arXiv subjects

Seong-Lyun Kim

Publications and source records attributed to Seong-Lyun Kim.

At least 19 recordsLinked to original sources

arXiv2026

Channel2World: A Wireless Foundation Model for RF Environment Representation

Wireless channels are commonly treated as link-specific observations, although their multipath structure is governed by the surrounding radio-frequency (RF) environment. In this paper, we propose Channel2World, a wireless foundation model that learns a reusable environment-level representation from multiple-input multiple-output (MIMO) channel-position observations. The model aggregates channels collected within the same base-station-centered environment into a wireless world embedding using a Transformer-based encoder. The encoder is pretrained through context-query prediction, where context channels condition user equipment (UE) position and relative path-gain prediction for disjoint query channels. After pretraining, the encoder is frozen and used as a task-agnostic environment-conditioning module for downstream wireless models, enabling adaptation to unseen environments without site-specific fine-tuning. To learn an environment-level latent space that generalizes across deployments, we pretrain Channel2World using ray-tracing data from 26,000 environments, with approximately 5,000 channel measurements per environment. Evaluations on UE localization, beam-domain channel state information (CSI) reconstruction, and RF-observable geometry reconstruction show that the learned embeddings provide effective conditioning in unseen environments. For localization and CSI reconstruction tasks, embedding-based conditioning outperforms or remains competitive with site-specific fine-tuning, although fine-tuning requires task-specific labeled data and additional gradient-based adaptation. The embeddings also support the reconstruction of dominant reflector structures, indicating their utility as reusable environmental priors across tasks.

eess.SP↗

arXiv2026

Breaking the Capacity Bottleneck in Model-Heterogeneous Federated Learning via Gradual Model Restoration

Federated learning (FL) enables distributed model training, yet in heterogeneous deployments, Bandwidth-Constrained Clients (BCCs) often contribute inefficiently due to limited uplink bandwidth. In model-heterogeneous FL with fixed small sub-models, BCCs with sub-models may improve quickly in early rounds but become under-parameterized later, resulting in slow convergence and poor generalization. To address this challenge, we propose FedGMR, a federated learning framework centered around Gradual Model Restoration (GMR), where GMR progressively increases each client's sub-model density during training, allowing BCCs to remain effective contributors throughout optimization. To make GMR practical under real-world heterogeneity, FedGMR is realized as an end-to-end workflow with asynchronous coordination and stable, mask-aware aggregation. We further establish convergence guarantees, showing that the aggregation error scales with the average sub-model density across clients and rounds, and that GMR provably narrows the gap toward full-model FL. Extensive experiments on FEMNIST, CIFAR-10, ImageNet-100, and StackOverflow demonstrate that FedGMR improves both convergence speed and final accuracy, especially under severe heterogeneity and non-IID data distributions. Code is available at https://github.com/machengjie321/ICML2026-FedGMR.

cs.DC↗

arXiv2026

Communication-Efficient Hybrid Language Model via Uncertainty-Aware Opportunistic and Compressed Transmission

To support emerging language-based applications using dispersed and heterogeneous computing resources, the hybrid language model (HLM) offers a promising architecture, where an on-device small language model (SLM) generates draft tokens that are validated and corrected by a remote large language model (LLM). However, the original HLM suffers from substantial communication overhead, as the LLM requires the SLM to upload the full vocabulary distribution for each token. Moreover, both communication and computation resources are wasted when the LLM validates tokens that are highly likely to be accepted. To overcome these limitations, we propose communication-efficient and uncertainty-aware HLM (CU-HLM). In CU-HLM, the SLM transmits truncated vocabulary distributions only when its output uncertainty is high. We validate the feasibility of this opportunistic transmission by discovering a strong correlation between SLM's uncertainty and LLM's rejection probability. Furthermore, we theoretically derive optimal uncertainty thresholds and optimal vocabulary truncation strategies. Simulation results show that, compared to standard HLM, CU-HLM achieves up to 206$\times$ higher token throughput by skipping 74.8% transmissions with 97.4% vocabulary compression, while maintaining 97.4% accuracy.

cs.DC↗

arXiv2026

Semantics-Aware Hierarchical Token Communication: Clustering, Bit Mapping, and Power Allocation

Despite the rise of token communication (TokCom) as a new paradigm beyond traditional bit communication, existing approaches have primarily adopted artificial intelligence (AI)-centric designs that rely on semantic recovery via large models. Meanwhile, their physical-layer designs, such as token-bit mapping and power allocation, remain conventional and do not reflect token-level semantics. These semantics-agnostic designs can lead to significant semantic loss, particularly at low signal-to-noise ratio (SNR) levels. To address this issue, we propose hierarchical TokCom (H-TokCom), a framework that embeds semantic structure directly into physical-layer design. The key idea is to group semantically similar tokens into clusters and hierarchically assign their bit representations, where each token is represented by a cluster-level prefix and a token-specific suffix. As long as the cluster bits are correctly delivered, errors in the suffix bits typically map the received token to another within the same semantic cluster, resulting in only limited semantic distortion. This robustness is further strengthened by allocating more transmit power to the prefix bits than to the suffix bits. Simulation results show that H-TokCom achieves substantial semantic-similarity gains over conventional TokCom across the considered SNR range, increasing the semantic similarity from $0.206$ to $0.279$ at $γ=3$ dB on COCO, corresponding to a gain of $0.073$ $(35.4\%)$.

eess.SP↗

arXiv2025

Action Deviation-Aware Inference for Low-Latency Wireless Robots

To support latency-sensitive AI applications ranging from autonomous driving to industrial robot manipulation, 6G envisions distributed ML with computational resources in mobile, edge, and cloud connected over hyper-reliable low-latency communication (HRLLC). In this setting, speculative decoding can facilitate collaborative inference of models distributively deployed: a lightweight on-device model locally generates drafts while a more capable remote target model on a server verifies and corrects them in parallel with speculative sampling, thus resulting in lower latency without compromising accuracy. However, unlike autoregressive text generation, behavior cloning policies, typically used for embodied AI applications, cannot parallelize verification and correction for multiple drafts as each generated action depends on observation updated by a previous action. To this end, we propose Action Deviation-Aware Hybrid Inference (ADAHI), wherein drafts are selectively transmitted and verified based on action deviation, which has a strong correlation with action's rejection probability by the target model. By invoking server operation only when necessary, communication and computational overhead can be reduced while accuracy gain from speculative sampling is preserved. Experiments on our testbed show that ADAHI reduces transmission and server operations by approximately 40%, lowers end-to-end latency by 39.2%, and attains up to 97.2% of the task-success rate of baseline that invokes speculative sampling for every draft embedding vector.

cs.RO↗

arXiv2025

Energy-Efficient Wireless LLM Inference via Uncertainty and Importance-Aware Speculative Decoding

To address the growing demand for on-device LLM inference in resource-constrained environments, hybrid language models (HLM) have emerged, combining lightweight local models with powerful cloud-based LLMs. Recent studies on HLM have primarily focused on improving accuracy and latency, while often overlooking communication and energy efficiency. We propose a token-level filtering mechanism for an energy-efficient importance- and uncertainty-aware HLM inference that leverages both epistemic uncertainty and attention-based importance. Our method opportunistically uploads only informative tokens, reducing LLM usage and communication costs. Experiments with TinyLlama-1.1B and LLaMA-2-7B demonstrate that our method achieves up to 87.5% BERT Score and token throughput of 0.37 tokens/sec while saving the energy consumption by 40.7% compared to standard HLM. Furthermore, compared to our previous U-HLM baseline, our method improves BERTScore from 85.8% to 87.0%, energy savings from 31.6% to 43.6%, and throughput from 0.36 to 0.40. This approach enables an energy-efficient and accurate deployment of LLMs in bandwidth-constrained edge environments.

cs.LG↗

arXiv2025

Deadline-Aware Bandwidth Allocation for Semantic Generative Communication with Diffusion Models

The importance of Radio Access Network (RAN) in support Artificial Intelligence (AI) application services has grown significantly, underscoring the need for an integrated approach that considers not only network efficiency but also AI performance. In this paper we focus on a semantic generative communication (SGC) framework for image inpainting application. Specifically, the transmitter sends semantic information, i.e., semantic masks and textual descriptions, while the receiver utilizes a conditional diffusion model on a base image, using them as conditioning data to produce the intended image. In this framework, we propose a bandwidth allocation scheme designed to maximize bandwidth efficiency while ensuring generation performance. This approach is based on our finding of a Semantic Deadline--the minimum time that conditioning data is required to be injected to meet a given performance threshold--within the multi-modal SGC framework. Given this observation, the proposed scheme allocates limited bandwidth so that each semantic information can be transmitted within the corresponding semantic deadline. Experimental results corroborate that the proposed bandwidth allocation scheme achieves higher generation performance in terms of PSNR for a given bandwidth compared to traditional schemes that do not account for semantic deadlines.

eess.SY↗

arXiv2025

Hybrid Semantic-Complementary Transmission for High-Fidelity Image Reconstruction

Recent advances in semantic communication (SC) have introduced neural network (NN)-based transceivers that convey semantic representation (SR) of signals such as images. However, these NNs are trained over diverse image distributions and thus often fail to reconstruct fine-grained image-specific details. To overcome this limited reconstruction fidelity, we propose an extended SC framework, hybrid semantic communication (HSC), which supplements SR with complementary representation (CR) capturing residual image-specific information. The CR is constructed at the transmitter, and is combined with the actual SC outcome at the receiver to yield a high-fidelity recomposed image. While the transmission load of SR is fixed due to its NN-based structure, the load of CR can be flexibly adjusted to achieve a desirable fidelity. This controllability directly influences the final reconstruction error, for which we derive a closed-form expression and the corresponding optimal CR. Simulation results demonstrate that HSC substantially reduces MSE compared to the baseline SC without CR transmission across various channels and NN architectures.

eess.SP↗

arXiv2025

Uncertainty-Aware Hybrid Inference with On-Device Small and Remote Large Language Models

This paper studies a hybrid language model (HLM) architecture that integrates a small language model (SLM) operating on a mobile device with a large language model (LLM) hosted at the base station (BS) of a wireless network. The HLM token generation process follows the speculative inference principle: the SLM's vocabulary distribution is uploaded to the LLM, which either accepts or rejects it, with rejected tokens being resampled by the LLM. While this approach ensures alignment between the vocabulary distributions of the SLM and LLM, it suffers from low token throughput due to uplink transmission and the computation costs of running both language models. To address this, we propose a novel HLM structure coined Uncertainty-aware opportunistic HLM (U-HLM), wherein the SLM locally measures its output uncertainty and skips both uplink transmissions and LLM operations for tokens that are likely to be accepted. This opportunistic skipping is enabled by our empirical finding of a linear correlation between the SLM's uncertainty and the LLM's rejection probability. We analytically derive the uncertainty threshold and evaluate its expected risk of rejection. Simulations show that U-HLM reduces uplink transmissions and LLM computations by 45.93%, while achieving up to 97.54% of the LLM's inference accuracy and 2.54$\times$ faster token throughput than HLM without skipping.

cs.LG↗

arXiv2024

Privacy-Preserving Split Learning with Vision Transformers using Patch-Wise Random and Noisy CutMix

In computer vision, the vision transformer (ViT) has increasingly superseded the convolutional neural network (CNN) for improved accuracy and robustness. However, ViT's large model sizes and high sample complexity make it difficult to train on resource-constrained edge devices. Split learning (SL) emerges as a viable solution, leveraging server-side resources to train ViTs while utilizing private data from distributed devices. However, SL requires additional information exchange for weight updates between the device and the server, which can be exposed to various attacks on private training data. To mitigate the risk of data breaches in classification tasks, inspired from the CutMix regularization, we propose a novel privacy-preserving SL framework that injects Gaussian noise into smashed data and mixes randomly chosen patches of smashed data across clients, coined DP-CutMixSL. Our analysis demonstrates that DP-CutMixSL is a differentially private (DP) mechanism that strengthens privacy protection against membership inference attacks during forward propagation. Through simulations, we show that DP-CutMixSL improves privacy protection against membership inference attacks, reconstruction attacks, and label inference attacks, while also improving accuracy compared to DP-SL and DP-MixSL.

cs.DC↗

arXiv2024

Knowledge Distillation from Language-Oriented to Emergent Communication for Multi-Agent Remote Control

In this work, we compare emergent communication (EC) built upon multi-agent deep reinforcement learning (MADRL) and language-oriented semantic communication (LSC) empowered by a pre-trained large language model (LLM) using human language. In a multi-agent remote navigation task, with multimodal input data comprising location and channel maps, it is shown that EC incurs high training cost and struggles when using multimodal data, whereas LSC yields high inference computing cost due to the LLM's large size. To address their respective bottlenecks, we propose a novel framework of language-guided EC (LEC) by guiding the EC training using LSC via knowledge distillation (KD). Simulations corroborate that LEC achieves faster travel time while avoiding areas with poor channel conditions, as well as speeding up the MADRL training convergence by up to 61.8% compared to EC.

cs.AI↗

arXiv2024

Energy-Efficient Edge Learning via Joint Data Deepening-and-Prefetching

The vision of pervasive artificial intelligence (AI) services can be realized by training an AI model on time using real-time data collected by internet of things (IoT) devices. To this end, IoT devices require offloading their data to an edge server in proximity. However, transmitting high-dimensional and voluminous data from energy-constrained IoT devices poses a significant challenge. To address this limitation, we propose a novel offloading architecture, called joint data deepening-and-prefetching (JD2P), which is feature-by-feature offloading comprising two key techniques. The first one is data deepening, where each data sample's features are sequentially offloaded in the order of importance determined by the data embedding technique such as principle component analysis (PCA). Offloading is terminated once the already transmitted features are sufficient for accurate data classification, resulting in a reduction in the amount of transmitted data. The criteria to offload data are derived for binary and multi-class classifiers, which are designed based on support vector machine (SVM) and deep neural network (DNN), respectively. The second one is data prefetching, where some features potentially required in the future are offloaded in advance, thus achieving high efficiency via precise prediction and parameter optimization. We evaluate the effectiveness of JD2P through experiments using the MNIST dataset, and the results demonstrate its significant reduction in expected energy consumption compared to several benchmarks without degrading learning accuracy.

cs.LG↗

arXiv2024

Generative AI Meets Semantic Communication: Evolution and Revolution of Communication Tasks

While deep generative models are showing exciting abilities in computer vision and natural language processing, their adoption in communication frameworks is still far underestimated. These methods are demonstrated to evolve solutions to classic communication problems such as denoising, restoration, or compression. Nevertheless, generative models can unveil their real potential in semantic communication frameworks, in which the receiver is not asked to recover the sequence of bits used to encode the transmitted (semantic) message, but only to regenerate content that is semantically consistent with the transmitted message. Disclosing generative models capabilities in semantic communication paves the way for a paradigm shift with respect to conventional communication systems, which has great potential to reduce the amount of data traffic and offers a revolutionary versatility to novel tasks and applications that were not even conceivable a few years ago. In this paper, we present a unified perspective of deep generative models in semantic communication and we unveil their revolutionary role in future communication frameworks, enabling emerging applications and tasks. Finally, we analyze the challenges and opportunities to face to develop generative models specifically tailored for communication systems.

cs.CL↗

arXiv2023

Mobility-Induced Graph Learning for WiFi Positioning

A smartphone-based user mobility tracking could be effective in finding his/her location, while the unpredictable error therein due to low specification of built-in inertial measurement units (IMUs) rejects its standalone usage but demands the integration to another positioning technique like WiFi positioning. This paper aims to propose a novel integration technique using a graph neural network called Mobility-INduced Graph LEarning (MINGLE), which is designed based on two types of graphs made by capturing different user mobility features. Specifically, considering sequential measurement points (MPs) as nodes, a user's regular mobility pattern allows us to connect neighbor MPs as edges, called time-driven mobility graph (TMG). Second, a user's relatively straight transition at a constant pace when moving from one position to another can be captured by connecting the nodes on each path, called a direction-driven mobility graph (DMG). Then, we can design graph convolution network (GCN)-based cross-graph learning, where two different GCN models for TMG and DMG are jointly trained by feeding different input features created by WiFi RTTs yet sharing their weights. Besides, the loss function includes a mobility regularization term such that the differences between adjacent location estimates should be less variant due to the user's stable moving pace. Noting that the regularization term does not require ground-truth location, MINGLE can be designed under semi- and self-supervised learning frameworks. The proposed MINGLE's effectiveness is extensively verified through field experiments, showing a better positioning accuracy than benchmarks, say root mean square errors (RMSEs) being 1.398 (m) and 1.073 (m) for self- and semi-supervised learning cases, respectively.

cs.LG↗

arXiv2023

Communication-Efficient On-Device Machine Learning: Federated Distillation and Augmentation under Non-IID Private Data

On-device machine learning (ML) enables the training process to exploit a massive amount of user-generated private data samples. To enjoy this benefit, inter-device communication overhead should be minimized. With this end, we propose federated distillation (FD), a distributed model training algorithm whose communication payload size is much smaller than a benchmark scheme, federated learning (FL), particularly when the model size is large. Moreover, user-generated data samples are likely to become non-IID across devices, which commonly degrades the performance compared to the case with an IID dataset. To cope with this, we propose federated augmentation (FAug), where each device collectively trains a generative model, and thereby augments its local data towards yielding an IID dataset. Empirical studies demonstrate that FD with FAug yields around 26x less communication overhead while achieving 95-98% test accuracy compared to FL.

cs.LG↗

arXiv2023

Towards Semantic Communication Protocols for 6G: From Protocol Learning to Language-Oriented Approaches

The forthcoming 6G systems are expected to address a wide range of non-stationary tasks. This poses challenges to traditional medium access control (MAC) protocols that are static and predefined. In response, data-driven MAC protocols have recently emerged, offering ability to tailor their signaling messages for specific tasks. This article presents a novel categorization of these data-driven MAC protocols into three levels: Level 1 MAC. task-oriented neural protocols constructed using multi-agent deep reinforcement learning (MADRL); Level 2 MAC. neural network-oriented symbolic protocols developed by converting Level 1 MAC outputs into explicit symbols; and Level 3 MAC. language-oriented semantic protocols harnessing large language models (LLMs) and generative models. With this categorization, we aim to explore the opportunities and challenges of each level by delving into their foundational techniques. Drawing from information theory and associated principles as well as selected case studies, this study provides insights into the trajectory of data-driven MAC protocols and sheds light on future research directions.

cs.IT↗

arXiv2023

Semantics Alignment via Split Learning for Resilient Multi-User Semantic Communication

Recent studies on semantic communication commonly rely on neural network (NN) based transceivers such as deep joint source and channel coding (DeepJSCC). Unlike traditional transceivers, these neural transceivers are trainable using actual source data and channels, enabling them to extract and communicate semantics. On the flip side, each neural transceiver is inherently biased towards specific source data and channels, making different transceivers difficult to understand intended semantics, particularly upon their initial encounter. To align semantics over multiple neural transceivers, we propose a distributed learning based solution, which leverages split learning (SL) and partial NN fine-tuning techniques. In this method, referred to as SL with layer freezing (SLF), each encoder downloads a misaligned decoder, and locally fine-tunes a fraction of these encoder-decoder NN layers. By adjusting this fraction, SLF controls computing and communication costs. Simulation results confirm the effectiveness of SLF in aligning semantics under different source data and channel dissimilarities, in terms of classification accuracy, reconstruction errors, and recovery time for comprehending intended semantics from misalignment.

cs.LG↗

arXiv2023

Language-Oriented Communication with Semantic Coding and Knowledge Distillation for Text-to-Image Generation

By integrating recent advances in large language models (LLMs) and generative models into the emerging semantic communication (SC) paradigm, in this article we put forward to a novel framework of language-oriented semantic communication (LSC). In LSC, machines communicate using human language messages that can be interpreted and manipulated via natural language processing (NLP) techniques for SC efficiency. To demonstrate LSC's potential, we introduce three innovative algorithms: 1) semantic source coding (SSC) which compresses a text prompt into its key head words capturing the prompt's syntactic essence while maintaining their appearance order to keep the prompt's context; 2) semantic channel coding (SCC) that improves robustness against errors by substituting head words with their lenghthier synonyms; and 3) semantic knowledge distillation (SKD) that produces listener-customized prompts via in-context learning the listener's language style. In a communication task for progressive text-to-image generation, the proposed methods achieve higher perceptual similarities with fewer transmissions while enhancing robustness in noisy communication channels.

eess.SP↗