Blockchain has been widely adopted across diverse sectors. Yet, enterprise-grade systems remain vulnerable to critical flaws that undermine stability and security. Although academic fuzzing tools such as LOKI and Tyr have shown effectiveness in detecting such issues, their integration into industrial practice remains challenging.

Temporal Awareness—the ability to reason dynamically based on the timestamp when a question is raised—is the key distinction between offline and online video LLMs. Unlike offline models, which rely on complete videos for static, post hoc analysis, online models process video streams incrementally and dynamically adapt their responses based on the timestamp at which the question is posed. Despite its significance, temporal awareness has not been adequately evaluated in existing benchmarks. To fill this gap, we present OVO-Bench (Online-VideO-Benchmark), a novel video benchmark that emphasizes the importance of timestamps for advanced online video understanding capability benchmarking. OVO-Bench evaluates the ability of video LLMs to reason and respond to events occurring at specific timestamps under three distinct scenarios: (1) Backward tracing: trace back to past events to answer the question. (2) Real-time understanding: understand and respond to events as they unfold at the current timestamp. (3) Forward active responding: delay the response until sufficient future information becomes available to answer the question accurately. OVO-Bench comprises 12 tasks, featuring 644 unique videos and approximately human-curated 2,800 fine-grained meta-annotations with precise timestamps. We combine automated generation pipelines with human curation. With these high-quality samples, we further developed an evaluation pipeline to systematically query video LLMs along the video timeline. Evaluations of eleven Video-LLMs reveal that, despite advancements on traditional benchmarks, current models struggle with on-line video understanding, showing a significant gap compared to human agents. We hope OVO-Bench will drive progress in video LLMs and inspire future research in online video reasoning. Our benchmark and code can be accessed at https://github.com/JoeLeelyf/OVO-Bench.

A distributed file system (DFS) is a file system that spans across multiple file servers or multiple locations. The load balancing mechanism in a DFS is crucial, as it optimizes resource utilization across all nodes and improves response times. However, incorrect load scheduling or implementation errors in load balancing algorithms can lead to system imbalance, hang-ups, and even crashes. Such imbalance failures may be critical and pose a significant threat to the availability and security of distributed file systems.

Blockchain systems have implemented various transaction parallelism mechanisms to improve the system throughput and reduce the latency. However, they inevitably introduce bugs. Such bugs can result in severe consequences such as asset loss, double spending, consensus failure, and DDoS. Unfortunately, they have been little analyzed about their symptoms and root causes, leading to a lack of effective detection methods. In this work, we conduct a thorough analysis of historical transaction parallelism bugs in four commercial blockchains. Results show that most of them arise from mishandling conflicting transactions and manifest without obvious phenomena. However, given the heterogeneity of blockchains, it is challenging to trigger conflict handling in a unified way. Effectively identifying these bugs is also hard. Inspired by the findings, we propose Chord, aiming at detecting blockchain transaction parallelism bugs. Chord proposes a unified conflict transaction model to generate various conflict transactions. Chord also dynamically adjust the transaction submission and inserts proactive reverts during transaction execution to conduct thorough testing. Besides, Chord incorporates a local-remote differential oracle and a TPS oracle to capture the bugs. Our evaluation shows that Chord successfully detects 54 transaction parallelism bugs. Besides, Chord outperforms the existing methods by decreasing the TPS by 49.7% and increasing the latency by 388.0%, showing its effectiveness in triggering various conflict scenarios and exposing the bugs.

With the explosive development of mobile devices, mobile crowdsensing (MCS) has emerged as a promising approach for large-scale sensing data collection. In the research of MCS, blockchain technology has been widely adopted to decentralize the traditional mobile crowdsensing and tackle the problem of single point of failure. Incentive mechanisms are devised to boost participation with fairness and truthfulness. However, to better determine the incentive strategy, participants’ privacy can be disclosed on top of the blockchain and obtained by adversaries during the transmission and execution of user data, leading to serious security issues. In this paper, we propose a two-stage incentive scheme with efficiency, robustness and privacy preservation considered based on the combination of blockchain technology and Trusted Execution Environment (TEE). Detailedly, we design two kinds of smart contracts, where on-chain public contracts support the procedure of general crowdsensing interactions, and off-chain private ones enabled by TEE complete the privacy-preserving computations, including an online incentive mechanism for worker recruitment decisions and a truth discovery algorithm for data aggregation. Recovery mechanism and hash check mechanism are introduced to avoid TEE provider failures and TEE providers’ attacks, respectively. Our scheme is proved to be theoretically secure in terms of private information protection, worker participation anonymity, and data aggregation privacy. Experimental results also verify the feasibility and superiority of our incentive scheme.

With the explosive development of mobile devices, mobile crowdsensing (MCS) has emerged as a promising approach for large-scale sensing data collection, while blockchain has been introduced to secure MCS. This research proposes a two-stage incentive scheme with efficiency, robustness and privacy preservation based on the combination of blockchain technology and Trusted Execution Environment (TEE). Detailedly, we design two kinds of smart contracts where on-chain public ones support the procedure of general crowdsensing interactions, and off-chain private contracts enabled by TEE completes the privacy-preserving computations, including an online incentive mechanism for worker recruitment decision and a truth discovery algorithm for data aggregation. Diverse mechanisms are introduced to avoid attacks and failures of TEE providers. Experimental results also verify the feasibility and superiority of our scheme.

In industrial recommendation systems, there are several mini-apps designed to meet the diverse interests and needs of users. The sample space of them is merely a small subset of the entire space, making it challenging to train an efficient model. In recent years, there have been many excellent studies related to cross-domain recommendation aimed at mitigating the problem of data sparsity. However, few of them have simultaneously considered the adaptability of both sample and representation continual transfer setting to the target task. To overcome the above issue, we propose a Entire space Continual and Adaptive Transfer learning framework called ECAT which includes two core components: First, as for sample transfer, we propose a two-stage method that realizes a coarse-to-fine process. Specifically, we perform an initial selection through a graph guided method, followed by a fine-grained selection using domain adaptation method. Second, we propose an adaptive knowledge distillation method for continually transferring the representations from a model that is well-trained on the entire space dataset. ECAT enables full utilization of the entire space samples and representations under the supervision of the target task, while avoiding negative migration. Comprehensive experiments on real-world industrial datasets from Taobao show that ECAT advances state-of-the-art performance on offline metrics, and brings +13.6% CVR and +8.6% orders for Baiyibutie, a famous mini-app of Taobao.

Delays are inevitable in complex distributed environments. Timeout mechanisms are commonly used to handle unexpected failures in distributed systems. However, incorrect timeout handling or implementation errors in timeout mechanisms can lead to system hang-ups or crashes. Such timeout bugs may be crucial and pose a significant threat to the availability and security of distributed systems.In this work, we introduce Chronos, a general testing framework for automatically detecting timeout bugs in distributed systems with deep-priority transient delays. First, we propose general runtime delayed libraries that dynamically inject fine-grained delays in a Distributed System Under Test (DSUT). To effectively trigger delays and constantly explore timeout bugs in deep paths, Chronos harnesses a deep-priority guided fuzzing that dynamically generates high-quality delay sequences in the runtime. Then, Chronos utilizes transient delays to eliminate the time overhead caused by actual delays and accelerate the test process. We implemented and evaluated Chronos on four widely used distributed systems, including ZooKeeper, MySQL-Cluster, HDFS, and Go-Ethereum. Compared with the state-of-the-art techniques, Random, Brute-Force, and Coverage-Guided fault injection, Chronos covers 26.40%, 21.69%, and 15.14% more timeout mechanism logic, respectively. Furthermore, Chronos has detected 27 timeout bugs in these real-world applications, which have been repaired by the corresponding maintainers.

Mobile crowdsensing (MCS) is an efficient approach for large-scale sensing data collection by leveraging the mobility and capability of mobile devices. To avoid the weaknesses of traditional centralized crowdsensing systems, blockchain has been introduced to secure the process of MCS. This paper studies a location-aware scenario, where privacy of users are protected in a blockchain- based MCS system, and formulates an optimization problem to maximize the coverage given a budget based on reverse auction. An incentive mechanism named MMCB is further proposed and implemented as smart contracts in blockchain to solve the problem. We demonstrate that the mechanism achieves a set of desirable properties, including computation efficiency, individual rationality, truthfulness, budget feasibility, approximation, and privacy preservation. To protect the identity privacy of workers and obtain anonymity, a linkable ring signature is employed in smart contracts. In addition, a Pedersen commitment is utilized for protecting workers’ bid profile and the submitted sensing data is encrypted and only accessible to the requester. We implement a prototype system based on the Hyperledger Fabric platform, and the evaluation results show that our privacy-preserving incentive mechanism architecture improves 36.2% coverage and reduces 53.1% payment with better security level compared to the state-of-the-art schemes.

In recent years, mobile crowdsensing (MCS) has been widely adopted as an efficient method for large-scale data collection. In MCS systems, insufficient participation and unstable data quality have become two crucial issues that prevent crowdsensing from further development. Thus designing a valid incentive mechanism is essentially significant. Most of the existing works on incentive mechanism design focus on single-objective optimization with various constraints. However, in the real-world crowdsensing, it is common that several objectives to be optimized exist. Furthermore, constraints on budget or cost are often seen in MCS systems as the feasibility of implementing incentive mechanism is indispensable. This paper studies a bi-objective optimization scenario of MCS to simultaneously optimize total value function and coverage function with budget/cost constraint through a set of problem transformations. Then a budget- or cost-feasible bi-objective incentive mechanism is further proposed to solve the aforementioned bi-objective optimization problem through the combination of binary search and greedy heuristic solution under budget or cost constraint, respectively. Through both rigorous theoretical analysis and extensive simulations, the obtained results demonstrate that the mechanisms achieve computation efficiency, individual rationality, truthfulness, and budget or cost feasibility, while one mechanism obtains an approximation.

Cryptography is a core component of many security applications, and flaws hidden in its implementation will affect the functional integrity or, more severely, pose threats to data security. Hence, guaranteeing the correctness of the implementation is important. However, the semantic characteristics (e.g., diverse input data and complex functional transformation) challenge those traditional program validation techniques (e.g., static analysis and dynamic fuzzing). In this article, we propose CLFuzz, a semantic-aware fuzzer for the vulnerability detection of cryptographic algorithm implementation. CLFuzz first extracts the semantic information of targeted algorithms including their cryptographic-specific constraints and function signatures. Based on them, CLFuzz generates high-quality input data adaptively to trigger error-prone situations efficiently. Furthermore, CLFuzz applies innovative logical cross-check that strengthens the logical bug detection ability. We evaluate CLFuzz on the widely used implementations of 54 cryptographic algorithms. It outperforms state-of-the-art cryptographic fuzzing tools. For example, compared with Cryptofuzz, it achieves a coverage speedup of 3.4× and increases the final coverage by 14.4%. Furthermore, CLFuzz has detected 12 previously unknown implementation bugs in 8 cryptographic algorithms (e.g., CMAC in OpenSSL and Message Digest in SymCrypt), most of which are security-critical and have been successfully collected in the national vulnerability database (7 in NVD/CNVD) and is awarded by the Microsoft bounty program (2 for $1,000).

Resilience is vital to blockchain systems and helps them automatically adapt and continue providing their service when adverse situations occur, e.g., node crashing and data discarding. However, due to the vulnerabilities in their implementation, blockchain systems may fail to recover from the error situations, resulting in permanent service disruptions. Such vulnerabilities are called resilience issues.

Although pre-trained language models~(PLMs) have recently advanced the research progress in mathematical reasoning, they are not specially designed as a capable multi-task solver, suffering from high cost for multi-task deployment (e.g. a model copy for a task) and inferior performance on complex mathematical problems in practical applications. To address these issues, we propose JiuZhang 2.0, a unified Chinese PLM specially for multi-task mathematical problem solving. Our idea is to maintain a moderate-sized model and employ the cross-task knowledge sharing to improve the model capacity in a multi-task setting. Specially, we construct a Mixture-of-Experts (MoE) architecture for modeling mathematical text, to capture the common mathematical knowledge across tasks. For optimizing the MoE architecture, we design multi-task continual pre-training and multi-task fine-tuning strategies for multi-task adaptation. These training strategies can effectively decompose the knowledge from the task data and establish the cross-task sharing via expert networks. To further improve the general capacity of solving different complex tasks, we leverage large language models (LLMs) as complementary models to iteratively refine the generated solution by our PLM, via in-context learning. Extensive experiments have demonstrated the effectiveness of our model.

Blockchain is a decentralized distributed system on which a large number of financial applications have been deployed. The consensus process in it plays an important role, which guarantees that legal transactions on the chain can be executed and recorded fairly and consistently. However, because of Consensus Failure Bugs (CFBs), many blockchain systems do not provide even this basic guarantee. The validity and consistency of blockchain systems rely on the soundness of complex consensus logic implementation. Any bugs which cause the blockchain consensus failure can be crucial.In this work, we introduce Tyr, an open-source tool for detecting CFBs in blockchain systems with a large number of abnormal divergent consensus behaviors. First, we design four oracle detectors to monitor the behaviors of nodes and analyze the violation of consensus properties. To trigger these oracles effectively, Tyr harnesses a behavior divergent model to constantly generate consensus messages and make nodes behave as differently as possible. We implemented and evaluated Tyr on six widely used commercial blockchain consensus systems, including IBM Fabric, WeBank FISCO-BCOS, ConsenSys Quorum, Facebook Diem, Go-Ethereum, and EOS. Compared with the state-of-the-art tools Peach, Fluffy, and Twins, Tyr covers 27.3%, 228.2%, and 297.1% more branches, respectively. Furthermore, Tyr has detected 20 serious previously unknown vulnerabilities, all of which have been repaired by the corresponding maintainers.

Recently, contrastive learning has been shown effective in fine-tuning pre-trained language models (PLM) to learn sentence representations, which incorporates perturbations into unlabeled sentences to augment semantically related positive examples for training. However, previous works mostly adopt heuristic perturbation methods that are independent of the sentence representations. Since the perturbations are unaware of the goal or process of sentence representation learning during training, it is likely to lead to sub-optimal augmentations for conducting constrative learning. To address this issue, we propose a new frameworkL2P-CSRthat adopts a learnable perturbation strategy for improving contrastive learning of sentence representations. In our L2P-CSR, we design a safer perturbation mechanism that only weakens the influence of tokens and features on the sentence representation, which avoids dramatically changing the semantics of the sentence representations. Besides, we devise a gradient-based algorithm to generate adaptive perturbations specially for the dynamically updated sentence representation during training. Such a way is more capable of augmenting high-quality examples that guide the sentence representation learning. Extensive experiments on diverse sentence-related tasks show that our approach outperforms competitive baselines.

This paper aims to advance the mathematical intelligence of machines by presenting the first Chinese mathematical pre-trained language model (PLM) for effectively understanding and representing mathematical problems. Unlike other standard NLP tasks, mathematical texts are difficult to understand, since they involve mathematical terminology, symbols and formulas in the problem statement. Typically, it requires complex mathematical logic and background knowledge for solving mathematical problems.

Conversational recommender systems (CRS) aim to recommend suitable items to users through natural language conversations. For developing effective CRSs, a major technical issue is how to accurately infer user preference from very limited conversation context. To address issue, a promising solution is to incorporate external data for enriching the context information. However, prior studies mainly focus on designing fusion models tailored for some specific type of external data, which is not general to model and utilize multi-type external data. To effectively leverage multi-type external data, we propose a novel coarse-to-fine contrastive learning framework to improve data semantic fusion for CRS. In our approach, we first extract and represent multi-grained semantic units from different data signals, and then align the associated multi-type semantic units in a coarse-to-fine way. To implement this framework, we design both coarse-grained and fine-grained procedures for modeling user preference, where the former focuses on more general, coarse-grained semantic fusion and the latter focuses on more specific, fine-grained semantic fusion. Such an approach can be extended to incorporate more kinds of external data. Extensive experiments on two public CRS datasets have demonstrated the effectiveness of our approach in both recommendation and conversation tasks.

In recent years, mobile crowdsensing has become an effective method for large-scale data collection. Incentive mechanism is fundamentally important for mobile crowdsensing systems. Many mobile crowdsensing systems expect to optimize multiple objectives simultaneously. Most of the existing works transform the multiobjective problem into a single objective problem through constraints or scalarization method. However, due to the uncertain importance (weights) of objectives and the instable quality of crowdsensed data, such transformation is usually unrealizable. In this article, we aim to optimize the worst performance of two objective functions in mobile crowdsensing in order to improve the system robustness. We model an auction-based biobjective robust mobile crowdsensing system, and design two independent objective functions to maximize the expected profit and coverage, respectively. We formulate the robust user selection (RUS) problem, and design an incentive mechanism, which utilizes the combination of binary search and greedy algorithm, to solve the RUS problem. Through both rigorous theoretical analysis and extensive simulations, we demonstrate that the designed incentive mechanisms satisfy desirable properties of computational efficiency, individual rationality, truthfulness, and constant approximation to the tightened RUS problem. Moreover, the proposed incentive mechanism can be easily extended to multiobjective robust mobile crowdsensing systems, and all desirable properties still hold. The simulation results reveal that our incentive mechanism achieves 11% improvement of the platform’s utility, compared with the greedy algorithm for biobjective mobile crowdsensing systems on average.

Conversational recommender systems (CRS) aim to recommend high-quality items to users through interactive conversations.To develop an effective CRS, the support of high-quality datasets is essential.Existing CRS datasets mainly focus on immediate requests from users, while lack proactive guidance to the recommendation scenario.In this paper, we contribute a new CRS dataset named TG-ReDial (Recommendation through Topic-Guided Dialog).Our dataset has two major features.First, it incorporates topic threads to enforce natural semantic transitions towards the recommendation scenario.Second, it is created in a semi-automatic way, hence human annotation is more reasonable and controllable.Based on TG-ReDial, we present the task of topic-guided conversational recommendation, and propose an effective approach to this task.Extensive experiments have demonstrated the effectiveness of our approach on three sub-tasks, namely topic prediction, item recommendation and response generation.TG-ReDial is available at https

Conversational recommender systems (CRS) aim to recommend high-quality items to users through interactive conversations. Although several efforts have been made for CRS, two major issues still remain to be solved. First, the conversation data itself lacks of sufficient contextual information for accurately understanding users' preference. Second, there is a semantic gap between natural language expression and item-level user preference.