The co-design of neural network architectures, quantization precisions, and hardware accelerators offers a promising approach to achieving an optimal balance between performance and efficiency, particularly for model deployment on resource-constrained edge devices. In this work, we propose the JAQ Framework, which jointly optimizes the three critical dimensions. However, effectively automating the design process across the vast search space of those three dimensions poses significant challenges, especially when pursuing extremely low-bit quantization. Specifical, the primary challenges include: (1) Memory overhead in software-side: Low-precision quantization-aware training can lead to significant memory usage due to storing large intermediate features and latent weights for backpropagation, potentially causing memory exhaustion. (2) Search time-consuming in hardware-side: The discrete nature of hardware parameters and the complex interplay between compiler optimizations and individual operators make the accelerator search time-consuming. To address these issues, JAQ mitigates the memory overhead through a channel-wise sparse quantization (CSQ) scheme, selectively applying quantization to the most sensitive components of the model during optimization. Additionally, JAQ designs BatchTile, which employs a hardware generation network to encode all possible tiling modes, thereby speeding up the search for the optimal compiler mapping strategy. Extensive experiments demonstrate the effectiveness of JAQ, achieving approximately 7% higher Top-1 accuracy on ImageNet compared to previous methods and reducing the hardware search time per iteration to 0.15 seconds.

With the rapid proliferation of the Internet of Things (IoT), the security of IoT devices has become an increasingly critical concern, particularly in the context of lightweight cryptographic algorithms designed for resource-constrained environments. Lightweight cryptography is indispensable for ensuring the confidentiality and integrity of communications across IoT networks. This paper presents a thorough security evaluation of the Sycon algorithm, focusing on its susceptibility to a range of cryptographic attacks. We begin by introducing Meet-in-the-Middle (MitM) preimage and collision attacks on 3-round and 4-round Sycon-Hash. These attacks are the first published analysis results on reduced-round Sycon-Hash. By employing an SMT-based modeling approach in conjunction with the STP solver, we successfully construct and validate MitM attack paths, thereby providing novel insights into the resilience of Sycon against such threats. In addition, we propose and implement a committing attack on the 2-round Sycon-AEAD-64 using the CMT-3 framework, which uncovers potential vulnerabilities within its authenticated encryption mechanism. Our analysis not only advances the understanding of Sycon’s cryptographic robustness but also offers valuable methodologies for the future assessment of lightweight cryptographic solutions in resource-constrained contexts.

With the widespread use of mobile networks and smart devices, opportunistic mobile crowdsensing (MCS) has emerged as one of the most promising sensing paradigms for intelligent data. In Opportunistic MCS, the real-time mobility of participants and requesters is a crucial feature, as it significantly impacts the quality of MCS services. However, most existing task allocation approaches focus on optimizing the overall system performance while disregarding the mobile attribute of participants and requesters. To remedy this issue, this article proposes a real-time route prediction-based multiobjective task allocation for Opportunistic MCS, called RRP-MOTA, which presents the participants’ route-considered task allocation scheme to maximize social welfare comprehensively. Specifically, instead of merely optimizing system performance, a two-stage mechanism is designed to comprehensively enhance task allocation efficiency by estimating and leveraging participants’ routes. Moreover, by utilizing participants’ spatio–temporal location information, an improved graph convolutional network-based participant route prediction method is developed to provide more accurate participant location information for task allocation. Furthermore, a reference vector-based multiobjective task allocation method is suggested to cater to diverse usage preferences by balancing quality of service and task cost. To validate the performance of our proposed method, extensive simulations are performed on synthetic and real datasets in two scenarios. Experimental results demonstrate that the proposed RRP-MOTA significantly outperforms the chosen existing designs.

The key collision attack was proposed as an open problem in key-committing security in Authenticated Encryption (AE) schemes like AES-GCM and ChaCha20Poly1305. In ASIACRYPT 2024, Taiyama et al. introduce a novel type of key collision—target-plaintext key collision (TPKC) for AES. Depending on whether the plaintext is fixed, TPKC can be divided into fixed-TPKC and free-TPKC, which can be directly converted into collision attacks and semi-free-start collision attacks on the Davies-Meyer (DM) hashing mode. In this paper, we propose a new rebound attack framework leveraging a time-memory tradeoff strategy, enabling practical key collision attacks with optimized complexity. We also present an improved automatic method for findingrebound-friendlydifferential characteristics by controlling the probabilities in the inbound and outbound phases, allowing the identified characteristics to be directly used inrebound-basedkey collision attacks. Our analysis reveals that the 2-round AES-128 fixed-TPKC attack proposed by Taiyama et al. is, in fact, a free-TPKC attack. This distinction is significant, as fixed-TPKC attacks are substantially more difficult than their free-TPKC counterparts. By integrating our improved automatic method with a new rebound attack framework, we successfully identify a new differential characteristic for the 2-round AES-128 fixed-TPKC attack and develope the first practical fixed-TPKC attack against 2-round AES-128. Additionally, we present practical fixed-TPKC attacks against 5-round AES-192 and 3-round Kiasu-BC, along with a practical free-TPKC attack against 6-round Kiasu-BC. Furthermore, we reduce time complexities for free-TPKC and fixed-TPKC attacks on other AES variants.

Mobile crowdsourcing is rapidly growing with Artificial Intelligent of Things. At the same time, the type and complexity of the tasks requested by the requester change and diversify. Therefore, how to design allocation algorithms for the situation of increasing task complexity is particularly critical. In this paper, to cope with this problem, the idea of worker coalition collaboration and reputation evaluation mechanisms are introduced into it. A two-stage allocation based on same-region and cross-region is performed in the divided regional grid. In the first stage, multi-worker and multi-task allocation is realized by combining the reverse auction theory based on the workers’ historical reputation value, which motivates the workers with high reputation value to choose their tasks and contribute data with high sensed quality. The second stage utilizes genetic algorithms to select a coalition of workers for cross-region sensed execution for tasks that do not meet sensed quality requirements. This process will provide additional payoff incentives to compensate for travel costs within the worker coalition, increasing the number of tasks completed and maximizing social welfare. Finally, multiple comparison experiments on the real dataset Yelp are conducted for validation.

The evolution of 5G and 6G technologies has boosted mobile network speed, reduced delays, and widened coverage, empowering Mobile Crowd Sensing (MCS) to overcome surface and terrain obstacles. However, this advancement brings new hurdles for online task assignment. While most MCS methods suit surface applications, they struggle with allocation in complex environments. This paper focuses on MCS in varied settings like surface, air, and high altitude. Currently, planning-based task assignment works better for one-to-one or one-to-many scenarios, with limited options for many-to-many situations. Improving platform utility, attracting top-quality crowd workers, and enhancing task completion efficiency are vital. To tackle these challenges, the paper introduces a spatial division algorithm using 3D Voronoi diagrams for complex environments. This algorithm utilizes task coordinates to delineate assignment spaces. Additionally, it introduces a two-stage many-to-many online task assignment algorithm (MOTA) that forecasts crowd workers’ arrival probabilities and combines auction-based incentives with differential evolution algorithms. MOTA ensures efficient matching of workers and tasks within spatio-temporal constraints, balancing both parties’ interests. Finally, comparative experiments on real datasets assess the proposed MOTA algorithm’s usability and effectiveness based on overall gain, running time, task count, and assignment rate.

With the rapid development of crowd sensing computing, Mobile crowdsourcing (MCS) has become an indispensable part of today’s society. While MCS brings convenience to people, it also exposes them to the risk of privacy leakage. In addition, the demand for data is increasing, and the personalized privacy requirements of crowd workers may affect the service quality. In order to address these problems, this paper proposes a personalized privacy protection incentive mechanism (PPPIM) for MCS based on homomorphic encryption and edge computing. Firstly, this paper designs a personalized privacy metric, using social attributes and private attributes of crowd workers to calculate the privacy level required by crowd workers. Then, based on homomorphic encryption and edge computing, a personalized residual federated security learning scheme (PRFSL) is proposed to ensure the security, timeliness, integrity of task data and the privacy of crowd workers’ needs to improve encryption efficiency. Finally, based on the evolutionary game, a personalized privacy incentive mechanism is proposed to improve the overall service utility. Experimental comparisons based on real datasets show that the proposed scheme can not only ensure the security, timeliness, and integrity of task data more effectively. It can also effectively reduce data processing time, improve the probability of crowd workers actively completing tasks and the overall service quality utility.

The advent of 6G communications technology will bring about a transition from the “Internet of Everything” to the “Intelligent Connection of Everything”. 6G-enabled vehicular ad hoc networks (VANETs) will enjoy lower latency, higher speed, and greater capacity network services. Nevertheless, achieving secure data sharing will be an even tougher challenge. Given this, we propose an attribute-based data sharing scheme with blockchain for 6G-enabled VANETs. First, we propose an efficient multi-tree-based user revocation mechanism. With the Chinese remainder theorem, our mechanism supports user batch revocation and batch joining. Second, we achieve distributed data storage by utilizing the blockchain and smart contracts. To solve the problem of insufficient storage capacity on the blockchain, we adopt a combination of on-chain and off-chain storage. Third, to reduce the computation burden on users, our proposal supports online/offline encryption and verifiable outsourced decryption. Meanwhile, our mechanism supports policy hiding, data revocation, and cross-domain data sharing. The proposed scheme is proven to satisfy the indistinguishability under chosen plaintext attack (IND-CPA) in the standard model. Theoretical analysis shows that our mechanism outperforms existing schemes in functionality and security. Simulation experiments show that our proposal is efficient and suitable for 6G-enabled VANETs.

Vehicular ad hoc network (VANET), as one of the bases of intelligent transport systems, plays an essential role in improving road traffic safety. Nevertheless, in such a complicated, distributed, and highly mobile network structure, how to achieve secure data sharing is a great challenge. The ciphertext-policy attribute-based encryption (CP-ABE) is a potential method to realize one-to-many data sharing for VANET. However, the key abuse problems of users and attribute authorities (AAs) incur many security concerns for VANET. Both issues are extremely important because the attribute keys directly affect users’ access to shared data. To solve the above issues, we propose an accountable attribute-based data-sharing scheme with the blockchain technology (AT-DS-VAHN, in short). For AAs key abuse, we use the consortium blockchain maintained by AAs to achieve distributed key storage and distribution. The attribute keys generated by each AA and its key distribution records are recorded on the blockchain in the form of transactions. Based on the traceability of blockchain, the key abuse behavior of AAs can be caught and prosecuted. For user key abuse, we achieve white-box traceability and efficient user revocation. Based on the principle of traceable-then-revocable, malicious users can be tracked and then revoked directly from the system without complex operations. Besides, to reduce the computation burden on users, our proposal supports online/offline encryption and verifiable outsourced decryption. Security and efficiency analyses show that our proposal is secure and efficient, with high practicability and reliability for VANET.

Vehicular fog computing (VFC), as an extended model of fog computing, combines fog computing with traditional in-vehicle networks to provide real-time response services for users. However, in such a dynamic system architecture, achieving secure and efficient data sharing is an enormous challenge. Ciphertext-policy attribute-based encryption (CP-ABE) is widely regarded as an excellent way of achieving one-to-many data sharing. Nevertheless, several practical challenges hinder its widespread application in VFC, such as inefficient attribute revocation, single-factor access control, and centralized data storage. For this purpose, we design a multifactor combined data sharing scheme for VFC with CP-ABE and blockchain (MC-DS-VFC, in short). We first propose an efficient attribute revocation mechanism that does not require complex key update operations. We then embed time, user attributes, and access interests into data sharing for more fine-grain access control, which enables users with sufficient attributes to efficiently access real-time shared data according to their access interests. Finally, we combine the interplanetary file system (IPFS) and the blockchain maintained by roadside units (RSUs) to achieve distributed collaborative storage. Furthermore, our mechanism also supports user traceability, attribute joining, online/offline encryption, and verifiable outsourced decryption. Our proposal is shown to satisfy the indistinguishability under chosen plaintext attack (IND-CPA) in the standard model. Theoretical analysis and simulation experiments indicate that the MC-DS-VFC scheme is efficient and practical for VFC.