Publications

We introduce a novel architectural method for FedDG, namely gPerXAN, which relies on a normalization scheme working with a guiding regularizer. In particular, we carefully design Personalized Xplicitly Assembled Normalization to enforce client models selectively filtering domain-specific features that are biased towards local data while retaining discrimination of those features.

Le Huy Khiem, Long Tuan Ho, Cuong Do, Danh Le-Phuoc, Kok-Seng Wong

Conference on Computer Vision and Pattern Recognition 2024 (CVPR’24)

[Paper] [Code]

We propose a novel framework, called Correlated Low-rank Structure (CoLR), which leverages the concept of adjusting lightweight trainable parameters while keeping most parameters frozen.

Ngoc-Hieu Nguyen, Tuan-Anh Nguyen, Tuan Nguyen, Vu Tien Hoang, Dung D Le, and Kok-Seng Wong

The Web Conference 2024 (WWW’24)

[Paper] [Code] [Talk]

We propose a simple and lightweight defense against black-box attacks by adding random noise to hidden features at intermediate layers of the model at inference time.

Quang H Nguyen, Yingjie Lao, Tung Pham, Kok-Seng Wong, and Khoa D Doan

Twelfth International Conference on Learning Representations (ICLR’24)

[Paper]

In this paper, we propose a novel framework, namely Personalized Privacy-Preserving Federated Learning (PPPFL), with a concentration on cross-silo FL to overcome these challenges.

Van-Tuan Tran, Huy-Hieu Pham, Kok-Seng Wong

IEEE Transactions on Emerging Topics in Computing (2024)

[Paper] [Code]

We develop a Federated Few-Shot Learning framework, FedFSLAR, that collaboratively learns the classification model from multiple FL clients to recognize unseen actions with a few labeled video samples

Nguyen Anh Tu, Assanali Abu, Nartay Aikyn, Nursultan Makhanov, Min-Ho Lee, Khiem Le-Huy, and Kok-Seng Wong

IEEE/CVF Winter Conference on Applications of Computer Vision (WACV2024) Workshop

[Paper]

We provide a comprehensive survey of current backdoor attack strategies and defenses in FL, including a comprehensive analysis of different approaches.

Thuy Dung Nguyen, Tuan Nguyen, Phi Le Nguyen, Hieu H Pham, Khoa Doan, Kok-Seng Wong

2024 Engineering Applications of Artificial Intelligence

[Paper]

We proposed a framework that offers a more effective, stealthy, and durable approach to backdoor attacks in Federated Learning (FL).

Dung Thuy Nguyen, Tuan Minh Nguyen, Anh Tuan Tran, Khoa D. Doan, and Kok-Seng Wong

Thirty-seventh Conference on Neural Information Processing Systems (NeurIPS’23)

[Paper] [Code] [Talk]

We present an empirical study to investigate the impacts of various unlearning methods in diverse scenarios in Federated Learning

Thai-Hung Nguyen, Hong-Phuc Vu, Dung Thuy Nguyen, Tuan Minh Nguyen, Khoa D Doan, Kok-Seng Wong

Proceedings of Machine Learning Research (ACML’23)

[Paper] [Code]

In this paper, we present the frst defensive strategy that utilizes a unique characteristic of ViTs against backdoor attacks.

Khoa D Doan, Yingjie Lao, Peng Yang, Ping Li

2023 AAAI Conference on Artificial Intelligence

[Paper]

We introduce FedDCT, a novel distributed learning paradigm that enables the usage of large, high-performance CNNs on resource-limited edge devices.

Quan Nguyen, Hieu H Pham, Kok-Seng Wong, Phi Le Nguyen, Truong Thao Nguyen, Minh N Do

2023 IEEE Transactions on Network and Service Management

[Paper] [Code]

In this paper, we show empirically that the proposed framework achieves high attack performance (e.g., 100% attack success rates in several experiments) while preserving the cleandata performance in several benchmark datasets, including MNIST, CIFAR10, GTSRB, and TinyImageNet.

Khoa D Doan, Yingjie Lao, Ping Li

2022 The Conference and Workshop on Neural Information Processing Systems

[Paper]

In this paper, we propose a computationally efficient distributional distance by leveraging the discrete property of the hash functions.

Doan Khoa D, Peng Yang, Ping Li

2022 Conference on Computer Vision and Pattern Recognition

[Paper]

The proposed model also exhibits significant robustness toward out-of-distribution query data and is able to overcome missing data in both the training and testing phase with minimal retrieval performance degradation. Extensive experiments on several real-world datasets demonstrate superior results in which the proposed model achieves up to 5% improvement over the current state-of-the-art supervised hashing methods and exhibits a significant performance boost and robustness in both out-of-distribution retrieval and missing data scenarios.

Khoa D Doan, Sarkhan Badirli, Chandan K Reddy

2022 Conference on Asian Conference on Computer Vision

[Paper]

This paper proposes a two-hop communication protocol with a dynamic resource allocation strategy to investigate the possibility of bandwidth allocation from a limited network resource to the maximum number of clients participating in FL.

Tu Viet Nguyen, Nhan Duc Ho, Hieu Thien Hoang, Cuong Danh Do, Kok-Seng Wong

IEEE Access 2022

[Paper]

This paper aims to find a trade-off between our privacy protection method and data utility for medical images. Specifically, we propose a solution to anonymize chest X-ray images by directly adding noise to the images to prevent verification attacks and evaluate how well those images can maintain good performance in the lung disease classification task.

Truong Giang Vu, Nursultan Makhanov, Nguyen Anh Tu, Kok-Seng Wong

2022 International Conference on Advanced Technologies for Communications (ATC)

[Paper]

This paper discusses the emerging privacy and trust issues that are essential to motivate the acceptance of autonomous vehicles operating on public roads.

Thai-Hung Nguyen, Truong Giang Vu, Huong-Lan Tran, Kok-Seng Wong

2022 International Conference on Information Networking (ICOIN)

[Paper]

In this paper, we transform the private integer comparison into a block comparison problem.In particular, we employ a block vectorization mechanism to encode the private inputs into blocks.

Thai-Hung Nguyen, Kok-Seng Wong, Thomas Oikonomou

IEEE Access 2021

[Paper]

In this paper, we propose a novel and stealthy backdoor attack framework, LIRA, which jointly learns the optimal, stealthy trigger injection function and poisons the model. We formulate such an objective as a non-convex, constrained optimization problem.

Khoa Doan, Yingjie Lao, Weijie Zhao, Ping Li

2021 International Conference on Computer Vision

[Paper] [Code]

In this paper, we extend the concept of imperceptible backdoor from the input space to the latent representation, which significantly improves the effectiveness against the existing defense mechanisms, especially those relying on the distinguishability between clean inputs and backdoor inputs in latent space.

Khoa Doan, Yingjie Lao, Weijie Zhao, Ping Li

2021 The Conference and Workshop on Neural Information Processing Systems

[Paper]