A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Filters
Keep It Simple: Fault Tolerance Evaluation of Federated Learning with Unreliable Clients
[article]
2023
arXiv
pre-print
Preserved Fulltext
Other Versions
2023
2023 IEEE 16th International Conference on Cloud Computing (CLOUD)
unpublished
doi:10.1109/cloud60044.2023.00024 fatcat:jzfipt75kfgwxlutopqcotyjaa
doi:10.1109/cloud60044.2023.00024 fatcat:jzfipt75kfgwxlutopqcotyjaa
Federated learning (FL), as an emerging artificial intelligence (AI) approach, enables decentralized model training across multiple devices without exposing their local training data. ...
While research works have been proposed to improve the fault tolerance of FL, the real impact of unreliable devices (e.g., dropping out, misconfiguration, poor data quality) in real-world applications ...
What is the effect of training data volume on FL with upload, download) Which client/site drops out (sites Classification accuracy Robustness unreliable clients? ...
arXiv:2305.09856v1
fatcat:m6kvjidmhrcmjixr4l5itrasbu
Loss Adapted Plasticity in Deep Neural Networks to Learn from Data with Unreliable Sources
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
In many applications, sources can have varied levels of noise or corruption that has negative effects on the learning of a robust deep learning model. ...
We show that applying this technique can significantly improve model performance when trained on a mixture of reliable and unreliable data sources, and maintain performance when models are trained on data ...
However, in our testing, we allow for the transfer of data to the server, since we are not concerned with this constraint.
III. ...
arXiv:2212.02895v1
fatcat:yimb7jhh2vh7tpoqikiy44smqm
Robust Deep Sensing Through Transfer Learning in Cognitive Radio
[article]
2019
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We incorporate transfer learning into the framework to improve the robustness. Results validate the effectiveness as well as the robustness of the proposed deep spectrum sensing framework. ...
We propose a robust spectrum sensing framework based on deep learning. ...
With no labeled target data, the transfer is unreliable and depends on whether QPSK or Gaussian signals are the source or target. ...
arXiv:1908.00658v1
fatcat:uq53lqa2mbawhb52k3wu6k4o6e
Bimodal Vein Recognition Based on Task-Specific Transfer Learning
2017
IEICE transactions on information and systems
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
Both gender and identity recognition task with hand vein information is solved based on the proposed cross-selected-domain transfer learning model. ...
State-of-the-art recognition results demonstrate the effectiveness of the proposed model for pattern recognition task, and the capability to avoid over-fitting of fine-tuning DCNN with small-scaled database ...
transfer learning. ...
doi:10.1587/transinf.2017edl8031
fatcat:k5wgfqjzxbda7ce7appcxbe2zy
Leveraging Transfer Learning for Reliable Intelligence Identification on Vietnamese SNSs (ReINTEL)
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2020 pre-print arXiv:2012.07557v1 fatcat:5vrzfi7svjdvzcc2pliirhzd24
Besides, we utilize the ensemble method to improve the robustness of different approaches. ...
Experimental results show that transfer learning method can provide high efficiency in detecting fake news. Along with that is the potential of the ensemble method can be exploited. ...
Our source code 1 is publicly available. 1 https://github.com/heraclex12/VLSP2020-Fake-News-Detection
A. Data Pre-processing We process the text contents in two phases. ...
arXiv:2012.07557v2
fatcat:z26ohhq2hjh25jsb4mjqhghwzu
The Effectiveness of Transfer Learning in Electronic Health Records Data
2017
International Conference on Learning Representations
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Here we present initial results on the application of transfer learning to this problem. ...
The application of machine learning to clinical data from Electronic Health Records is limited by the scarcity of meaningful labels. ...
We aim to overcome this challenge through the application of transfer learning: we first train a neural network to predict a source task with ubiquitous labels that, while not directly related to our phenotype ...
dblp:conf/iclr/DuboisRJSK17
fatcat:73r3cwfnwfgbzncommwbc62oem
Domain Adaptive Video Semantic Segmentation via Cross-Domain Moving Object Mixing
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Other Versions
FATC exploits the robust source domain features, which are trained with ground truth labels, to learn discriminative target domain features in an unsupervised manner by filtering unreliable predictions ...
To address this problem, we propose Cross-domain Moving Object Mixing (CMOM) that cuts several objects, including hard-to-transfer classes, in the source domain video clip and pastes them into the target ...
FATC exploits the robust source domain features, which are trained with ground truth labels, to learn discriminative target domain features in an unsupervised manner by filtering unreliable predictions ...
arXiv:2211.02307v2
fatcat:dottxeitljhvlpng2ws2sfn6ue
Transfer Learning with Pretrained Neural Network Between Unrelated Tasks for Machine Health Diagnosis
2020
International journal of recent technology and engineering
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Another approach, the Transfer Learning (TL), had demonstrated in the literature that he can overcome these weaknesses. ...
Our method doesn't require a high amount of input data and thus saves a lot of time in retraining the network in another task, which can be related or unrelated to the source task. ...
requires the availability of labeled target data. Transductive transfer learning is used when labeled source data are available while labeled target data do not exist. And unsupervised transfer learning ...
doi:10.35940/ijrte.f8489.038620
fatcat:j46y7uprqbexdhjjyoyagvakum
Reliability Exploration with Self-Ensemble Learning for Domain Adaptive Person Re-identification
2022
AAAI Conference on Artificial Intelligence
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
In this paper, we propose a Reliability Exploration with Self-ensemble Learning (RESL) framework for domain adaptive person Re-ID. ...
Person re-identification (Re-ID) based on unsupervised domain adaptation (UDA) aims to transfer the pre-trained model from one labeled source domain to an unlabeled target domain. ...
Learning with Noise Recent studies on learning with noisy labels can broadly group to three categories: robust loss design (Wang et al. 2019) , label correction (Lee et al. 2018 ) and re-weighting methods ...
dblp:conf/aaai/LiSLCWZ22
fatcat:cchq6s5v3feoffsgpf3bep2zde
Transferring model structure in Bayesian transfer learning for Gaussian process regression
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
By successfully transferring higher moments of the source, the target can reject unreliable source knowledge (i.e. it achieves robust transfer). ...
Bayesian transfer learning (BTL) is defined in this paper as the task of conditioning a target probability distribution on a transferred source distribution. ...
The paper has presented evidence to show that this extension of the target ensures that unreliable source knowledge is rejected (i.e. robust transfer is achieved (Remark 2)), and our positive transfer ...
arXiv:2101.06884v1
fatcat:vlrq6kz27fertou2g4fenautli
AUGCO: Augmentation Consistency-guided Self-training for Source-free Domain Adaptive Semantic Segmentation
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2021 pre-print arXiv:2107.10140v1 fatcat:prvcvgjoqjgc7lyawcp6qcptxa
We focus on source-free domain adaptation for semantic segmentation, wherein a source model must adapt itself to a new target domain given only unlabeled target data. ...
Most modern approaches for domain adaptive semantic segmentation rely on continued access to source data during adaptation, which may be infeasible due to computational or privacy constraints. ...
SFDA makes use of data-free distillation via a dual attention module for knowledge transfer from the source Table 2 : SYNTHIA→Cityscapes: IoU on the Cityscapes validation set. ...
arXiv:2107.10140v2
fatcat:ligkvixukfbojdafzor6vxuydy
Communication Support for Knowledge-intensive Services
[chapter]
1998
Broadband Communications
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2018; you can also visit the original URL.
The file type is application/pdf.
Data sequences are transferred by using the unreliable IP-Multicast protocol. ...
This could be ensured by using sliding windows methods with numbered data packets to obtain the so-called source order. ...
doi:10.1007/978-0-387-35378-4_16
fatcat:h26rwznzare2pfwwcfuvgve6hi
FedGEMS: Federated Learning of Larger Server Models via Selective Knowledge Fusion
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2021 pre-print arXiv:2110.11027v1 fatcat:jz4cngjkxvan5irsd6ysoyjo7e
Today data is often scattered among billions of resource-constrained edge devices with security and privacy constraints. ...
Federated Learning (FL) has emerged as a viable solution to learn a global model while keeping data private, but the model complexity of FL is impeded by the computation resources of edge nodes. ...
RELATED WORK Federated Learning with larger server models. FL is a collaborating learning framework without sharing private data among the clients. ...
arXiv:2110.11027v2
fatcat:u533ihbv2veeji574pvqfmjfka
Secure and Reliable Transfer Learning Framework for 6G-enabled Internet of Vehicles
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Transfer Learning (TL) has great potential to empower promising 6G-enabled IoV, such as smart driving assistance, with its outstanding features including enhancing quality and quantity of training data ...
For instance, malicious vehicles in source domains may transfer and share untrustworthy models (i.e., knowledge) about connection availability to target domains, thus adversely affecting the performance ...
In contrast, labeled source data is available in multi-task learning. ...
arXiv:2111.05804v1
fatcat:ve7tneopdjadfov7ygwuititpe
Focused Adaptation of Dynamics Models for Deformable Object Manipulation
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Additionally, we combine this adaptation method with prior work on planning with unreliable dynamics to make a method for data-efficient online adaptation, called FOCUS. ...
In order to efficiently learn a dynamics model for a task in a new environment, one can adapt a model learned in a similar source environment. ...
The idea that transfer is easier when the source and target data are similar is well-supported in the transfer learning literature [4] , [5] . ...
arXiv:2209.14261v2
fatcat:au5afooc3jakvlnmlsybeo7ohi
« Previous
Showing results 1 — 15 out of 31,429 results