Qin Wang

Rui Gong, Qin Wang, Dengxin Dai et al. · eth-zurich

Unsupervised sim-to-real domain adaptation (UDA) for semantic segmentation aims to improve the real-world test performance of a model trained on simulated data. It can save the cost of manually labeling data in real-world applications such as robot vision and autonomous driving. Traditional UDA often assumes that there are abundant unlabeled real-world data samples available during training for the adaptation. However, such an assumption does not always hold in practice owing to the collection difficulty and the scarcity of the data. Thus, we aim to relieve this need on a large number of real data, and explore the one-shot unsupervised sim-to-real domain adaptation (OSUDA) and generalization (OSDG) problem, where only one real-world data sample is available. To remedy the limited real data knowledge, we first construct the pseudo-target domain by stylizing the simulated data with the one-shot real data. To mitigate the sim-to-real domain gap on both the style and spatial structure level and facilitate the sim-to-real adaptation, we further propose to use class-aware cross-domain transformers with an intermediate domain randomization strategy to extract the domain-invariant knowledge, from both the simulated and pseudo-target data. We demonstrate the effectiveness of our approach for OSUDA and OSDG on different benchmarks, outperforming the state-of-the-art methods by a large margin, 10.87, 9.59, 13.05 and 15.91 mIoU on GTA, SYNTHIA$\rightarrow$Cityscapes, Foggy Cityscapes, respectively.

Linh T. Nguyen, Lam Duc Nguyen, Thong Hoang et al.

Various data-sharing platforms have emerged with the growing public demand for open data and legislation mandating certain data to remain open. Most of these platforms remain opaque, leading to many questions about data accuracy, provenance and lineage, privacy implications, consent management, and the lack of fair incentives for data providers. With their transparency, immutability, non-repudiation, and decentralization properties, blockchains could not be more apt to answer these questions and enhance trust in a data-sharing platform. However, blockchains are not good at handling the four Vs of big data (i.e., volume, variety, velocity, and veracity) due to their limited performance, scalability, and high cost. Given many related works proposes blockchain-based trustworthy data-sharing solutions, there is increasing confusion and difficulties in understanding and selecting these technologies and platforms in terms of their sharing mechanisms, sharing services, quality of services, and applications. In this paper, we conduct a comprehensive survey on blockchain-based data-sharing architectures and applications to fill the gap. First, we present the foundations of blockchains and discuss the challenges of current data-sharing techniques. Second, we focus on the convergence of blockchain and data sharing to give a clear picture of this landscape and propose a reference architecture for blockchain-based data sharing. Third, we discuss the industrial applications of blockchain-based data sharing, ranging from healthcare and smart grid to transportation and decarbonization. For each application, we provide lessons learned for the deployment of Blockchain-based data sharing. Finally, we discuss research challenges and open research directions.

Ismail Nejjar, Qin Wang, Olga Fink

Unsupervised Domain Adaptation Regression (DAR) aims to bridge the domain gap between a labeled source dataset and an unlabelled target dataset for regression problems. Recent works mostly focus on learning a deep feature encoder by minimizing the discrepancy between source and target features. In this work, we present a different perspective for the DAR problem by analyzing the closed-form ordinary least square~(OLS) solution to the linear regressor in the deep domain adaptation context. Rather than aligning the original feature embedding space, we propose to align the inverse Gram matrix of the features, which is motivated by its presence in the OLS solution and the Gram matrix's ability to capture the feature correlations. Specifically, we propose a simple yet effective DAR method which leverages the pseudo-inverse low-rank property to align the scale and angle in a selected subspace generated by the pseudo-inverse Gram matrix of the two domains. We evaluate our method on three domain adaptation regression benchmarks. Experimental results demonstrate that our method achieves state-of-the-art performance. Our code is available at https://github.com/ismailnejjar/DARE-GRAM.

Guangsheng Yu, Xu Wang, Caijun Sun et al.

Federated learning (FL) provides an effective machine learning (ML) architecture to protect data privacy in a distributed manner. However, the inevitable network asynchrony, the over-dependence on a central coordinator, and the lack of an open and fair incentive mechanism collectively hinder its further development. We propose \textsc{IronForge}, a new generation of FL framework, that features a Directed Acyclic Graph (DAG)-based data structure and eliminates the need for central coordinators to achieve fully decentralized operations. \textsc{IronForge} runs in a public and open network, and launches a fair incentive mechanism by enabling state consistency in the DAG, so that the system fits in networks where training resources are unevenly distributed. In addition, dedicated defense strategies against prevalent FL attacks on incentive fairness and data privacy are presented to ensure the security of \textsc{IronForge}. Experimental results based on a newly developed testbed FLSim highlight the superiority of \textsc{IronForge} to the existing prevalent FL frameworks under various specifications in performance, fairness, and security. To the best of our knowledge, \textsc{IronForge} is the first secure and fully decentralized FL framework that can be applied in open networks with realistic network and training settings.

Yuan Tian, Klaus-Rudolf Kladny, Qin Wang et al.

In multi-agent reinforcement learning, multiple agents learn simultaneously while interacting with a common environment and each other. Since the agents adapt their policies during learning, not only the behavior of a single agent becomes non-stationary, but also the environment as perceived by the agent. This renders it particularly challenging to perform policy improvement. In this paper, we propose to exploit the fact that the agents seek to improve their expected cumulative reward and introduce a novel \textit{Time Dynamical Opponent Model} (TDOM) to encode the knowledge that the opponent policies tend to improve over time. We motivate TDOM theoretically by deriving a lower bound of the log objective of an individual agent and further propose \textit{Multi-Agent Actor-Critic with Time Dynamical Opponent Model} (TDOM-AC). We evaluate the proposed TDOM-AC on a differential game and the Multi-agent Particle Environment. We show empirically that TDOM achieves superior opponent behavior prediction during test time. The proposed TDOM-AC methodology outperforms state-of-the-art Actor-Critic methods on the performed experiments in cooperative and \textbf{especially} in mixed cooperative-competitive environments. TDOM-AC results in a more stable training and a faster convergence.

Qin Wang, Guangsheng Yu, Yilin Sai et al.

As Artificial Intelligence (AI) integrates into diverse areas, particularly in content generation, ensuring rightful ownership and ethical use becomes paramount, AI service providers are expected to prioritize responsibly sourcing training data and obtaining licenses from data owners. However, existing studies primarily center on safeguarding static copyrights, which simply treat metadata/datasets as non-fungible items with transferable/trading capabilities, neglecting the dynamic nature of training procedures that can shape an ongoing trajectory. In this paper, we present \textsc{IBis}, a blockchain-based framework tailored for AI model training workflows. Our design can dynamically manage copyright compliance and data provenance in decentralized AI model training processes, ensuring that intellectual property rights are respected throughout iterative model enhancements and licensing updates. Technically, \textsc{IBis} integrates on-chain registries for datasets, licenses and models, alongside off-chain signing services to facilitate collaboration among multiple participants. Further, \textsc{IBis} provides APIs designed for seamless integration with existing contract management software, minimizing disruptions to established model training processes. We implement \textsc{IBis} using Daml on the Canton blockchain. Evaluation results showcase the feasibility and scalability of \textsc{IBis} across varying numbers of users, datasets, models, and licenses.

Zeke Xiao, Yuekang Li, Qin Wang et al.

We explore the feasibility of using LLMs for Automated Exploit Generation (AEG) against vulnerable smart contracts. We present \textsc{ReX}, a framework integrating LLM-based exploit synthesis with the Foundry testing suite, enabling the automated generation and validation of proof-of-concept (PoC) exploits. We evaluate five state-of-the-art LLMs (GPT-4.1, Gemini 2.5 Pro, Claude Opus 4, DeepSeek, and Qwen3 Plus) on both synthetic benchmarks and real-world smart contracts affected by known high-impact exploits. Our results show that modern LLMs can reliably generate functional PoC exploits for diverse vulnerability types, with success rates reaching up to 92\%. Notably, Gemini 2.5 Pro and GPT-4.1 consistently outperform others in both synthetic and real-world scenarios. We further analyze factors influencing AEG effectiveness, including model capabilities, contract structure, and vulnerability types. We also collect the first curated dataset of real-world PoC exploits to support future research.

Yuchen Lei, Yuexin Xiang, Qin Wang et al.

Cryptocurrencies are widely used, yet current methods for analyzing transactions often rely on opaque, black-box models. While these models may achieve high performance, their outputs are usually difficult to interpret and adapt, making it challenging to capture nuanced behavioral patterns. Large language models (LLMs) have the potential to address these gaps, but their capabilities in this area remain largely unexplored, particularly in cybercrime detection. In this paper, we test this hypothesis by applying LLMs to real-world cryptocurrency transaction graphs, with a focus on Bitcoin, one of the most studied and widely adopted blockchain networks. We introduce a three-tiered framework to assess LLM capabilities: foundational metrics, characteristic overview, and contextual interpretation. This includes a new, human-readable graph representation format, LLM4TG, and a connectivity-enhanced transaction graph sampling algorithm, CETraS. Together, they significantly reduce token requirements, transforming the analysis of multiple moderately large-scale transaction graphs with LLMs from nearly impossible to feasible under strict token limits. Experimental results demonstrate that LLMs have outstanding performance on foundational metrics and characteristic overview, where the accuracy of recognizing most basic information at the node level exceeds 98.50% and the proportion of obtaining meaningful characteristics reaches 95.00%. Regarding contextual interpretation, LLMs also demonstrate strong performance in classification tasks, even with very limited labeled data, where top-3 accuracy reaches 72.43% with explanations. While the explanations are not always fully accurate, they highlight the strong potential of LLMs in this domain. At the same time, several limitations persist, which we discuss along with directions for future research.

Rujia Li, Qin Wang, Qi Wang et al.

Smart contracts have recently been adopted by many security protocols. However, existing studies lack satisfactory theoretical support on how contracts benefit security protocols. This paper aims to give a systematic analysis of smart contract (SC)-based security protocols to fulfill the gap of unclear arguments and statements. We firstly investigate \textit{state of the art studies} and establish a formalized model of smart contract protocols with well-defined syntax and assumptions. Then, we apply our formal framework to two concrete instructions to explore corresponding advantages and desirable properties. Through our analysis, we abstract three generic properties (\textit{non-repudiation, non-equivocation, and non-frameability}) and accordingly identify two patterns. (1) a smart contract can be as an autonomous subscriber to assist the trusted third party (TTP); (2) a smart contract can replace traditional TTP. To the best of our knowledge, this is the first study to provide in-depth discussions of SC-based security protocols from a strictly theoretical perspective.

Qin Wang, Shiping Chen, Yang Xiang

Blockchain brings various advantages to online transactions. However, the total transparency of these transactions may leakage users' sensitive information. Requirements on both cooperation and anonymity for companies/organizations become necessary. In this paper, we propose a Multi-center Anonymous Blockchain-based (MAB) system, with joint management for the consortium and privacy protection for the participants. To achieve that, we formalize the syntax used by the MAB system and present a general construction based on a modular design. By applying cryptographic primitives to each module, we instantiate our scheme with anonymity and decentralization. Furthermore, we carry out a comprehensive formal analysis of the proposed solution. The results demonstrate our constructed scheme is secure and efficient.

Qin Wang, Rujia Li, Shiping Chen et al.

NEO is one of the top public chains worldwide. We focus on its backbone consensus protocol, called delegated Byzantine Fault Tolerance (dBFT). The dBFT protocol has been adopted by a variety of blockchain systems such as ONT. dBFT claims to guarantee the security when no more than $f = \lfloor \frac{n}{3} \rfloor$ nodes are Byzantine, where $n$ is the total number of consensus participants. However, we identify attacks to break the claimed security. In this paper, we show our results by providing a security analysis on its dBFT protocol. First, we evaluate NEO's source code and formally present the procedures of dBFT via the state machine replication (SMR) model. Next, we provide a theoretical analysis with two example attacks. These attacks break the security of dBFT with no more than $f$ nodes. Then, we provide recommendations on how to fix the system against the identified attacks. The suggested fixes have been accepted by the NEO official team. Finally, we further discuss the reasons causing such issues, the relationship with current permissioned blockchain systems, and the scope of potential influence.

Rujia Li, Qin Wang, Xinrui Zhang et al.

Blockchain-based cryptocurrencies, facilitating the convenience of payment by providing a decentralized online solution, have not been widely adopted so far due to slow confirmation of transactions. Offline delegation offers an efficient way to exchange coins. However, in such an approach, the coins that have been delegated confront the risk of being spent twice since the delegator's behaviour cannot be restricted easily on account of the absence of effective supervision. Even if a third party can be regarded as a judge between the delegator and delegatee to secure transactions, she still faces the threat of being compromised or providing misleading assure. Moreover, the approach equipped with a third party contradicts the real intention of decentralized cryptocurrency systems. In this paper, we propose \textit{DelegaCoin}, an offline delegatable cryptocurrency system to mitigate such an issue. We exploit trusted execution environments (TEEs) as decentralized "virtual agents" to prevent malicious delegation. In DelegaCoin, an owner can delegate his coins through offline-transactions without interacting with the blockchain network. A formal model and analysis, prototype implementation, and further evaluation demonstrate that our scheme is provably secure and practically feasible.

Bozhi Wang, Qin Wang, Shiping Chen et al.

The Tangle-based structure becomes one of the most promising solutions when designing DAG-based blockchain systems. The approach improves the scalability by directly confirming multiple transactions in parallel instead of single blocks in linear. However, the performance gain may bring potential security risks. In this paper, we construct three types of attacks with comprehensive evaluations, namely parasite attack (PS), double spending attack (DS), and hybrid attack (HB). To achieve that, we deconstruct the Tangle-based projects (e.g. IOTA) and abstract the main components to rebuild a simple but flexible network for the simulation. Then, we informally define three smallest actions to build up the attack strategies layer by layer. Based on that, we provide analyses to evaluate different types of attacks. To the best of our knowledge, this is the first study to provide a comprehensive security analysis of Tangle-based blockchains.

Abdoulaye Diop, Yue Qi, Qin Wang et al.

Wireless Sensor Networks (WSNs) are often deployed in hostile environments, which make such networks highly vulnerable and increase the risk of attacks against this type of network. WSN comprise of large number of sensor nodes with different hardware abilities and functions. Due to the limited memory resources and energy constraints, complex security algorithms cannot be used in sensor networks. Therefore, it is necessary to balance between the security level and the associated energy consumption overhead to mitigate the security risks. Hierarchical routing protocol is more energy-efficient than other routing protocols in WSNs. Many secure cluster-based routing protocols have been proposed in the literature to overcome these constraints. In this paper, we discuss Secure Energy-Efficient Hierarchical Routing Protocols in WSNs and compare them in terms of security, performance and efficiency. Security issues for WSNs and their solutions are also discussed.