Abstract Recent progress in deep learning for automated multi-modal medical diagnosis heavily depend on extensive expert annotations, which is time-intensive and impractical. To mitigate this, masked image modeling (MIM)-based pre-training strategies have emerged, effectively learning generalized representations from unlabelled data for various downstream tasks. Nevertheless, these approaches are tailored for natural images while neglect the distinct characteristics of medical data, resulting in suboptimal generalization in medical diagnosis applications. In this work, we attempt to harness the complementary information of multi-modal medical data to perform self-supervised pre-training and propose MRM++, an enhanced masked relation modeling paradigm. Different from the previous MIM methods that randomly mask input data, causing potentially missing of disease-relevant semantics, we devise prior-guided relation masking to break token-wise feature relation guided by anatomy-aware prior in both self- and cross-modal aspects. This can preserve complete input semantics and enable the model to learn abundant disease-related knowledge. Furthermore, to boost semantic relation modeling, the relation matching is introduced, which aligns sample-wise relations among unmasked and masked features. By exploiting inter-sample relations, the relation matching imposes the global constraints in the feature space, ensuring ample semantic relation for robust feature representation. Additionally, considering that the model may overfit to the pre-training dataset and lead to inherent gap between pre-training and downstream fine-tuning, we conceive task-oriented adapting as a pre-stage before fine-tuning to simultaneously perform self-supervised and task-supervised learning on downstream dataset. It can adaptively transform knowledge from the pre-trained model to be compatible with downstream tasks while maintaining transferable information. Extensive experiments on medical image-text and image-genome benchmarks validate the effectiveness and transfer ability of the proposed framework, outperforming state-of-the-art methods across various downstream diagnostic tasks. Source codes are made publicly available on https://github.com/CUHK-AIM-Group/MRM_plus .

The Industrial Internet of Things (IIoT) has been widely deployed to enable real-time monitoring and automation. Within IIoT-driven production, supply chain management plays a critical role, necessitating verifiable provenance to ensure the authenticity and traceability of goods across multi-stakeholder networks. While blockchain provides a tamper-proof foundation, traditional storage structures suffer from unsecured data integrity, poor query efficiency, and scalability over provenance data. To address these challenges, we propose vProChain, an efficient provenance verification system to support verifiable and parallel queries over graph-structured provenance data. First, we design an Adaptive DAG Verkle Tree (ADVT) that deterministically maps supply chain dependencies into a graph-native authenticated data structure, enabling constant-size proofs and low-overhead verification. Second, we introduce the Merkle Inverted Patricia Trie (MIPT) to facilitate fast, verifiable multi-dimensional Boolean queries. Third, we develop a parallel provenance query algorithm that accelerates multi-hop path retrieval via consistent hashing and weighted bipartite matching. Finally, formal security analysis and extensive empirical evaluations demonstrate that vProChain can provide provable cryptographic guarantees for the soundness of provenance proofs and the completeness of query retrievals, while achieving high query efficiency in a large-scale IIoT environment.

Millimeter-wave (mmWave) 5G-V2X communications face significant challenges in real-time beam alignment within high-mobility vehicular networks. While environmentaware beam prediction methods mitigate channel estimation overhead, their efficacy is severely compromised by modality data loss stemming from lighting variations, adverse weather, or sensor failures. To address this issue, we propose a Cross-modal Losstolerant Beam Prediction model (CLBP). CLBP robustly fuses RGB camera and LiDAR data, employing a novel cross-modal attention mechanism to achieve resilient feature alignment across these heterogeneous modalities. Furthermore, a Branch Features Dynamic Fusion (BFDF) module adaptively reweights modality features, suppressing noise from degraded inputs and promoting effective information propagation to enhance resilience. To facilitate realistic evaluation, we introduce a Data-Conditioned Missingness Mechanism (DCMM), which augments the DeepSense 6G V2V dataset with meticulously simulated sensor failure scenarios. Experimental results demonstrate CLBP's superior performance, achieving 94.48% Top-5 beam prediction accuracy even under 10% modality loss, and a 29% reduction in average power loss compared to baseline methods. These findings demonstrate CLBP's significant robustness in dynamic vehicular environments and its capacity to maintain consistent, high-performance beam prediction despite challenging data imperfections.

Current research on autonomous driving trajectory prediction algorithms continues to face challenges. One key issue is that the encoding of road topology may fail to simultaneously capture both local detail and global representation. Secondly, the separate processing of spatial and temporal information for traffic participants can result in the loss of critical temporal feature information. To address these issues, we propose a trajectory prediction algorithm named TrajPred-MSM using the high-definition map by multi-scale. The core idea of this algorithm is to encode high-definition maps at two scales: nodes for local representation and lane segments for global representation. The designed dual-scale approach captures fine-grained microscopic connection relationships in local areas while preserving macroscopic connections from a global perspective. Experimental results on the Argoverse dataset demonstrate that TrajPred-MSM, by effectively extracting map information, outperforms traditional trajectory prediction algorithms.

Nowadays, checkpoints have been proven to be an effective solution to ensure slashability in proof-of-stake (PoS) consensus, and Tas et al.'s cutting-edge solution in S&P 2023 is a typical example. Unfortunately, despite progress, hour-level transaction delays and annually around 10 K dollar checkpoint costs make existing related solutions still unacceptable in realworld PoS applications. In this paper, we propose SecPoS, a slashable PoS consensus with second-level transaction delays and one-time checkpoint costs. To achieve these design goals, we draw inspiration from Pixel+ signatures and chameleon hash functions to design a novel bilateral blockchain structure, achieving twoblock transaction finalization via only uploading the first block of our chain as checkpoints. Next, considering practical application requirements, we address a series of following challenges, such as bilateral immutability, blockchain forks, determination of the main chain, and malicious attacks from PoS members. In detail, we propose two constructions of SecPoS, i.e., SecPoS – A and SecPoS – B. SecPoS – A and SecPoS – B have a tradeoff between transaction delays and block numbers packed in an epoch. Compatible with most existing one-way blockchains, we implement and outsource a prototype SecPoS to facilitate research11https://github.com/Academic-Paper-Codes/SecPoS-Consensus, and prove the security of SecPoS. Experiments on this prototype show that SecPoS – A and SecPoS – B require around 5s and 100s transaction delays, respectively, and both require 2 dollars one-time checkpoint costs.

Abstract Federated semi-supervised learning (FSSL) target to address the increasing privacy concerns for the practical scenarios, where data holders are limited in labeling capability. Latest FSSL approaches leverage the prediction consistency between the local model and global model to exploit knowledge from partially labeled or completely unlabeled clients. However, they merely utilize data-level augmentation for prediction consistency and simply aggregate model parameters through the weighted average at the server, which leads to biased classifiers and suffers from skewed unlabeled clients. To remedy these issues, we present a novel FSSL framework, Relation-guided Versatile Regularization (FedRVR), consisting of versatile regularization at clients and relation-guided directional aggregation strategy at the server. In versatile regularization, we propose the model-guided regularization together with the data-guided one, and encourage the prediction of the local model invariant to two extreme global models with different abilities, which provides richer consistency supervision for local training. Moreover, we devise a relation-guided directional aggregation at the server, in which a parametric relation predictor is introduced to yield pairwise model relation and obtain a model ranking. In this manner, the server can provide a superior global model by aggregating relative dependable client models, and further produce an inferior global model via reverse aggregation to promote the versatile regularization at clients. Extensive experiments on three FSSL benchmarks verify the superiority of FedRVR over state-of-the-art counterparts across various federated learning settings.

The manufacturing environment and activities with a large volume and variety of product data have put forward higher requirements for the proof and verification of identity information. Achieving decentralized digital identity management in the Industrial Internet of Things (IIoT) helps to improve the performance of relevant proofs and authentication. The Decentralized Identity (DID) system serves as a bridge between the physical and digital worlds, assigning digital identities to physical entities to facilitate their participation in online activities. However, faced with the huge number of manufacturing entities accessing the DID system, the number of DID documents in the system has proliferated. It is still a big challenge to improve the scalability of the system while ensuring the efficiency of information access and verification. In this paper, we propose a blockchain-based verifiable decentralized identity system for IIoT. First, we propose a blockchain-based system architecture with a specially designed storage structure for DID documents. Specifically, we design a structure based on Merkle Tree that visually summarises the physical associations of manufacturing entities and reduces access overhead. Second, we design a multiblock storage structure within the blockchain, which establishes inter-block jumps based on the associated DID, effectively improving the query efficiency of the system. Finally, we design a verification scheme that enables users to verify the integrity of the identity data of the proof provider. We implemented the system framework and conducted experiments to evaluate the performance of our system. The experimental results proved the effectiveness of the system.

Dynamic searchable symmetric encryption (DSSE) schemes empower data owners to outsource their encrypted data to clouds while retaining the ability to update or search on it. Despite a lot of efforts devoted in recent years, there are still several challenges that have not been well addressed. First, the confidentiality of data might be compromised if forward privacy and backward privacy cannot be ensured. Second, only the traditional single keyword-file search has attracted tremendous attention, while other popular queries like Boolean queries and range queries are not fully investigated. Lastly, how to solve these problems on untrusted servers that may deviate from pre-defined protocols is also challenging. In this article, aiming to tackle the above problems, we propose a novel DSSE scheme named${\sf HeX}$based on Trusted Execution Environment (TEE) that supports rich queries on untrusted servers while guaranteeing forward and backward privacy. We achieve strong forward and backward security by designing a deferred obfuscating read-write technique atop the bitmap index. We further extend the basic scheme to realize Boolean queries and range queries by reducing them to basic keyword queries. Strict theoretical analysis is conducted to prove the security of${\sf HeX}$, and extensive evaluations illustrate its efficiency and practicality.

Adversarial training (AT) is widely considered the state-of-the-art technique for improving the robustness of deep neural networks (DNNs) against adversarial examples (AEs). Nevertheless, recent studies have revealed that adversarially trained models are prone to unfairness problems. Recent works in this field usually apply class-wise regularization methods to enhance the fairness of AT. However, this paper discovers that these paradigms can be sub-optimal in improving robust fairness. Specifically, we empirically observe that the AEs that are already robust (referred to as “easy AEs” in this paper) are useless and even harmful in improving robust fairness. To this end, we propose the hard adversarial example mining (HAM) technique which concentrates on mining hard AEs while discarding the easy AEs in AT. Specifically, HAM identifies the easy AEs and hard AEs with a fast adversarial attack method. By discarding the easy AEs and reweighting the hard AEs, the robust fairness of the model can be efficiently and effectively improved. Extensive experimental results on four image classification datasets demonstrate the improvement of HAM in robust fairness and training efficiency compared to several state-of-the-art fair adversarial training methods. Our code is available athttps://github.com/yyl-github-1896/HAM.

Blockchain has become a popular paradigm for secure and immutable data storage. Despite its numerous applications across various fields, concerns regarding the user privacy and result integrity during data queries persist. Additionally, the need for rich query functionalities to harness the full potential of blockchain data remains an area ripe for exploration. In order to address these challenges, our paper first utilizes a framework based on the Trusted Execution Environment (TEE) and oblivious RAM technique to achieve both privacy and data integrity. To enhance the query efficiency over the entire blockchain, we then devise a two-level learned indexing methodology named TELEX within the TEE for both integer and string keys. We also propose different query processing algorithms for versatile query types, including exact queries, aggregate queries, Boolean queries, and range queries. By implementing the prototype and conducting extensive evaluation, we demonstrate the feasibility and remarkable improvement in efficiency compared to existing solutions.

A neural processing unit (NPU) is a microprocessor which is specially designed for various types of neural network applications. Because of its high acceleration efficiency and lower power consumption, the airborne embedded system has widely deployed NPU to replace GPU as the new accelerator. Unfortunately, the inherent scheduler of NPU does not consider real-time scheduling. Therefore, it cannot meet real-time requirements of airborne embedded systems. At present, there is less research on the multi-task real-time scheduling of the NPU device. In this article, we first design an NPU resource management framework based on Kubernetes. Then, we propose WAMSPRES, a workload-aware NPU performance model based soft preemptive real-time scheduling method. The proposed workload-aware NPU performance model can accurately predict the remaining execution time of the task when it runs with other tasks concurrently. The soft preemptive real-time scheduling algorithm can provide approximate preemption capability by dynamically adjusting the NPU computing resources of tasks. Finally, we implement a prototype NPU scheduler of the airborne embedded system for the fixed-wing UAV. The proposed models and algorithms are validated on both the simulated and realistic task sets. Experimental results illustrate that WAMSPRES can achieve low prediction error and high scheduling success rate.

Web 3.0 is viewed as the next generation of the Internet, with the aim of establishing a decentralized network where users can control their digital identities and data. Due to its decentralization feature, blockchain has become a promising solution for secure data storage and retrieval for abundant de-centralized applications in Web 3.0. In this context, decentralized identifiers (DIDs) are rapidly emerging as a key infrastructure for blockchain-based Web 3.0. However, with more and more DIDs generated and stored on the blockchain, it is challenging to support efficient retrieval of DIDs with data integrity assurance. In this paper, we propose a novel authenticated DID retrieval system for blockchain-based Web 3.0. Specifically, a new authenticated data structure (ADS) with the corresponding data verification algorithm is designed to enable efficient retrieval and verification for both DID records and their historical updates. Theoretical analysis has been performed to prove the security and efficiency of our proposed system. We implement our system and conduct experiments to evaluate the performance. Experimental results demonstrate that our proposed scheme exhibits higher system efficiency compared to the baseline solution.

Web 3.0 has been proposed as a new generation of the Internet, which shifts towards system decentralization, improved data security, and self-sovereign identity. With the proliferation of networked entities, the proper management and verification of their identities play a vital role in Web 3.0. Decentralized identity is a promising paradigm to enhance data security and restore sovereignty over personal data to users. However, the data security in existing centralized solutions is often severely limited. In this paper, we propose vDID, a novel blockchain-enabled verifiable decentralized identity management system for Web 3.0. First, we design a generic verifiable DID structure, which is capable of capturing and expressing the inherent relationships between different entities with high granularity. Second, we develop an identity verification scheme to support efficient integrity verification for identities and their relationships in the decentralized framework. We implement vDID and conduct experiments to evaluate the system performance. Experimental results demonstrate the effectiveness of our proposed system.

The learning capabilities of single-modality images are often severely limited and fail to meet the requirements of complexity defect detection in industrial settings. For instance, traditional visible light images are susceptible to environmental factors such as lighting and occlusions, while infrared images cannot capture texture details due to their low spatial resolution. Consequently, employing multiple image modalities typically yields better results than relying on a single modality. However, utilizing data from multiple modalities inevitably introduces additional computational costs, posing high hardware demands on edge computing devices, and the need for real-time detection in industrial environments is critical. To address these challenges, we propose a multimodal distillation approach that uses visible and infrared images as inputs to train a complex teacher model, while the student model continues to operate with a single-modal image input. Through knowledge transfer, the student model is enhanced, and model light-weighting is implemented to ensure that it can acquire multi-modal feature information while still meeting real-time performance requirements.

To reduce the transmission cost of blockchain confidential transactions, we propose SymmeProof, a novel communication efficient non-interactive zero-knowledge range proof protocol without a trusted setup. We design and integrate two new techniques in SymmeProof, namely vector compression and inner-product range proof. The proposed vector compression is able to reduce the communication cost to log(n) for n-size vectors. The proposed inner-product range proof converts a range proof relation into an inner-product form, which can further reduce the range proof size with the vector compression technique. Based on these two techniques, SymmeProof can eventually achieve a log(n)-size range proof. The proposed SymmeProof can be used in many important applications such as blockchain confidential transactions as well as arguments for arithmetic circuits satisfiability. We evaluate the performance of SymmeProof. The results show that SymmeProof substantially outperforms representative methods such as Bulletproofs in the proof size without a trusted setup.

Blockchain has recently gained massive attention thanks to the success of cryptocurrencies and decentralized applications. With immutability and tamper-resistance features, it can be seen as a promising secure database solution. To address the need of searches over blockchain databases, prior work vChain proposed a novel verifiable processing framework that ensures query integrity without maintaining a full copy of the blockchain database. It however suffers from several limitations, including linear-scan search performance in the worst case and impractical public key management. In this paper, we propose a new searchable blockchain system, vChain+, that supports efficient verifiable boolean range queries with additional features. Specifically, we propose a sliding window accumulator index to achieve efficient query processing even for the worst case. We also design an object registration index to enable practical public key management without compromising the security guarantee. To support richer queries, we employ optimal tree-based indexes to index both keywords and numerical attributes of the data objects. Several optimizations are also proposed to further improve the query performance. Security analysis and empirical study validate the robustness and performance improvement of the proposed system. Compared with vChain, vChain+ improves the query performance by up to 913x.

Augmented reality (AR) defines a new information-delivery paradigm by overlaying computer-generated information on the perception of the real world. AR-integrated robot has become an appealing concept in terms of enhanced human-robot interaction. Despite intensive research on AR, existing indoor location-based AR systems are vulnerable to attacks and can hardly meet the security and privacy requirements in practice. The problem of designing a secure AR framework to ensure the efficiency and privacy of location-based AR has not been sufficiently studied. In this paper, we holistically study this problem and propose EPAR, an efficient and privacy-aware AR framework for indoor location-based services. EPAR distinguishes itself from the existing work by being the first to address the issues of AR delivery in terms of system scalability, accuracy, privacy, and efficiency. First, an effective indoor location cloaking scheme is presented to safeguard user's privacy while improving system scalability and accuracy. Then, a novel privacy-aware localization scheme is proposed to hierarchically localize the user with privacy concerns. Finally, for the AR content delivery, a new authenticated data structure is tailored to save the data transmission cost and improve system efficiency. We implement EPAR and conduct extensive experiments in real-world scenarios. Evaluation results demonstrate the effectiveness of our EPAR system.

Recently, sparse arrays such as nested array and coprime array have attracted much attention in the field of array signal processing. In this letter, we develop a symmetric coprime array (SCA) whose sensor locations satisfy the nesting property, so it can be used as a dense subarray of nested array. Based on this observation, we propose a new sparse array named coprime nested array, which can achieve the same number of uniform degrees of freedom (uDOFs) as the prototype nested array, while the mutual coupling effect is at the same level as the coprime arrays. Moreover, an improved coprime nested array (ICNA) is proposed by rearranging some sensors in SCA to the right side of the sparse subarray. ICNA possesses more uDOFs than the existing nested arrays with further reduced mutual coupling effect. Numerical simulations verify the effectiveness of the proposed configurations.

Clinically oriented deep learning algorithms, combined with large-scale medical datasets, have significantly promoted computer-aided diagnosis. To address increasing ethical and privacy issues, Federated Learning (FL) adopts a distributed paradigm to collaboratively train models, rather than collecting samples from multiple institutions for centralized training. Despite intensive research on FL, two major challenges are still existing when applying FL in the real-world medical scenarios, including the performance degradation (i.e., retrogress) after each communication and the intractable class imbalance. Thus, in this paper, we propose a novel personalized FL framework to tackle these two problems. For the retrogress problem, we first devise a Progressive Fourier Aggregation (PFA) at the server side to gradually integrate parameters of client models in the frequency domain. Then, at the client side, we design a Deputy-Enhanced Transfer (DET) to smoothly transfer global knowledge to the personalized local model. For the class imbalance problem, we propose the Conjoint Prototype-Aligned (CPA) loss to facilitate the balanced optimization of the FL framework. Considering the inaccessibility of private local data to other participants in FL, the CPA loss calculates the global conjoint objective based on global imbalance, and then adjusts the client-side local training through the prototype-aligned refinement to eliminate the imbalance gap with such a balanced goal. Extensive experiments are performed on real-world dermoscopic and prostate MRI FL datasets. The experimental results demonstrate the advantages of our FL framework in real-world medical scenarios, by outperforming state-of-the-art FL methods with a large margin. The source code is available at https://github.com/CityU-AIM-Group/PRR-Imbalancehttps://github.com/CityU-AIM-Group/PRR-Imbalance.

Despite increasingly emerging applications, a primary concern for blockchain to be fully practical is the inefficiency of data query. Direct queries on the blockchain take much time by searching every block, while indirect queries on a blockchain database greatly degrade the authenticity of query results. To conquer the authenticity problem, we propose a Verifiable Query Layer (VQL) that can be deployed in the cloud to provide both efficient and verifiable data query services for blockchain systems. The middleware layer extracts data from the underlying blockchain system and efficiently reorganizes them in databases. To prevent falsified data from being stored in the middleware, a cryptographic fingerprint is calculated based on each constructed database. The database fingerprint will be first verified by miners and then written into the blockchain. Moreover, public users can verify the entire databases or several databases that interest them in the middleware layer. We implement VQL together with the verification schemes and conduct extensive experiments based on a practical blockchain system. The evaluation results demonstrate that VQL can efficiently support various data query services and guarantee the authenticity of query results for blockchain systems.

The eruption of a pandemic, such as COVID-19, can cause an unprecedented global crisis. Contact tracing, as a pillar of communicable disease control in public health for decades, has shown its effectiveness on pandemic control. Despite intensive research on contact tracing, existing schemes are vulnerable to attacks and can hardly simultaneously meet the requirements of data integrity and user privacy. The design of a privacy-preserving contact tracing framework to ensure the integrity of the tracing procedure has not been sufficiently studied and remains a challenge. In this paper, we propose P2B-Trace, a privacy-preserving contact tracing initiative based on blockchain. First, we design a decentralized architecture with blockchain to record an authenticated data structure of the user's contact records, which prevents the user from intentionally modifying his local records afterward. Second, we develop a zero-knowledge proximity verification scheme to further verify the user's proximity claim while protecting user privacy. We implement P2B-Trace and conduct experiments to evaluate the cost of privacy-preserving tracing integrity verification. The evaluation results demonstrate the effectiveness of our proposed system.

The introduction of software-defined networking (SDN) has emerged as a new network paradigm for network innovations. By decoupling the control plane from the data plane in traditional networks, SDN provides high programmability to control and manage networks. However, the communication between the two planes can be a bottleneck of the whole network. SDN-aimed DoS attacks can cause long packet delay and high packet loss rate by using massive table-miss packets to jam links between the two planes. To detect and mitigate SDN-aimed DoS attacks, this paper presents FloodDefender, an efficient and protocol-independent defense framework for SDN/OpenFlow networks. FloodDefender stands between the controller platform and other controller apps, and conforms to the OpenFlow policy without additional devices. The detection module in FloodDefender utilizes new frequency features to precisely identify SDN-aimed DoS attacks. The mitigation module uses three new techniques to efficiently mitigate attack traffic: table-miss engineering to prevent the communication bandwidth from being exhausted; packet filter to filter out attack traffic and save computational resources of the control plane; and flow rule management to eliminate most of useless flow entries in the switch flow table. Our evaluation on a prototype implementation of FloodDefender shows that the defense framework can precisely identify and efficiently mitigate the SDN-aimed DoS attacks with very little overhead.

Mining attacks allow attackers to gain an unfair share of the mining reward by deviating from the honest mining strategy in the Bitcoin system. Among the most well-known are block withholding (BWH), fork after withholding (FAW), and selfish mining. In this paper, we propose two new strategies: power adjusting and bribery racing, and introduce two novel mining attacks, Power Adjusting Withholding (PAW) and Bribery Selfish Mining (BSM) adopting the new strategies. Both attacks can increase the reward of attackers. Furthermore, we show PAW can avoid the "miner's dilemma" in BWH attacks. BSM introduces a new "venal miner's dilemma", which results in all targets (bribes) willing to help the attacker but getting less reward finally. Quantitative analyses and simulations are conducted to verify the effectiveness of our attacks. We propose some countermeasures to mitigate the new attacks, but a practical and efficient solution remains to be an open problem.

Information about urban safety, e.g., the safety index of a position, is of great importance to protect humans and support safe walking route planning. Despite some research on urban safety analysis, the accuracy and granularity of safety index inference are both very limited. The problem of analyzing urban safety to predict safety index throughout a city has not been sufficiently studied and remains open. In this paper, we propose U-Safety, an urban safety analysis system to infer safety index by leveraging multiple cross-domain urban location-based data. We first extract spatially-related and temporally-related features from various urban location-based data, including urban map, housing rent and density, population, positions of police stations, point of interests (POIs), crime event records, and taxi GPS trajectories. Then, these features are fed into a novel sparse auto-encoder (SAE) framework with feature correlation constraint to obtain the final discriminative feature representation. Finally, we design a new co-training-based learning method, which consists of two separated classifiers, to calculate safety index accurately. We implement U-Safety and conduct extensive experiments by utilizing various real data sources obtained in New York City. The evaluation results demonstrate the advantages of U-Safety over other methods.

In this paper, we present a new design, named MWS, for application offloading in mobile environments. MWS implements a mobility-aware WiFi selection policy on smartphones with the goal of achieving low delay and energy efficiency. The essence of this work lies in the emphasis of minimizing the number of occurrences of WiFi disconnects by utilizing human mobility habits and cloud-assisted WiFi information profiling. The WiFi access points (AP) selected by MWS are predicted to maintain the longest connection with smartphones among all available APs. As a result, MWS manages to avoid unnecessary or unsuccessful handoff that may otherwise be caused by the default signal strength oriented WiFi selection policy. The benefits brought by MWS are validated in our real-world evaluation. Compared with the default policy, MWS effectively reduces the number of occurrences of WiFi disconnects, and achieves a reduction of up to 50% in energy consumption and up to 66% in data communication time for application offloading in mobile environments.

With the development of sensing technology, smartphones can provide various kinds of data, including inertial sensing data, WiFi data, depth data, and images. These data make it possible to construct accurate indoor floor plans that are the critical foundations of flourishing indoor location-based services for smartphone. However, even with the popular crowdsourcing approach, the wide construction of indoor floor plans has not yet to be realized due to the intensive time consumption. In this paper, we utilize deep learning techniques to build PlanSketcher, a system that enables one user to construct fine-grained and facility-labeled indoor floor plans accurately. First, the proposed system extracts novel integrated features to recognize diverse landmarks. Second, traverse-independent hallway topologies are constructed based on the sensing data, depth data, and images through the proposed hallway construction algorithms. Finally, PlanSketcher constructs the room shape and labels recognized facilities in their corresponding positions to generate a complete indoor floor plan. Because PlanSketcher exploits different kinds of data collected from smartphones with new feature extraction method, it can obtain accurate indoor floor plan topology and facility labels. We implement PlanSketcher and conduct extensive experiments in three large indoor settings. The evaluation results show that the 90th percentile accuracy of positions and orientations of facilities are 1 m–2.5 m and 4°–6°, while 85%–95% facilities are recognized and labeled precisely.

Similarity search is a primitive operation in various database applications. Thus far, a large number of access methods have been proposed to accelerate the similarity query processing. Nonetheless, these methods mostly focus on developing standalone systems by proposing new indices. Given the fact that existing RDBMS merely support traditional indices, it is of great necessity and practical importance to develop a standard RDBMS built-in index based approach to speeding up the query processing. In this demonstration, we introduce MSQL+, a plugin toolkit that enable users to answer similarity queries in metric spaces simply using standard SQL statements. This toolkit can help existing RDBMS to effectively and efficiently handle with big data due to the following three advantages. First, MSQL+ enables users to find similar objects by submitting SELECT-FROM-WHERE statements so that it can be easily integrated into existing RDBMS. Second, MSQL+ works in a more general data space. Objects of any type can be indexed by B + -trees and the query processing can be boosted by using index seeks, as long as the similarity function is metric. Third, MSQL+ supports the parallelization of both pre-processing and query processing in distributed RDBMS.

Accurate digital maps play a crucial role in various location-based services and applications. However, store information is usually missing or outdated in current maps. In this paper, we propose CrowdGIS, an automatic store selfupdating system for digital maps that leverages street views and sensing data crowdsourced from mobile users. We first develop a new weighted artificial neural network to learn the underlying relationship between estimated positions and real positions to localize user's shooting positions. Then, a novel text detection method is designed by considering two valuable features, including the color and texture information of letters. In this way, we can recognize complete store name instead of individual letters as in the previous study. Furthermore, we transfer the shooting position to the location of recognized stores in the map. Finally, CrowdGIS considers three updating categories (replacing, adding, and deleting) to update changed stores in the map based on the kernel density estimate model. We implement CrowdGIS and conduct extensive experiments in a real outdoor region for 1 month. The evaluation results demonstrate that CrowdGIS effectively accommodates store variations and updates stores to maintain an up-to-date map with high accuracy.

Information about urban safety, e.g., the safety index of a position, is of great importance to protect humans and support safe walking route planning. Despite some research on urban safety analysis, the accuracy and granularity of safety index inference are both very limited. The problem of analyzing urban safety to predict safety index throughout a city has not been sufficiently studied and remains open. In this paper, we propose U-Safety, an urban safety analysis system to infer safety index by leveraging multiple cross-domain urban data. We first extract spatially-related and temporally-related features from various urban data, including urban map, housing rent and density, population, positions of police stations, point of interests (POIs), crime event records, and taxi GPS trajectories. Then, these features are feeded into a sparse auto-encoder (SAE) model to obtain the final discriminative feature representation. Finally, we design a new co-training-based learning method, which consists of two separated classifiers, to calculate safety index accurately. We implement U-Safety and conduct extensive experiments based on real data sources obtained in New York City. The evaluation results demonstrate the advantages of U-Safety over other methods.

Route planning is an important service in the map navigation. However, most of commercial map applications provide an optimal path that only minimize a single metric such as distance, time or other costs, while ignoring a critical criterion: safety. When citizens or travellers walk in a city, they may prefer to find a safe walking route to avoid the potential crime risk and to have a short distance, which can be formulated as a multi-objective optimization problem. Many previous methods are proposed to solve the multi-objective route planning, however, most of them are not efficient or optimized in a large-scale road network. In this paper, we propose a reinforcement learning based Multi-Objective Hyper-Heuristic (MOHH) approach to route planning in a smart city. We conduct experiments on the safety index map constructed based on the historical urban data of the New York city. Comprehensive experimental results show that the proposed approach is almost 34 and 1.4 times faster than the exact multi-objective optimization algorithm and the NSGA-II algorithm respectively. Moreover, it can obtain more than 80% Pareto optimal solutions in a large-scale road network.

The separated control and data planes in software-defined networking (SDN) with high programmability introduce a more flexible way to manage and control network traffic. However, SDN will experience long packet delay and high packet loss rate when the communication link between two planes is jammed by SDN-aimed DoS attacks with massive table-miss packets. In this paper, we propose FloodDefender, an efficient and protocol-independent defense framework for SDN/OpenFlow networks to mitigate DoS attacks. It stands between the controller platform and other controller apps, and can protect both the data and control plane resources by leveraging three new techniques: table-miss engineering to prevent the communication bandwidth from being exhausted; packet filter to identify attack traffic and save computational resources of the control plane; and flow rule management to eliminate most of useless flow entries in the switch flow table. All designs of FloodDefender conform to the OpenFlow policy, requiring no additional devices. We implement a prototype of FloodDefender and evaluate its performance in both software and hardware environments. Experimental results show that FloodDefender can efficiently mitigate the SDN-aimed DoS attacks, incurring less than 0.5% CPU computation to handle attack traffic, only 18ms packet delay and 5% packet loss rate under attacks.

Energy saving solutions on smartphone devices can greatly extend a smartphone's lasting time. However, today's push services require keep-alive connections to notify users of incoming messages, which cause costly energy consuming and drain a smartphone's battery quickly in cellular communications. Most keep-alive connections force smartphones to frequently send heartbeat packets that create additional energy-consuming radio-tails. No previous work has addressed the high-energy consumption of keep-alive connections in smartphones push services. In this paper, we propose Single Connection Proxy (SCoP) system based on fog computing to merge multiple keep-alive connections into one, and push messages in an energy-saving way. The new design of SCoP can satisfy a predefined message delay constraint and minimize the smartphone energy consumption for both real-time and delay-tolerant apps. SCoP is transparent to both smartphones and push servers, which does not need any changes on today's push service framework. Theoretical analysis shows that, given the Poisson distribution of incoming messages, SCoP can reduce the energy consumption by up to 50%. We implement SCoP system, including both the local proxy on the smartphone and remote proxy on the “Fog”. Experimental results show that the proposed system consumes 30% less energy than the current push service for real-time apps, and 60% less energy for delay-tolerant apps.

Energy efficient prefetching systems for smart-phones can greatly reduce energy consumption and data transmission, and maintain the timely response when information is prefetched. However, the proxy structure of the system can cause security problem to reveal private information to the third party. The end-to-end encryption (SSL) in traditional prefetching systems cannot solve the security problem in this new, complex energy efficient prefetching system. In this paper, we propose Secure and Energy Efficient Prefetching (SEEP) to meet the security requirement of HTTPS connections and to save smartphone's energy consumption and data transmission. The new design of SEEP includes two parts: the local proxy on the smartphone to verify the validity of prefetched responses, and the remote proxy (e.g. on the cloudlet) to store encrypted prefetched responses. SEEP is transparent to both smartphones and web servers, which does not need to change today's Browser/Server framework. Security analysis shows that SEEP protects the confidentiality of requests and responses, and is able to resist replay attack from malicious proxy. Experimental results show that the proposed system consumes 25% less energy and 95% less data when prefetching 10 outbound webpages than the traditional prefetching system in Wi-Fi networks.

Wearable cameras require connecting to cellular-capable devices (e.g., smartphones) so as to provide live broadcast services for worldwide users when Wi-Fi is unavailable. However, the constantly changing cellular network conditions may substantially slow down the upload of recorded videos. In this paper, we consider the scenario where wearable cameras upload live videos to remote distribution servers under cellular networks, aiming at maximizing the quality of uploaded videos while meeting the delay requirements. To attain the goal, we propose a dynamic video coding approach that utilizes dynamic video recording resolution adjustment on wearable cameras and Lyapunov based video preprocessing on smartphones. Our proposed resolution adjustment algorithm adapts to network condition changes, and reduces the overheads of video preprocessing. Due to the property of Lyapunov optimization framework, our proposed video preprocessing algorithm delivers near-optimal video quality while meeting the upload delay requirements. Our evaluation results show that our approach achieves up to 50% reduction in power consumption on smartphones and up to 60% reduction in average delay, at the cost of slightly compromised video quality.

The benefit of offloading applications from smart-phones to cloud servers is undermined by the significant energy consumption in data transmission. Most previous approaches attempt to improve the energy efficiency only by choosing a more energy efficient network. However, we find that for computer vision applications, pre-processing the data before offloading can also substantially lower the energy consumption in data transmission at the cost of lower result accuracy. In this paper, we propose a novel online decision making approach to determining the pre-processing level for either higher result accuracy or better energy efficiency in a mobile environment. Different from previous work that maximizes the energy efficiency, our work takes the energy consumption as a constraint. Since people usually charge their smartphones daily, it is unnecessary to extend the battery life to last more than a day. Under both the energy and time constraints, we attempt to solve the problem of maximizing the result accuracy in an online way. Our real-world evaluation shows that the implemented prototype of our approach achieves a near-optimal accuracy for application execution results (nearly 99% correct detection rate for face detection), and sufficiently satisfies the energy constraint.

The complexity of real-world industrial control systems is growing rapidly. This raises significant challenges for the use of embedded systems in control applications, since embedded platforms are generally resource limited. In this work we develop a Scilab/Scicos based embedded controller on which various control software can be easily modeled, simulated, implemented, and evaluated to meet the ever-expanding requirements of complex industrial control applications. It is built on the Cirrus Logic EP9315 ARM9 systems-on-chip board. With the developed platform, it is possible to design and implement complex embedded control systems that employ advanced control strategies in a rapid and cost-efficient fashion. Thanks to the free and open source nature of the software packages used, the cost of the embedded controller is minimized.