Xi Chen 0009

dblp:16/3283-9 · DBLP profile ↗
← Back
38ranked-venue papers
12as first author
23since 2021 · last 2026
—ORCID · conflict

Domains — the database's venue-derived domains; a paper can count in several

Computer networks · 20 · 6 first-author · 15 since 2021Artificial intelligence and machine learning · 9 · 7 since 2021Graphics, computer vision, multimedia, augmented reality and games · 4 · 1 first-author · 3 since 2021Systems, architecture and hardware · 3 · 2 first-authorSecurity and privacy · 1 · 1 first-authorSoftware engineering, systems software and programming languages · 1Databases, data management, data science and information retrieval · 1 · 1 first-authorHuman-computer interaction and ubiquitous computing · 1Applied, interdisciplinary, general and emerging computing · 1 · 1 first-author
YearPublicationVenuePosition
2026 MaestroBot: Generalized Gesture-Driven Hierarchical Coordination for Robotic Formations
abstract
Robotic swarm coordination holds transformative potential for applications such as warehouse automation, search & rescue, and entertainment. However, approaches relying on wearable devices or vision-based systems are often constrained by hardware-intensive, high computational requirements, reliance on line-of-sight, and privacy concerns. Wireless sensing, particularly using Channel State Information (CSI), offers a promising alternative by translating environmental perturbations into CSI variation data. Nevertheless, existing CSI-based systems face significant challenges in domain adaptation, resource limitation, and scalability issues. This paper introduces MaestroBot, a hierarchical motion coordination system that combines distributed CSI-based wireless sensing with domain-adaptive learning to address these limitations. For leader robots, the system features a lightweight hand gesture recognition model, built on a “Hybrid-Single” knowledge distillation framework, achieving up to 95.87% accuracy while maintaining adaptability across diverse domains. For follower robots, the hierarchical motion propagation model leverages localized CSI analysis and dual-layer error correction mechanisms to deliver 97.2% accuracy with a low latency of 0.085 seconds, even in multi-row formations. Additionally, its cost-effective hardware design ensures practical scalability and real-world deployability. These results position MaestroBot as an efficient, robust, and privacy-preserving solution for large-scale robotic swarm coordination in dynamic environments.
Zhiye Wang, Yuhan Xu, Haiming Jin, Linghe Kong, Rui Li 0098, Xi Chen 0009, Qiao Xiang, Guihai Chen
IEEE Trans. Mob. Comput.8
2025 NoT: Federated Unlearning via Weight Negation
abstract
Federated unlearning (FU) aims to remove a participant’s data contributions from a trained federated learning (FL) model, ensuring privacy and regulatory compliance. Traditional FU methods often depend on auxiliary storage on either the client or server side or require direct access to the data targeted for removal—a dependency that may not be feasible if the data is no longer available. To overcome these limitations, we propose NoT, a novel and efficient FU algorithm based on weight negation (multiplying by -1), which circumvents the need for additional storage and access to the target data. We argue that effective and efficient unlearning can be achieved by perturbing model parameters away from the set of optimal parameters, yet being well-positioned for quick re-optimization. This technique, though seemingly contradictory, is theoretically grounded: we prove that the weight negation perturbation effectively disrupts inter-layer co-adaptation, inducing unlearning while preserving an approximate optimality property, thereby enabling rapid recovery. Experimental results across three datasets and three model architectures demonstrate that NoT significantly outperforms existing baselines in unlearning efficacy as well as in communication and computational efficiency.
Yasser H. Khalil, Leo Maxime Brunswic, Soufiane Lamghari, Xu Li 0001, Mahdi Beitollahi, Xi Chen 0009
CVPR6
2025 FedOne: Query-Efficient Federated Learning for Black-box Discrete Prompt Learning
abstract
Black-Box Discrete Prompt Learning (BDPL) is a prompt-tuning method that optimizes discrete prompts without accessing model parameters or gradients, making the prompt tuning on a cloud-based Large Language Model (LLM) feasible. Adapting Federated Learning (FL) to BDPL could further enhance prompt tuning performance by leveraging data from diverse sources. However, all previous research on federated black-box prompt tuning had neglected the substantial query cost associated with the cloud-based LLM service. To address this gap, we conducted a theoretical analysis of query efficiency within the context of federated black-box prompt tuning. Our findings revealed that degrading FedAvg to activate only one client per round, a strategy we called \textit{FedOne}, enabled optimal query efficiency in federated black-box prompt learning. Building on this insight, we proposed the FedOne framework, a federated black-box discrete prompt learning method designed to maximize query efficiency when interacting with cloud-based LLMs. We conducted numerical experiments on various aspects of our framework, demonstrating a significant improvement in query efficiency, which aligns with our theoretical results.
Ganyu Wang, Jinjie Fang, Maxwell J. Yin, Bin Gu 0001, Xi Chen 0009, Boyu Wang 0004, Yi Chang 0001, Charles Ling 0001
ICML5
2025 Connector-S: A Survey of Connectors in Multi-modal Large Language Models
abstract
With the rapid advancements in multi-modal large language models (MLLMs), connectors play a pivotal role in bridging diverse modalities and enhancing model performance. However, the design and evolution of connectors have not been comprehensively analyzed, leaving gaps in understanding how these components function and hindering the development of more powerful connectors. In this survey, we systematically review the current progress of connectors in MLLMs and present a structured taxonomy that categorizes connectors into atomic operations (mapping, compression, mixture of experts) and holistic designs (multi-layer, multi-encoder, multi-modal scenarios), highlighting their technical contributions and advancements. Furthermore, we discuss several promising research frontiers and challenges, including high-resolution input, dynamic compression, guide information selection, combination strategy, and interpretability. This survey is intended to serve as a foundational reference and a clear roadmap for researchers, providing valuable insights into the design and optimization of next-generation connectors to enhance the performance and adaptability of MLLMs.
Xi Chen 0009, Yiming Shi, Miao Li 0003, Ji Wu 0002
IJCAI3
2024 Less or More From Teacher: Exploiting Trilateral Geometry For Knowledge Distillation
abstract
Knowledge distillation aims to train a compact student network using soft supervision from a larger teacher network and hard supervision from ground truths. However, determining an optimal knowledge fusion ratio that balances these supervisory signals remains challenging. Prior methods generally resort to a constant or heuristic-based fusion ratio, which often falls short of a proper balance. In this study, we introduce a novel adaptive method for learning a sample-wise knowledge fusion ratio, exploiting both the correctness of teacher and student, as well as how well the student mimics the teacher on each sample. Our method naturally leads to the \textit{intra-sample} trilateral geometric relations among the student prediction ($\mathcal{S}$), teacher prediction ($\mathcal{T}$), and ground truth ($\mathcal{G}$). To counterbalance the impact of outliers, we further extend to the \textit{inter-sample} relations, incorporating the teacher's global average prediction ($\mathcal{\bar{T}})$ for samples within the same class. A simple neural network then learns the implicit mapping from the intra- and inter-sample relations to an adaptive, sample-wise knowledge fusion ratio in a bilevel-optimization manner. Our approach provides a simple, practical, and adaptable solution for knowledge distillation that can be employed across various architectures and model sizes. Extensive experiments demonstrate consistent improvements over other loss re-weighting methods on image classification, attack detection, and click-through rate prediction.
Chengming Hu, Haolun Wu, Chen Ma 0001, Xi Chen 0009, Boyu Wang 0004, Jun Yan 0007, Xue (Steve) Liu
ICLR5
2024 Latent Trajectory Learning for Limited Timestamps under Distribution Shift over Time
abstract
Distribution shifts over time are common in real-world machine-learning applications. This scenario is formulated as Evolving Domain Generalization (EDG), where models aim to generalize well to unseen target domains in a time-varying system by learning and leveraging the underlying evolving pattern of the distribution shifts across domains. However, existing methods encounter challenges due to the limited number of timestamps (every domain corresponds to a timestamp) in EDG datasets, leading to difficulties in capturing evolving dynamics and risking overfitting to the sparse timestamps, which hampers their generalization and adaptability to new tasks. To address this limitation, we propose a novel approach SDE-EDG that collects the Infinitely Fined-Grid Evolving Trajectory (IFGET) of the data distribution with continuous-interpolated samples to bridge temporal gaps (intervals between two successive timestamps). Furthermore, by leveraging the inherent capacity of Stochastic Differential Equations (SDEs) to capture continuous trajectories, we propose their use to align SDE-modeled trajectories with IFGET across domains, thus enabling the capture of evolving distribution trends. We evaluate our approach on several benchmark datasets and demonstrate that it can achieve superior performance compared to existing state-of-the-art methods.
Qiuhao Zeng, Changjian Shui, Long-Kai Huang, Xi Chen 0009, Charles Ling 0001, Boyu Wang 0004
ICLR5
2024 FedSwarm: An Adaptive Federated Learning Framework for Scalable AIoT
abstract
Federated learning (FL) is a key solution for datadriven the Artificial Intelligence of Things (AIoT). Although much progress has been made, scalability remains a core challenge for real-world FL deployments. Existing solutions either suffer from accuracy loss or do not fully address the connectivity dynamicity of FL systems. In this article, we tackle the scalability issue with a novel, adaptive FL framework called FedSwarm, which improves system scalability for AIoT by deploying multiple collaborative edge servers. FedSwarm has two novel features: 1) adaptiveness on the number of local updates and 2) dynamicity of the synchronization between edge devices and edge servers. We formulate FedSwarm as a local update adaptation and perdevice dynamic server selection problem and prove FedSwarm‘s convergence bound. We further design a control mechanism consisting of a learning-based algorithm for collaboratively providing local update adaptation on the servers’ side and a bonus-based strategy for spurring dynamic per-device server selection on the devices’ side. Our extensive evaluation shows that FedSwarm significantly outperforms other studies with better scalability, lower energy consumption, and higher model accuracy.
Haizhou Du, Chengdong Ni, Chaoqian Cheng, Qiao Xiang, Xi Chen 0009, Xue (Steve) Liu
IEEE Internet Things J.5
2023 Hyperspherical Quantization: Toward Smaller and More Accurate Models
abstract
Model quantization enables the deployment of deep neural networks under resource-constrained devices. Vector quantization aims at reducing the model size by indexing model weights with full-precision embeddings, i.e., codewords, while the index needs to be restored to 32-bit during computation. Binary and other low-precision quantization methods can reduce the model size up to 32×, however, at the cost of a considerable accuracy drop. In this paper, we propose an efficient framework for ternary quantization to produce smaller and more accurate compressed models. By integrating hyperspherical learning, pruning and reinitialization, our proposed Hyperspherical Quantization (HQ) method reduces the cosine distance between the full-precision and ternary weights, thus reducing the bias of the straight-through gradient estimator during ternary quantization. Compared with existing work at similar compression levels (~30×, ~40×), our method significantly improves the test accuracy and reduces the model size.
Xi Chen 0009, Chen Ma 0001, Xue (Steve) Liu
WACV2
2023 Dynamic Consolidation for Continual Learning
abstract
Training deep learning models from a stream of nonstationary data is a critical problem to be solved to achieve general artificial intelligence. As a promising solution, the continual learning (CL) technique aims to build intelligent systems that have the plasticity to learn from new information without forgetting the previously obtained knowledge. Unfortunately, existing CL methods face two nontrivial limitations. First, when updating a model with new data, existing CL methods usually constrain the model parameters within the vicinity of the parameters optimized for old data, limiting the exploration ability of the model; second, the important strength of each parameter (used to consolidate the previously learned knowledge) is fixed and thus is suboptimal for the dynamic parameter updates. To address these limitations, we first relax the vicinity constraints with a global definition of the important strength, which allows us to explore the full parameter space. Specifically, we define the important strength as the sensitivity of the global loss function to the model parameters. Moreover, we propose adjusting the important strength adaptively to align it with the dynamic parameter updates. Through extensive experiments on popular data sets, we demonstrate that our proposed method outperforms the strong baselines by up to 24% in terms of average accuracy.
Chen Ma 0001, Xi Chen 0009, Xue (Steve) Liu
Neural Comput.3
2023 Learning From FM Communications: Toward Accurate, Efficient, All-Terrain Vehicle Localization
abstract
Vehicle localization service is a fundamental component of intelligent transportation systems. The widely used satellite navigation systems perform poorly in urban areas because the lines of sight to satellites are blocked by complex terrain characteristics, e.g., buildings, elevated streets and interchanges. In this paper, we design RadioLoc, a novel system achieving accurate, efficient, all-terrain vehicle localization with two key design points. First, RadioLoc harvests the frequency modulation (FM) signal, which has higher availability than satellite signal in complex terrains, as the signal source for localization. Second, RadioLoc integrates modern machine learning techniques into the processing of FM signals to efficiently learn the accurate vehicle localization in all-terrain environments. We validate the feasibility of FM-based vehicle localization and corresponding challenges and practical issues via field tests (e.g., signal distortion, signal inconsistency and limited in- vehicle radio bandwidth), and develop a series of advanced techniques in RadioLoc to address them, including adaptive batching, frequency sweeping, a novel multipath delay spread filter, a reconstructive PCA denoiser and a tailored FM feature extractor. We then develop a generic, modular localization module in RadioLoc, and design different learning-based 3D position identification algorithms for this module. We implement a prototype of RadioLoc and perform extensive field experiments to evaluate its efficiency and efficacy. Results show that (1) RadioLoc achieves a real-time localization latency of less than 100 milliseconds; (2) RadioLoc achieves a worst-case localization accuracy of 99.6% even in an underground parking lot, and (3) the horizontal error of RadioLoc is only one sixth of a dedicated GPS device even when the vehicle is moving at a high-speed (i.e., 80 km/h) in a complex highway scenario.
Xi Chen 0009, Qiao Xiang, Linghe Kong, Huisan Xu, Xue (Steve) Liu
IEEE/ACM Trans. Netw.1
2023 Eliminating Space Scanning: Fast mmWave Beam Alignment with UWB Radios
abstract
Due to their large bandwidth and impressive data speed, millimeter-wave (mmWave) radios are expected to play a key role in the 5G and beyond (e.g., 6G) communication networks. Yet, to release mmWave’s true power, the highly directional mmWave beams need to be aligned perfectly. Most existing beam alignment methods adopt an exhaustive or semi-exhaustive space scanning, which introduces up to seconds of delays. To eliminate the need for complex space scanning, this article presents an Ultra-wideband (UWB)-assisted mmWave communication framework, which leverages the co-located UWB antennas to estimate the best angles for mmWave beam alignment. One major challenge of applying this idea in the real world is the barrier of limited antenna numbers. Commercial-Off-The-Shelf (COTS) devices are usually equipped with only a small number of UWB antennas, which are not enough for the existing algorithms to provide an accurate angle estimation. To solve this challenge, we design a novel Multi-Frequency MUltiple SIgnal Classification (MF-MUSIC) algorithm, which extends the classic MUltiple SIgnal Classification (MUSIC) algorithm to the frequency domain and overcomes the antenna limitation barrier in the spatial domain. Extensive real-world experiments and numerical simulations illustrate the advantage of the proposed MF-MUSIC algorithm. MF-MUSIC uses only three antennas to achieve an accurate angle estimation, which is a mere 0.15° (or a relative difference of 3.6%) different from the state-of-the-art 16-antenna-based angle estimation method.
Ju Wang 0003, Xi Chen 0009, Xue (Steve) Liu, Gregory Dudek
ACM Trans. Sens. Networks2
2022 A Provably Secure ECC-based Multi-factor 5G-AKA Authentication Protocol
abstract
Due to the constant penetration of various security attacks, it is highly important to secure the underlying communication networks between the IoT, Fog and Cloud in the next generation of mobile communication system (5G). Thus, secure authentication and key agreement protocol, namely 5G-AKA, has been proposed in the literature to safely and stably access the 5G mobile services. However, some recent findings reveal that 5G-AKA and its numerous versions based on symmetric or asymmetric encryption are either vulnerable to different attacks such as perfect forward secrecy violation, malicious Serving Network (SN), de-synchronization attack, privacy theft, stolen device, or are computationally intensive. Apart from that, these protocols use single-factor authentication. Considering the above demerits of these protocols and the necessity to provide enhanced security, we propose an Elliptic Curve-Cryptography (ECC)-based multi-factor 5G-AKA authentication protocol. It provides additional security and achieves cost-effectiveness in terms of computational, communication, storage costs and energy consumption. The formal security analysis using Real-Or-Random (ROR) logic has been done to confirm its security. Moreover, we evaluate the performance of the proposed protocol in terms of computational, communication, storage costs and energy consumption. The evaluation results show that the proposed protocol requires less cost than its counterparts, reducing computational cost by up to 57%, communication cost by up to 59%, storage cost by up to 52%, and energy consumption by up to 51%.
Awaneesh Kumar Yadav, Manoj Misra, Pradumn Kumar Pandey, Kuljeet Kaur, Sahil Garg, Xi Chen 0009
GLOBECOM6
2022 Resource Management for Heterogeneous Aerial Networks with Backhaul Constraints
abstract
In this paper, we study the coverage maximization problem in the aerial networks. Specifically, we propose a heterogeneous aerial network (HetAN) consisting of a high-altitude base station (HBS) acting as a hub to provide wireless backhaul and multiple low-altitude BSs (LBSs) acting as access points to provide on-demand wireless coverage. Besides, we adopt the non-orthogonal multiple access (NOMA) technique for the uplink transmissions of the terrestrial users so as to support massive connections. Then, we formulate a joint power control, channel assignment, and rate control problem with the objective to maximize user connectivity and network throughput. Based on the graph methods and theoretical analysis, we propose an efficient iterative algorithm to solve the formulated problem. Simulation results demonstrate that our algorithm outperforms the other schemes in terms of connectivity and throughput.
Daosen Zhai, Qiqi Shi, Haotong Cao, Sahil Garg, Xi Chen 0009, Rongxing Lu
GLOBECOM5
2022 Data-Efficient Communication Traffic Prediction With Deep Transfer Learning
abstract
Prediction of future traffic load is a crucial task to support the automatic Operations, Administration, and Management (OAM) of communication networks. Existing Machine Learning (ML) models require big data to accomplish this task. However, large data sets are not always available, due to the limited storage capacity and the high storage cost at Base Stations (BSs). To solve the problem, we leverage the spatial-temporal correlation among different BSs, which allows other BSs’ data to be used for the prediction of the target BS. One major challenge in realizing this idea is the imbalance of data amounts between neighbor BSs and a prediction target BS. If one simply aggregates the data from both neighbors and target, the target’s traffic features would be overwhelmed by the neighbors’ data. To address this challenge, we propose a Spatial-Temporal Transfer (STT) framework, which trains a base model with an aggregated data set from multiple BSs, and then carefully refines the base model to serve a target BS. To strike a perfect balance between general tendency and individual features, STT adopts an advanced transfer learning technique that exploits regularization on model parameters. Experiments show the efficiency of the proposed STT framework.
Ju Wang 0003, Xi Chen 0009, Xue Liu 0004, Gregory Dudek
ICC3
2022 Communication Traffic Prediction with Continual Knowledge Distillation
abstract
Accurate traffic volume estimation and prediction are essential for advanced communication network functions, such as automatic operations and predictive resource allocation. Although machine learning (ML)-based approaches achieve great success in accomplishing this goal, existing approaches suffer from two drawbacks that limit their real-world applications. First, the ML-based prediction models developed in the past might be obsolete now, since the communication traffic patterns and volumes keep changing in the real world, leading to prediction errors. Second, most Base Stations (BSs) can only save a small amount of data due to the limited storage capacity and high storage costs, which prevents from training an accurate prediction model. In this paper, we propose a novel framework that adapts the prediction model to the constantly changing traffic with only a few current traffic data. Specifically, the framework first learns the knowledge of historical traffic data as much as possible by using a proposed two-branch neural network design, which includes a prediction and a reconstruction module. Then, the framework transfers the knowledge from an old (past) prediction model to a new (current) model for the model update by using a proposed continual knowledge distillation technique. Evaluations on a real-world dataset show that the proposed framework reduces the Mean Absolute Error (MAE) of traffic prediction by up to 9.62% compared to the state-of-the-art prediction methods.
Ju Wang 0003, Chengming Hu, Xi Chen 0009, Xue Liu 0004, Seowoo Jang, Gregory Dudek
ICC4
2022 Fidora: Robust WiFi-Based Indoor Localization via Unsupervised Domain Adaptation
abstract
Emerging Internet of Things (IoT) applications, such as cashier-less shopping, mobile ads targeting, and geo-based augmented reality (AR), are expected to bring us much more convenience and infotainment. To realize this amazing future, we need to feed these applications with user locations of (sub)meter-level resolution anytime and anywhere. Unfortunately, many widely used location sources are either unavailable indoor (e.g., global positioning system) or coarse grained (e.g., user check-ins). In order to provide ubiquitous localization services, the widespread WiFi signals are being leveraged to establish (sub)meter-level localization systems. Fine-grained WiFi propagation characteristics, which are sensitive to human body locations, have been employed to create location fingerprints. However, these WiFi characteristics are also sensitive to: 1) the body shapes of different users and 2) the objects in the background environment. Consequently, systems based on WiFi fingerprints are vulnerable in the presence of: 1) new users with different body shapes and 2) daily changes of the environment, e.g., opening/closing doors. To tackle this issue, this article proposes a WiFi-based localization system based on domain-adaptation with cluster assumption, named Fidora. Fidora is able to: 1) localize different users with labeled data from only one or two example users and 2) localize the same user in a changed environment without labeling any new data. To achieve these, Fidora integrates two major modules. It first adopts a data augmenter that introduces data diversity using a variational autoencoder (VAE). It then trains a domain-adaptive classifier that adjusts itself to newly collected unlabeled data using a joint classification-reconstruction structure. We conducted real-world experiments to evaluate Fidora against the state of the art. It is demonstrated that when tested on an unlabeled user, Fidora increases the average$F1$score by 17.8% and improves the worst case accuracy by 20.2%. Moreover, when applied in a varied environment, Fidora outperforms the state of the art by 23.1%.
Xi Chen 0009, Chenyi Zhou, Xue Liu 0004, Di Wu 0044, Gregory Dudek
IEEE Internet Things J.1
2021 One for All: Traffic Prediction at Heterogeneous 5G Edge with Data-Efficient Transfer Learning
abstract
By placing the computing, storage and networking resources close to the end users, distributed edge computing greatly benefits the performance of 5G communication systems. However, as a tradeoff, resources on the edge are usually limited and imbalanced among the heterogeneous edge nodes. To overcome this drawback, this paper proposes a Transfer Learning based Prediction (TLP) framework that allows the edge nodes to share their resources and data in an efficient manner. In particular, the TLP framework focuses on the prediction of the future traffic load, which is a key reference for many automated network functions. To enhance the efficiency of data and bandwidth, TLP first learns a base model on a data-abundant edge node (the source), and then transfers this model (instead of data) to other data-limited nodes (the targets). To achieve a delicate balance between maintaining common features and learning target-specific features, we develop a new transfer learning technique named Similarity-based Elastic Weight Con-solidation (SEWC), and integrate it into TLP. Experiments on real-world data illustrate that, compared to the state-of-the-art methods, TLP-SEWC reduces the Mean Absolute Error (MAE) of traffic prediction by up to 57.9%.
Xi Chen 0009, Ju Wang 0003, Yi Tian Xu, Di Wu 0044, Xue Liu 0004, Gregory Dudek, Taeseop Lee, Intaik Park
GLOBECOM1
2021 AFB: Improving Communication Load Forecasting Accuracy with Adaptive Feature Boosting
abstract
Prediction of key system characteristics, such as the communication load, is required to overcome the delays in wireless communication systems. State-of-The-Art (SOTA) approaches mostly apply existing Neural Network (NN) structures, and extract latent features purely based on their sensitivity to the forecasting accuracy. This way of feature extraction may neglect some non-obvious yet informative dimensions in the model input, leading to inaccurate forecasting results. In this paper, we present an Adaptive Feature Boosting (AFB) approach, which integrates multiple AutoEncoders (AEs) to automatically extract robust and comprehensive latent features for communication load forecasting. The recurrent and residual connections among the AEs make sure that the extracted latent features are representative for all input dimensions. With more comprehensive information extracted from the history, the forecasting accuracy is thus improved. We evaluate AFB against existing approaches on a real-world dataset that contains Call Detail Records (CDRs) of the Milan city over a period of two months. The evaluation shows that our AFB-based approach achieves 35.2% more accurate load forecasting results than the SOTA deep approaches.
Chengming Hu, Xi Chen 0009, Ju Wang 0003, Jikun Kang, Yi Tian Xu, Xue Liu 0004, Di Wu 0044, Seowoo Jang, Intaik Park, Gregory Dudek
GLOBECOM2
2021 Learning Assisted Identification of Scenarios Where Network Optimization Algorithms Under-Perform
abstract
We present a generative adversarial method that uses deep learning to identify network load traffic conditions in which network optimization algorithms under-perform other known algorithms: the Deep Convolutional Failure Generator (DCFG). The spatial distribution of network load presents challenges for network operators for tasks such as load balancing, in which a network optimizer attempts to maintain high quality communication while at the same time abiding capacity constraints. Testing a network optimizer for all possible load distributions is challenging if not impossible. We propose a novel method that searches for load situations where a target network optimization method underperforms baseline, which are key test cases that can be used for future refinement and performance optimization. By modeling a realistic network simulator's quality assessments with a deep network and, in parallel, optimizing a load generation network, our method efficiently searches the high dimensional space of load patterns and reliably finds cases in which a target network optimization method under-performs a baseline by a significant margin.
Dmitriy Rivkin, David Meger, Di Wu 0044, Xi Chen 0009, Xue Liu 0004, Gregory Dudek
GLOBECOM4
2021 Load Balancing for Communication Networks via Data-Efficient Deep Reinforcement Learning
abstract
Within a cellular network, load balancing between different cells is of critical importance to network performance and quality of service. Most existing load balancing algorithms are manually designed and tuned rule-based methods where near-optimality is almost impossible to achieve. These rule-based meth-ods are difficult to adapt quickly to traffic changes in real-world environments. Given the success of Reinforcement Learning (RL) algorithms in many application domains, there have been a number of efforts to tackle load balancing for communication systems using RL-based methods. To our knowledge, none of these efforts have addressed the need for data efficiency within the RL framework, which is one of the main obstacles in applying RL to wireless network load balancing. In this paper, we formulate the communication load balancing problem as a Markov Decision Process and propose a data-efficient transfer deep reinforcement learning algorithm to address it. Experimental results show that the proposed method can significantly improve the system performance over other baselines and is more robust to environmental changes.
Di Wu 0044, Jikun Kang, Yi Tian Xu, Jimmy Li 0001, Xi Chen 0009, Dmitriy Rivkin, Michael R. M. Jenkin, Taeseop Lee, Intaik Park, Xue Liu 0004, Gregory Dudek
GLOBECOM6
2021 Hierarchical Policy Learning for Hybrid Communication Load Balancing
abstract
Due to the uneven demographic distribution and people’s daily activities, communication systems usually experience highly imbalanced load across different cells. This imbalance leads to unsatisfied users in the congested cells and under-utilized resources in the less-loaded cells. To deal with this issue, existing work migrates the load from heavily loaded cells to lightly loaded cells, by either handing over active mode User Equipment (UEs) to other serving cells, or re-selecting the camping cells for idle mode UEs. In this paper, we further advance the research on Load Balancing (LB) with a hybrid control of both active and idle UEs. This task is challenging, due to the conflicts between Active-UE LB (AULB) and Idle-UE LB (IULB) policies. To overcome this challenge, we propose a Hierarchical Policy Learning (HPL) framework, which coordinates the actions between LB policies with a two-level learning structure. In this way, HPL produces AULB and IULB policies that are better aligned with each other. Extensive simulation results illustrate the efficiency and efficacy of the proposed HPL.
Jikun Kang, Xi Chen 0009, Di Wu 0044, Yi Tian Xu, Xue Liu 0004, Gregory Dudek, Taeseop Lee, Intaik Park
ICC2
2021 UWB-Assisted Fast mmWave Beam Alignment
abstract
Due to their large bandwidth and impressive data speed, millimeter-wave (mmWave) radios are expected to play a key role in the 5G and beyond (e.g., 6G) communication networks. Yet, to release mmWave’s true power, the highly directional mmWave beams need to be aligned perfectly. Most existing beam alignment methods adopt an exhaustive or semi-exhaustive space scanning, which introduces up to seconds of delays.To eliminate the need of a complex space scanning, this paper presents an Ultra-wideband (UWB)-assisted mmWave communication framework, which leverages the co-located UWB antennas to estimate the best angles for mmWave beam alignment. One major challenge to apply this idea in real-world is the barrier of limited antenna numbers. Commercial-Off-The-Shelf (COTS) devices are usually equipped with only a few number of UWB antennas, which are not enough for the existing algorithms to provide an accurate angle estimation. To solve this challenge, we design a novel Multi-Frequency MUSIC (MF-MUSIC) algorithm, which extends the classic MUSIC algorithm to the frequency domain and overcomes the antenna limitation barrier in the spatial domain. By doing this, our framework uses only 3 antennas to achieve an accurate angle estimation, which is merely 0.15° different from the state-of-the-art 16-antenna method.
Ju Wang 0003, Xi Chen 0009, Xue Liu 0004, Gregory Dudek
ICC2
2021 Generalized DataWeighting via Class-Level Gradient Manipulation
abstract
Label noise and class imbalance are two major issues coexisting in real-world datasets. To alleviate the two issues, state-of-the-art methods reweight each instance by leveraging a small amount of clean and unbiased data. Yet, these methods overlook class-level information within each instance, which can be further utilized to improve performance. To this end, in this paper, we propose Generalized Data Weighting (GDW) to simultaneously mitigate label noise and class imbalance by manipulating gradients at the class level. To be specific, GDW unrolls the loss gradient to class-level gradients by the chain rule and reweights the flow of each gradient separately. In this way, GDW achieves remarkable performance improvement on both issues. Aside from the performance gain, GDW efficiently obtains class-level weights without introducing any extra computational cost compared with instance weighting methods. Specifically, GDW performs a gradient descent step on class-level weights, which only relies on intermediate gradients. Extensive experiments in various settings verify the effectiveness of GDW. For example, GDW outperforms state-of-the-art methods by $2.56\%$ under the $60\%$ uniform noise setting in CIFAR10. Our code is available at https://github.com/GGchen1997/GDW-NIPS2021.
Can Chen 0005, Shuhao Zheng, Xi Chen 0009, Erqun Dong, Xue (Steve) Liu, Hao Liu 0026, Dejing Dou
NeurIPS3
2020 PresSense: Passive Respiration Sensing via Ambient WiFi Signals in Noisy Environments
abstract
Passive sensing with ambient WiFi signals is a promising technique that will enable new types of human-robot interactions while preserving users' privacy. Here, we present PresSense, a system for human respiration sensing in noisy environments. Unlike existing WiFi-based respiration sensors, we employ a human presence detector, improving the robustness in scenarios where no human is present in an Area Of Interest (AOI). We also integrate our novel feature, Peak Distance Histogram (PDH), with other classic WiFi features to achieve better accuracy when someone is present in the AOI. We tested our system using commodity WiFi devices in an office room. Our PresSense outperforms the state of the arts in both respiration rate estimation and presence detection.
Yi Tian Xu, Xi Chen 0009, Xue Liu 0004, David Meger, Gregory Dudek
IROS2
2020 FiDo: Ubiquitous Fine-Grained WiFi-based Localization for Unlabelled Users via Domain Adaptation
abstract
To fully support the emerging location-aware applications, location information with meter-level resolution (or even higher) is required anytime and anywhere. Unfortunately, most of the current location sources (e.g., GPS and check-in data) either are unavailable indoor or provide only house-level resolutions. To fill the gap, this paper utilizes the ubiquitous WiFi signals to establish a (sub)meter-level localization system, which employs WiFi propagation characteristics as location fingerprints. However, an unsolved issue of these WiFi fingerprints lies in their inconsistency across different users. In other words, WiFi fingerprints collected from one user may not be used to localize another user. To address this issue, we propose a WiFi-based Domain-adaptive system FiDo, which is able to localize many different users with labelled data from only one or two example users. FiDo contains two modules: 1) a data augmenter that introduces data diversity using a Variational Autoencoder (VAE); and 2) a domain-adaptive classifier that adjusts itself to newly collected unlabelled data using a joint classification-reconstruction structure. Compared to the state of the art, FiDo increases average F1 score by 11.8% and improves the worst-case accuracy by 20.2%.
Xi Chen 0009, Chenyi Zhou, Xue (Steve) Liu, Di Wu 0044, Gregory Dudek
WWW1
2019 RadioLoc: Learning Vehicle Locations with FM Signal in All-Terrain Environments
abstract
Vehicle localization service is a fundamental component of intelligent transportation systems. The widely used satellite navigation systems perform poorly in urban areas because the lines of sight to satellites are blocked by complex terrain characteristics, e.g., buildings, elevated streets and interchanges. In this paper, we design RadioLoc, a novel system achieving accurate, efficient, all-terrain vehicle localization with two key design points. First, RadioLoc harvests the frequency modulation (FM) signal, which has a higher availability than satellite signal in complex terrains, as the signal source for localization. Second, RadioLoc integrates modern machine learning techniques into the processing of FM signals to efficiently learn the accurate vehicle localization in all-terrain environments. We validate the feasibility of FM-based vehicle localization and corresponding challenges and practical issues via field tests (e.g., signal distortion, signal inconsistency and limited in-vehicle radio bandwidth), and develop a series of advanced techniques in RadioLoc to address them, including a new multipath delay spread filter, a reconstructive PCA denoiser, a tailored FM feature extractor, an adaptive batching technique and a frequency sweep technique. We implement a prototype of RadioLoc and perform extensive field experiments to evaluate its efficiency and efficacy. Results show that (1) RadioLoc achieves a real-time localization latency of less than 100 milliseconds; (2) RadioLoc achieves a worst-case localization accuracy of 99.6% even in an underground parking lot, and (3) the horizontal error of RadioLoc is only one sixth of a dedicated GPS device even when the vehicle is moving at a high-speed (i.e., 80 km/h) in a complex highway scenario.
Xi Chen 0009, Qiao Xiang, Linghe Kong, Xue (Steve) Liu
MASS1
2017 Taming the inconsistency of Wi-Fi fingerprints for device-free passive indoor localization
abstract
Device-free Passive (DfP) indoor localization releases the users from the burden of wearing sensors or carrying smartphones. Instead of locating devices, DfP technology directly locates human bodies. This promising technology upgrades and even redefines many services, such as intruder alarm, fire rescue, fall detection, baby monitoring, etc. Using Wi-Fi based fingerprints, DfP approaches can achieve a nearly perfect accuracy with a resolution less than one meter. However, Wi-Fi localization profiles may easily drift with a minor environment change, resulting in an inconsistency between fingerprints and new profiles. This inconsistency issue could lead to large errors, and may quickly ruin the whole system. To address this issue, we propose a approach named AutoFi to automatically calibrate the localization profiles in an unsupervised manner. AutoFi embraces a new technique that online estimates and cancels profile contaminants introduced by environment changes. It applies an autoencoder to preserve critical features of fingerprints, and reproduces them later in new localization profiles. Experiment results demonstrate that AutoFi indeed rescues the Wi-Fi fingerprints from variations in the surrounding. The localization accuracy is improved from 18.8% (before auto-calibration) to 84.9% (after auto-calibration).
Xi Chen 0009, Chen Ma 0001, Michel Allegue, Xue (Steve) Liu
INFOCOM1
2017 Joint adaptation framework in mobile ad hoc networks: A control theory perspective
Linghe Kong, Xi Chen 0009, Xue (Steve) Liu, Xiao-Yang Liu, Jiadi Yu, Guangtao Xue, Guihai Chen
Neurocomputing3
2016 How cars talk louder, clearer and fairer: Optimizing the communication performance of connected vehicles via online synchronous control
abstract
The connected vehicles have been considered as a remedy for modern traffic issues, potentially saving hundreds of thousands of lives every year worldwide. The Dedicated Short-Range Communications (DSRC) technology is an essential building block of this promising vision. DSRC faces volatile vehicular environments, where not only wireless propagation channels but also network topologies vary rapidly. Moreover, traffic congestions during rush hours may lead to an unprecedentedly high density of broadcasting radios, resulting in compromised reliability, efficiency and fairness of DSRC. In order to optimize the performance of DSRC, we develop a novel Online Control Approach of power and Rates (OnCAR). Supported by systematic control theories, OnCAR performs stably even in the dynamic and unpredictable vehicular environments. To the best of our knowledge, OnCAR is the first solution to address the strong coupling between communication variables. It adopts a multi-variable control model to synchronously adjust transmission power and data rates, which are two major variables determining the performance of DSRC. In addition, OnCAR leverages receiver-side measurements of performance metrics to strike a balance between overall performance and fairness. Compared with the state of the art, OnCAR enhances the overall reliability and efficiency of DSRC by 23.7% and 30.1%, respectively. Meanwhile, these numbers are achieved with a 40.1% improvement in fairness.
Xi Chen 0009, Linghe Kong, Xue (Steve) Liu, Lei Rao, Fan Bai 0002, Qiao Xiang
INFOCOM1
2016 DRIVING: Distributed Scheduling for Video Streaming in Vehicular Wi-Fi Systems
abstract
Video streaming has been dominating the mobile bandwidth, and is still expanding drastically. Its tremendous economic benefits have driven the automobile industry to equip vehicles with video streaming capacity. As a result, the new in-cabin Wi-Fi systems have been deployed, enabling each vehicle as a streaming hotspot on the wheels. A built-in Access Point (AP) bridges the communications between Wi-Fi devices inside and cellular networks outside. Distinct advantages offered by this system include a more powerful antenna array to improve multimedia quality, a constant energy source to power the streaming, etc. However, there exist two challenging features that may jeopardize the system performance. (1) The in-cabin Wi-Fi hotspots are mostly deployed on private vehicles, and thus are completely decentralized. (2) Video packets need to be delivered before their deadlines with small delays. Due to these features, existing algorithms may fail to efficiently schedule the in-cabin Wi-Fi video streaming. To fill the gap, we propose the Delay-awaRe dIstributed Video schedulING (DRIVING) framework. Being fully distributed and delay-aware, DRIVING not only increases the streaming goodput, but also reduces the delivery latency and deadline missing ratio. %In order to optimize this new framework, we establish cross-layer analytical models, which help us tune the framework parameters for better performance. In a typical scenario, DRIVING increases the goodput by up to 27.0%, while reducing the queueing delay and the deadline missing ratio by up to 40.0% and 38.4%, respectively.
Xi Chen 0009, Lei Rao, Qiao Xiang, Xue (Steve) Liu, Fan Bai 0002
ACM Multimedia1
2016 Software Reliability Analysis Using Weakest Preconditions in Linear Assignment Programs
abstract
Weakest preconditions derived from triple axiomatic semantics have been widely used to prove the correctness of programs. They can also be applied to evaluate the reliability of software. However, deducing a weakest precondition, as well as determining its propagation path, encounters challenges such as unknown constraint conditions, symbol computation and means of representation. To address these challenges, in this paper, we utilize the disjunctive normal form of if-else branch structure to capture reasonable propagation paths of the weakest precondition. Meanwhile, by removing the sequential dependencies, we demonstrate how to get the weakest precondition of loop-structure by leveraging program function. Moreover, we extensively explore three modeling characteristics (i.e., path extension, innermost connection and condition leap) for deducing the weakest precondition of structured programs. Finally, taking the definition of program node and storage structure of weakest precondition as bases, we design a serial of modeling algorithms. Based on symbol computation and recursive call technology with Depth-First Search (DFS), our algorithms can not only be used to deduce the weakest precondition, but also to capture the propagate path of the weakest precondition. Experiments illustrate the efficacy and effectiveness of our proposed models and designed deductive algorithms.
Xue (Steve) Liu, Xi Chen 0009, Ting Long, Ronghua Jiang
IEEE Trans. Software Eng.3
2015 Solving the performance puzzle of DSRC multi-channel operations
abstract
Dedicated Short Range Communication (DSRC) protocol is a key enabling technology for enhancing road safety and transportation efficiency. Wireless Access in Vehicular Environments (WAVE) 1609.4 is a new amendment that enables multi-channel operations in DSRC. Operating intervals are divided into alternating Control Channel (CCH) Intervals and Service Channel (SCH) Intervals with an identical length. This alternating feature causes high packet losses in CCH and low throughput in SCH, and thus hinders the deployment of this protocol. The goal of our work is to provision sufficient reliability for safety messages in CCH while optimising non-safety service delivery in SCH. We develop analytical models to explore the relationship among traffic density, CCH packet loss ratio, SCH throughput, and the duration of each kind of intervals. We also design a multi-channel coordination algorithm which adaptively adjusts the duration of intervals to achieve better performance and reliability based on these models. Theoretical analysis and extensive simulation results demonstrate the accuracy of our model and the efficacy of the proposed algorithm.
Xi Chen 0009, Lei Rao, Xue (Steve) Liu, Yuan Yao 0004
ICC2
2015 Data preference matters: A new perspective of safety data dissemination in vehicular ad hoc networks
abstract
Vehicle-to-vehicle safety data dissemination plays an increasingly important role in ensuring the safety and efficiency of vehicle transportation. When collecting safety data, vehicles always prefer data generated at a closer location over data generated at a distant location, and prefer recent data over outdated data. However, these data preferences have been overlooked in most of existing safety data dissemination protocols, preventing vehicles getting more precise traffic information. In this paper, we explore the feasibility and benefits of incorporating the data preferences of vehicles in designing efficient safety data dissemination protocols. In particular, we propose the concept of packet-value to quantify these data preferences. We then design PVCast, a packet-value-based safety data dissemination protocol in VANET. PVCast makes the dissemination decision for each packet based on its packet-value and effective dissemination coverage in order to satisfy the data preferences of all the vehicles in the network. In addition, PVCast is lightweight and fully distributed. We evaluate the performance of PVCast on the ns-2 platform by comparing it with three representative data dissemination protocols. Simulation results in a typical highway scenario show that PVCast provides a significant improvement on per-vehicle throughput, per-packet dissemination coverage with small per-packet delay. Our findings demonstrate the importance and necessity of comprehensively considering the data preferences of vehicles when designing an efficient safety data dissemination protocol for VANET.
Qiao Xiang, Xi Chen 0009, Linghe Kong, Lei Rao, Xue (Steve) Liu
INFOCOM2
2015 FINE: Frequency-divided instantaneous neighbors estimation system in vehicular networks
abstract
In this paper, we present a novel Frequency-divided Instantaneous Neighbors Estimation (FINE) system specifically designed for density estimation in Dedicated Short Range Communication (DSRC) based vehicular networks. A large amount of vehicular applications such as navigation, traffic control, and data dissemination substantially rely on the density information. Recent works pay great attention to obtain the real-time density information and reduce the occupation time of DSRC channel. The state-of-the-art approach is the Framed Slotted ALOHA (FSA) framework, which benefits from its fine-grained time division design. However, FSA considers only time resource and is unaware of the frequency resource in DSRC. For further accelerating the density acquisition, we propose a frequency-divided approach. The core idea of FINE is to resort fine-grained channel division for parallel neighbors counting. Extensive simulations are conducted to evaluate FINE. The results demonstrate that FINE significantly outperforms existing methods. In a typical dense scenario, FINE reduces the time cost from 2 ms (FSA) to 50 μs, while maintains the accuracy at the same level as FSA.
Linghe Kong, Xi Chen 0009, Xue (Steve) Liu, Lei Rao
PerCom2
2014 DTS: Dynamic TDMA scheduling for Networked Control Systems
Xi Chen 0009, Akramul Azim, Xue (Steve) Liu, Sebastian Fischmeister
J. Syst. Archit.1
2013 SyRaFa: Synchronous Rate and Frequency Adjustment for Utilization Control in Distributed Real-Time Embedded Systems
abstract
To efficiently utilize the computing resources and provide good quality of service (QoS) to the end-to-end tasks in the distributed real-time systems, we can enforce the utilization bounds on multiple processors. The utilization control is challenging especially when the workload in the system is unpredictable. To handle the workload uncertainties, current research favors feedback control techniques, and recent work combines the task rate adaptation and processor frequency scaling in an asynchronous way for CPU utilization control, where task rates and the processor frequencies are tuned asynchronously in two decoupled control loops for control convenience. Since the two manipulated variables, task rates and processor frequencies, contribute to the CPU utilizations together with strong coupling, adjusting them asynchronously may degrade the utilization control performance. In this paper, we provide a novel scheme to make synchronous rate and frequency adjustment to enforce the utilization setpoint, referred to as SyRaFa scheme. SyRaFa can handle the workload uncertainties by identifying the system model online and can simultaneously adjust the manipulated variables by solving an optimization problem in each sampling period. Extensive evaluation results demonstrate SyRaFa outperforms the existing schemes especially under severe workload uncertainties.
Xi Chen 0009, Xiao-Wen Chang, Xue (Steve) Liu
IEEE Trans. Parallel Distributed Syst.1
2012 CSS: Conditional State-Based Scheduling for Networked Control Systems
abstract
Modern industrial networked control systems(NCSs) tend to be complicated and have dynamic workload by holding a variety of applications via a shared network. The static network scheduling algorithms fit most NCSs due to their deterministic characteristics and timing guarantees, but they cannot handle dynamic workloads for lack of making on the-fly decisions. The conditional state-based scheduling adds the dynamism in the static scheduling algorithms by automata or more explicitly state chart like formalisms with conditional transitions. In this paper, we propose CSS scheme that applies the conditional state-based scheduling to dynamically schedule different applications in the industrial NCSs. CSS aims at the time-triggered network in the NCSs and uses time division multiple access (TDMA) method to let the applications access the network. To enhance the scalability of the NCSs, we design CSS as a decentralized scheme where each application in NCSs has a local scheduler to make its schedule decisions. Appropriate algorithms are applied to ensure the scheduling decisions made by the local schedulers are consistent and the desired system performance can be achieved. Simulation results demonstrate the effectiveness of the proposed scheme compared to the static TDMA used in real-time networks.
Xi Chen 0009, Akramul Azim, Xue (Steve) Liu, Sebastian Fischmeister
RTCSA1
2012 TailCon: Power-Minimizing Tail Percentile Control of Response Time in Server Clusters
abstract
To provide satisfactory customer experience, modern server clusters like Amazon usually set Service Level Agreement (SLA) as guaranteeing a certain percentile (i.e. 99%) of the customer requests to have a response time within a threshold (i.e. 1s). One way to meet the SLA constraint is to serve the customer requests with sufficient computing capacity based on the worst case workload estimation in the server cluster. However, this may cause unnecessary power consumption in the server cluster due to over-provision of the computing capacity especially when the workload is highly dynamic. In this paper, we propose an adaptive computing capacity allocation scheme referred to as TailCon. TailCon aims at minimizing the power consumption in the server cluster while satisfying the SLA constraint by adjusting the number of active servers and the CPU frequencies of the turn on machines online. In TailCon, we analyze the distribution of the request response time dynamically and leverage the measured request response time to estimate the workload intensity in the server cluster, which is used as a continuous feedback to find the proper provision of the computing capacity online based on optimization techniques. We conduct both the emulation using the real-word HTTP traces and the experiments to evaluate the performance of TailCon. The experimental results demonstrate the effectiveness of TailCon scheme in enforcing the SLA constraint while saving the power consumption.
Xi Chen 0009, Xue (Steve) Liu, Shengquan Wang, Xiao-Wen Chang
SRDS1