A5088981076- Privacy-Preserving Technologies in Data
- Cryptography and Data Security
- Stochastic Gradient Optimization Techniques
- Domain Adaptation and Few-Shot Learning
- Adversarial Robustness in Machine Learning
- Access Control and Trust
- Privacy, Security, and Data Protection
- Ovarian function and disorders
- Advanced X-ray and CT Imaging
- Medical Imaging Techniques and Applications
- Ocular and Laser Science Research
- Anomaly Detection Techniques and Applications
- Data Quality and Management
- Advanced battery technologies research
- Advanced Computational Techniques and Applications
- Advanced Graph Neural Networks
- Advanced Battery Materials and Technologies
- Digital Radiography and Breast Imaging
- AI in cancer detection
- Explainable Artificial Intelligence (XAI)
- Personal Information Management and User Behavior
- Ethics and Social Impacts of AI
- Neural Networks and Reservoir Computing
- Text and Document Classification Technologies
- Photoreceptor and optogenetics research
National University of Defense Technology
2021-2025
West Bengal Electronics Industry Development Corporation Limited (India)
2020-2023
Hong Kong University of Science and Technology
2022
University of Hong Kong
2022
Beijing Academy of Artificial Intelligence
2022
Hubei University of Technology
2022
Beijing University of Posts and Telecommunications
2022
Samsung (United States)
2021
Xi'an Institute of Optics and Precision Mechanics
2019
University of Chinese Academy of Sciences
2019
Machine learning relies on the availability of a vast amount data for training. However, in reality, most are scattered across different organizations and cannot be easily integrated under many legal practical constraints. In this paper, we introduce new technique framework, known as federated transfer (FTL), to improve statistical models federation. The federation allows knowledge shared without compromising user privacy, enables complimentary transferred network. As result, target-domain...
Federated learning (FL) is a rapidly growing research field in machine learning. However, existing FL libraries cannot adequately support diverse algorithmic development; inconsistent dataset and model usage make fair algorithm comparison challenging. In this work, we introduce FedML, an open library benchmark to facilitate development performance comparison. FedML supports three computing paradigms: on-device training for edge devices, distributed computing, single-machine simulation. also...
How is it possible to allow multiple data owners collaboratively train and use a shared prediction model while keeping all the local training private? Traditional machine learning approaches n
Vertical Federated Learning (VFL) is a federated learning setting where multiple parties with different features about the same set of users jointly train machine models without exposing their raw data or model parameters. Motivated by rapid growth in VFL research and real-world applications, we provide comprehensive review concept algorithms VFL, as well current advances challenges various aspects, including effectiveness, efficiency, privacy. We an exhaustive categorization for settings...
Machine Learning models require a vast amount of data for accurate training. In reality, most is scattered across different organizations and cannot be easily integrated under many legal practical constraints. Federated Transfer (FTL) was introduced in [1] to improve statistical federation that allow knowledge shared without compromising user privacy, enable complementary transferred the network. As result, target-domain party can build more flexible powerful by leveraging rich labels from...
We introduce a novel federated learning framework allowing multiple parties having different sets of attributes about the same user to jointly build models without exposing their raw data or model parameters. Conventional approaches are inefficient for cross-silo problems because they require exchange messages gradient updates at every iteration, and raise security concerns over sharing such during learning. propose <i>Federated Stochastic Block Coordinate Descent (FedBCD)</i> algorithm,...
Federated learning (FL) enables participating parties to collaboratively build a global model with boosted utility without disclosing private data information. Appropriate protection mechanisms have be adopted fulfill the opposing requirements in preserving privacy and maintaining high . In addition, it is mandate for federated system achieve efficiency order enable large-scale training deployment. We propose unified framework that reconciles horizontal vertical learning. Based on this...
We introduce a collaborative learning framework allowing multiple parties having different sets of attributes about the same user to jointly build models without exposing their raw data or model parameters. In particular, we propose Federated Stochastic Block Coordinate Descent (FedBCD) algorithm, in which each party conducts local updates before communication effectively reduce number rounds among parties, principal bottleneck for problems. analyze theoretically impact and show that when...
Federated learning allows multiple parties to build machine models collaboratively without exposing data. In particular, vertical federated (VFL) enables participating a joint model based on distributed features of aligned samples. However, VFL requires all share sufficient amount reality, the set samples may be small, leaving majority non-aligned data unused. this article, we propose Cross-view Training (FedCVT), semi-supervised approach that improves performance with limited More...
We present a novel privacy-preserving federated adversarial domain adaptation approach ($\textbf{PrADA}$) to address an under-studied but practical cross-silo problem, in which the party of target is insufficient both samples and features. lack-of-feature issue by extending feature space through vertical learning with feature-rich tackle sample-scarce performing from sample-rich source party. In this work, we focus on financial applications where interpretability critical. However, existing...
Highly concentrated salts, like 30 m ZnCl₂, can reduce free water molecules in aqueous electrolytes but also increase acidity, causing severe acid-catalyzed corrosion of the Zn anode, current collector, and...
Machine learning (ML) training data is often scattered across disparate collections of datasets, called silos. This fragmentation poses a major challenge for data-intensive ML applications: integrating and transforming residing in different sources demand lot manual work computational resources. With privacy security constraints, cannot leave the premises silos, hence model should proceed decentralized manner. In this work, we present vision how to bridge traditional integration (DI)...
Vertical federated learning (VFL) allows an active party with labeled data to leverage auxiliary features from the passive parties improve model performance. Concerns about private feature and label leakage in both training inference phases of VFL have drawn wide research attention. In this paper, we propose a general privacy-preserving vertical deep framework called FedPass, which leverages adaptive obfuscation protect simultaneously. Strong capabilities labels are theoretically proved (in...
Trustworthy federated learning typically leverages protection mechanisms to guarantee privacy. However, inevitably introduce utility loss or efficiency reduction while protecting data Therefore, and their parameters should be carefully chosen strike an optimal tradeoff among privacy leakage , . To this end, practitioners need tools measure the three factors optimize between them choose mechanism that is most appropriate application at hand. Motivated by requirement, we propose a framework...
Personalized Federated Continual Learning (PFCL) is a new practical scenario that poses greater challenges in sharing and personalizing knowledge. PFCL not only relies on knowledge fusion for server aggregation at the global spatial-temporal perspective but also needs model improvement each client according to local requirements. Existing methods, whether (PFL) or (FCL), have overlooked multi-granularity representation of knowledge, which can be utilized overcome Spatial-Temporal...
Federated learning allows multiple parties to build machine models collaboratively without exposing data. In particular, vertical federated (VFL) enables participating a joint model based upon distributed features of aligned samples. However, VFL requires all share sufficient amount reality, the set samples may be small, leaving majority non-aligned data unused. this article, we propose Cross-view Training (FedCVT), semi-supervised approach that improves performance with limited More...
Conventionally, federated learning aims to optimize a single objective, typically the utility. However, for system be trustworthy, it needs simultaneously satisfy multiple objectives, such as maximizing model performance, minimizing privacy leakage and training costs, being robust malicious attacks. Multi-Objective Optimization (MOO) aiming conflicting objectives is quite suitable solving optimization problem of Trustworthy Federated Learning (TFL). In this paper, we unify MOO TFL by...