: pytorch implement for the paper Few-Shot Adversarial Domain AdaptationContributions处理只有少量数据的情形并且提供高效的训练。 扩展对抗学习的方法

原创 Boosting the Generalization Capability in Cross-Domain Few-shot Learning via Supervised Autoencoder

原文链接 https://openaccess.thecvf.com/content/ICCV2021/papers/Liang_Boosting_the_Generalization_Capability_in_Cross-Domain_Few-Shot_Learning_via_Noise-Enhanced_ICCV_2021_paper.pdfMotivationDifferent from general few-shot learning (FSL) where large-scal.

原创 [论文笔记] Domain-Adaptive Few-Shot Learning

原文链接：https://openaccess.thecvf.com/content/WACV2021/papers/Zhao_Domain-Adaptive_Few-Shot_Learning_WACV_2021_paper.pdfgithub链接：GitHub - dingmyu/DAPN: A pytorch implementation of "Domain-Adaptive Few-Shot Learning"（有bug和文件缺失，无法运行）Methodology (论文原图，侵删).

转载 [Pytorch 常用函数] 激活函数Relu, Leaky Relu

修正线性单元（Rectified linear unit，ReLU）是神经网络中最常用的激活函数。它保留了 step 函数的生物学启发（只有输入超出阈值时神经元才激活），不过当输入为正的时候，导数不为零，从而允许基于梯度的学习（尽管在 x=0 的时候，导数是未定义的）。使用这个函数能使计算变得很快，因为无论是函数还是其导数都不包含复杂的数学运算。然而，当输入为负值的时候，ReLU 的学习速度可能会变得很慢，甚至使神经元直接无效，因为此时输入小于零而梯度为零，从而其权重无法得到更新，在剩下的训练过程中会一..

原创 [论文笔记]CrossSense: Towards Cross-Site and Large-Scale WiFi Sensing

论文链接： FewSense, Towards a Scalable and Cross-Domain Wi-Fi Sensing System Using Few-Shot LearningIntroduction虽然当前的deep learning-based WiFi sensing 相关研究取得了不小的进展，但是依旧有不少的limitations限制了实际研究应用的部署 Limited Scalability：训练deep learning model需要采集大量的数据 Do

原创 [论文笔记] Context-Aware Wireless-Based Cross-DomainGesture Recognition

论文链接：Context-Aware Wireless-Based Cross-Domain Gesture Recognition | IEEE Journals & Magazine | IEEE XploreObjectiveMethodologyTrain ModelModel inputA matrix with dimension , where is the number of sampling points in the frequency domai...

原创 Issue: Resolving BrokenPipeError and Unknown CUDA Error

Problem:BrokenPipeError : is not going to be frozen to produce an executable.''')[Errno 32] Broken pipeSolution:insert if __name__=='_main_': for the main bodyProblem:RuntimeError: CUDA error: unknown errorSolution: - Running out of memor...

原创 Issue: Resolving new pip backtracking runtime issue

Problem:Run into a new pipbacktracking runtime issue when using pip3 install -r requirements.txt to install the required packagesSolution: Try adding--use-deprecated=legacy-resolverto the install command to use the old pip behavior :pip3 install...

原创 Issue: torch.cuda.is_available() returns False

Cuda Info:Run nvidia-smi:Problem: It turned out that using conda willinstalla CPU-only version by running the conda command provided by https://pytorch.org/get-started/locally/Solution: Use the pip command to install the torch. In my case it sho...