Ye Lin Tun

Federated Learning (FL) has emerged as a decentralized machine learning technique, allowing clients to train a global model collaboratively without sharing private data. However, most FL studies ignore the crucial challenge of heterogeneous domains where each client has a distinct feature distribution, which is popular in real-world scenarios. Prototype learning, which leverages the mean feature vectors within the same classes, has become a prominent solution for federated learning under domain shift. However, existing federated prototype learning methods focus soley on inter-domain prototypes and neglect intra-domain perspectives. In this work, we introduce a novel federated prototype learning method, namely I2PFL, which incorporates Intra-domain and Inter-domain Prototypes, to mitigate domain shift from both perspectives and learn a generalized global model across multiple domains in federated learning. To construct intra-domain prototypes, we propose feature alignment with MixUp-based augmented prototypes to capture the diversity within local domains and enhance the generalization of local features. Additionally, we introduce a reweighting mechanism for inter-domain prototypes to generate generalized prototypes that reduce domain shift while providing inter-domain knowledge across multiple clients. Extensive experiments on the Digits, Office-10, and PACS datasets illustrate the superior performance of our method compared to other baselines.

Beyond the success of Contrastive Language-Image Pre-training (CLIP), recent trends have increasingly marked a shift toward lightweight vision-language models for resource-constrained systems. However, such models often exhibit suboptimal performance when relying solely on a single image-text contrastive objective, spotlighting the need for more effective training mechanisms that guarantee robust cross-modal feature alignment. In this work, we propose CLIP-PING: Contrastive Language-Image Pre-training with Proximus Intrinsic Neighbors Guidance, a novel yet simple and efficient training paradigm designed to improve the performance of lightweight vision-language models with minimal computational overhead and lower data demands. Specifically, CLIP-PING bootstraps unimodal features extracted from arbitrary pre-trained encoders to leverage intrinsic guidance of proximus neighbor samples, i.e., nearest-neighbors (NN) and cross nearest-neighbors (XNN). We find that the additional contrastive supervision induced by these neighbors substantially strengthens cross-modal alignment, enabling lightweight models to learn more generic features with enriched semantic diversity. Extensive experiments reveal that CLIP-PING notably surpasses its peers in zero-shot generalization and cross-modal retrieval tasks. In particular, a 5.5% improvement on zero-shot ImageNet1K classification, along with gains of 10.7% image-to-text (I2T) and 5.7% text-to-image (T2I) on Flickr30K retrieval, compared to the original CLIP when using ViT-XS image encoder trained on 3 million (image, text) pairs. Moreover, CLIP-PING showcases strong downstream transferability under the linear evaluation protocol across multiple image classification benchmarks.

Combining different data modalities enables deep neural networks to tackle complex tasks more effectively, making multimodal learning increasingly popular. To harness multimodal data closer to end users, it is essential to integrate multimodal learning with privacy-preserving approaches like federated learning (FL). However, compared to conventional unimodal learning, multimodal setting requires dedicated encoders for each modality, resulting in larger and more complex models. Training these models requires significant resources, presenting a substantial challenge for FL clients operating with limited computation and communication resources. To address these challenges, we introduce LW-FedMML, a layer-wise federated multimodal learning approach which decomposes the training process into multiple stages. Each stage focuses on training only a portion of the model, thereby significantly reducing the memory and computational requirements. Moreover, FL clients only need to exchange the trained model portion with the central server, lowering the resulting communication cost. We conduct extensive experiments across various FL and multimodal learning settings to validate the effectiveness of our proposed method. The results demonstrate that LW-FedMML can compete with conventional end-to-end federated multimodal learning (FedMML) while significantly reducing the resource burden on FL clients. Specifically, LW-FedMML reduces memory usage by up to 2.7×, computational operations (FLOPs) by 2.4×, and total communication cost by 2.3×. We also explore a progressive training approach called Prog-FedMML. While it offers lesser resource efficiency than LW-FedMML, Prog-FedMML has the potential to surpass the performance of end-to-end FedMML, making it a viable option for scenarios with fewer resource constraints.

Semantic Communication (SemCom), notable for ensuring quality of service by jointly optimizing source and channel coding, effectively extracts data semantics, eliminates redundant information, and mitigates noise effects from wireless channel. However, most studies overlook multiple user scenarios and resource availability, limiting real-world applications. This paper addresses this gap by focusing on downlink communication from a base station to multiple users with varying computing capacities. Users employ variants of Swin transformer models for source decoding and a simple architecture for channel decoding. We propose a novel training procedure FRENCA, incorporating transfer learning and knowledge distillation to improve low-computing users’ performance. Extensive simulations validate the proposed methods.

Semantic communication has emerged as a pillar for the next generation of communication systems due to its capabilities in alleviating data redundancy. Most semantic communication systems are built upon advanced deep learning models whose training performance heavily relies on data availability. Existing studies often make unrealistic assumptions of a readily accessible data source, where in practice, data is mainly created on the client side. This circumstance limits the transmission of data due to privacy concerns, which is necessary for conventional centralized training schemes. To address this challenge, we explore semantic communication in a federated learning (FL) setting, which can harness the data of clients without compromising their privacy. Additionally, we design our system to lower the communication overhead by reducing the quantity of information delivered in each FL training round. This not only allows us to save significant bandwidth for resource-constrained devices but also reduces overall network traffic. Finally, we introduce a mechanism to aggregate the global model based on the client's performance, which we have referred to as FedLol. Extensive simulation results demonstrate the effectiveness of our proposed technique compared to baseline methods.

Semantic communication has gained significant attention from researchers as a promising technique to replace conventional communication in the next generation of communication systems, primarily due to its ability to reduce communication costs. However, little literature has studied its effectiveness in multi-user scenarios, particularly when there are variations in the model architectures used by users and their computing capacities. To address this issue, we explore a semantic communication system that caters to multiple users with different model architectures by using a multi-purpose transmitter at the base station (BS). Specifically, the BS in the proposed framework employs semantic and channel encoders to encode the image for transmission, while the receiver utilizes its local channel and semantic decoder to reconstruct the original image. Our joint source-channel encoder at the BS can effectively extract and compress semantic features for specific users by considering the signal-to-noise ratio (SNR) and computing capacity of the user. Based on the network status, the joint source-channel encoder at the BS can adaptively adjust the length of the transmitted signal. A longer signal ensures more information for high-quality image reconstruction for the user, while a shorter signal helps avoid network congestion. In addition, we propose a hybrid loss function for training, which enhances the perceptual details of reconstructed images. Finally, we conduct a series of extensive evaluations and ablation studies to validate the effectiveness of the proposed system.

Federated learning (FL) has emerged as a promising approach to collaboratively train machine learning models across multiple edge devices while preserving privacy. The success of FL hinges on the efficiency of participating models and their ability to handle the unique challenges of distributed learning. While several variants of Vision Transformer (ViT) have shown great potential as alternatives to modern convolutional neural networks (CNNs) for centralized training, the unprecedented size and higher computational demands hinder their deployment on resource-constrained edge devices, challenging their widespread application in FL. Since client devices in FL typically have limited computing resources and communication bandwidth, models intended for such devices must strike a balance between model size, computational efficiency, and the ability to adapt to the diverse and non-IID data distributions encountered in FL. To address these challenges, we propose OnDev-LCT: Lightweight Convolutional Transformers for On-Device vision tasks with limited training data and resources. Our models incorporate image-specific inductive biases through the LCT tokenizer by leveraging efficient depthwise separable convolutions in residual linear bottleneck blocks to extract local features, while the multi-head self-attention (MHSA) mechanism in the LCT encoder implicitly facilitates capturing global representations of images. Extensive experiments on benchmark image datasets indicate that our models outperform existing lightweight vision models while having fewer parameters and lower computational demands, making them suitable for FL scenarios with data heterogeneity and communication bottlenecks.

Diffusion models have shown great potential for vision-related tasks, particularly for image generation. However, their training is typically conducted in a centralized manner, relying on data collected from publicly available sources. This approach may not be feasible or practical in many domains, such as the medical field, which involves privacy concerns over data collection. Despite the challenges associated with privacy-sensitive data, such domains could still benefit from valuable vision services provided by diffusion models. Federated learning (FL) plays a crucial role in enabling decentralized model training without compromising data privacy. Instead of collecting data, an FL system gathers model parameters, effectively safeguarding the private data of different parties involved. This makes FL systems vital for managing decentralized learning tasks, especially in scenarios where privacy-sensitive data is distributed across a network of clients. Nonetheless, FL presents its own set of challenges due to its distributed nature and privacy-preserving properties. Therefore, in this study, we explore the FL strategy to train diffusion models, paving the way for the development of federated diffusion models. We conduct experiments on various FL scenarios, and our findings demonstrate that federated diffusion models have great potential to deliver vision services to privacy-sensitive domains.