Yongqing Zheng
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yongqing Zheng has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Guoxian Yu
Shandong University
4 shared publications - Jun Wang
Shandong University
3 shared publications - Qingzhong Li
Shandong University
2 shared publications - Xiaoqiang Gui
School of Software, Shandong University, Jinan, China
2 shared publications Yinyin Gong
School of Management, University of Shanghai for Science and Technology, Shanghai, China
2 shared publications- Baowen Sun
China Center for Internet Economy Research, Central University of Finance and Economics, Beijing, China
1 shared publication
Latest Publications
- Open Access
- 25 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 44 (3) , 1-40
Federated Recommendation via Stochastic Aggregation and Consistency Inference
With growing concerns over user privacy, federated recommendation (FedRec) has emerged as a mainstream solution for personalized recommendation services. FedRec trains user-private parameters on local clients while collaboratively updating global parameters on a centralized server. However, despite advances in optimizing these local and global parameters, existing methods overlook two key challenges: tradeoff training and distribution discrepancy . Tradeoff training balances timely local updates with diverse global parameters, limiting the model’s learning ability. Distribution discrepancy arises from the divergence between locally trained global parameters and those aggregated by the server, corrupting inference performance. To fill in the gap, we propose FedSC , a principled federated recommendation framework that boosts FedRec’s training and inference processes with minimal yet nontrivial efforts. During training, FedSC employs a stochastic aggregation strategy where all users participate in every round, while only a random subset is selected for aggregation, preserving the diversity of global parameters and ensuring timely local updates. During inference, FedSC makes recommendations with a consistency inference mechanism that uses the most recent locally trained global parameters of each user to improve the model’s understanding of user preferences. Extensive experiments on multiple benchmark datasets demonstrate the superiority of FedSC, achieving up to a 20% improvement in most evaluation scenarios.
- 25 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
- Vol. 44 (3) , 1-40
Open AccessFederated Recommendation via Stochastic Aggregation and Consistency Inference
With growing concerns over user privacy, federated recommendation (FedRec) has emerged as a mainstream solution for personalized recommendation services. FedRec trains user-private parameters on local clients while collaboratively updating global parameters on a centralized server. However, despite advances in optimizing these local and global parameters, existing methods overlook two key challenges: tradeoff training and distribution discrepancy . Tradeoff training balances timely local updates with diverse global parameters, limiting the model’s learning ability. Distribution discrepancy arises from the divergence between locally trained global parameters and those aggregated by the server, corrupting inference performance. To fill in the gap, we propose FedSC , a principled federated recommendation framework that boosts FedRec’s training and inference processes with minimal yet nontrivial efforts. During training, FedSC employs a stochastic aggregation strategy where all users participate in every round, while only a random subset is selected for aggregation, preserving the diversity of global parameters and ensuring timely local updates. During inference, FedSC makes recommendations with a consistency inference mechanism that uses the most recent locally trained global parameters of each user to improve the model’s understanding of user preferences. Extensive experiments on multiple benchmark datasets demonstrate the superiority of FedSC, achieving up to a 20% improvement in most evaluation scenarios.
- Open Access
- 15 October 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-18
Synergistic effects of digital technology and environmental regulation on the green transformation of China’s manufacturing industry
This study builds a framework for the green transformation of manufacturing from the dual perspectives of “technological empowerment” and “institutional regulation.” It integrates digital technology, environmental regulation, and their synergy. Using panel data from 30 Chinese provinces during 2012–2023, we apply a PVAR model to explore the dynamic transmission paths among these factors and the green and low-carbon transformation of manufacturing. The results show: (1) Digital technology has a phased effect on transformation: it inhibits in the short term, improves in the mid-term, and promotes in the long term, reflecting an accumulating response mechanism. (2) Environmental regulation mainly drives transformation efficiency in the short term through incentives, but its effect weakens in the long run. (3) A strong synergy between digital technology and environmental regulation has not yet emerged in the medium and long term, and their joint effects remain limited. (4) The main driving force of green and low-carbon transformation comes from internal development momentum. Compared with environmental regulation, digital technology shows much stronger explanatory power, while the marginal contribution of regulation remains small. This study suggests embedding digital tools into regulatory mechanisms, enhancing the long-term incentives of environmental regulation, and promoting cross-regional data integration to build a dual-driven model for green and low-carbon manufacturing.
- 15 October 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-18
Open AccessSynergistic effects of digital technology and environmental regulation on the green transformation of China’s manufacturing industry
This study builds a framework for the green transformation of manufacturing from the dual perspectives of “technological empowerment” and “institutional regulation.” It integrates digital technology, environmental regulation, and their synergy. Using panel data from 30 Chinese provinces during 2012–2023, we apply a PVAR model to explore the dynamic transmission paths among these factors and the green and low-carbon transformation of manufacturing. The results show: (1) Digital technology has a phased effect on transformation: it inhibits in the short term, improves in the mid-term, and promotes in the long term, reflecting an accumulating response mechanism. (2) Environmental regulation mainly drives transformation efficiency in the short term through incentives, but its effect weakens in the long run. (3) A strong synergy between digital technology and environmental regulation has not yet emerged in the medium and long term, and their joint effects remain limited. (4) The main driving force of green and low-carbon transformation comes from internal development momentum. Compared with environmental regulation, digital technology shows much stronger explanatory power, while the marginal contribution of regulation remains small. This study suggests embedding digital tools into regulatory mechanisms, enhancing the long-term incentives of environmental regulation, and promoting cross-regional data integration to build a dual-driven model for green and low-carbon manufacturing.
- Open Access
DAG-AFL:Directed Acyclic Graph-based Asynchronous Federated Learning
Due to the distributed nature of federated learning (FL), the vulnerability of the global model and the need for coordination among many client devices pose significant challenges. As a promising decentralized, scalable and secure solution, blockchain-based FL methods have attracted widespread attention in recent years. However, traditional consensus mechanisms designed for Proof of Work (PoW) similar to blockchain incur substantial resource consumption and compromise the efficiency of FL, particularly when participating devices are wireless and resource-limited. To address asynchronous client participation and data heterogeneity in FL, while limiting the additional resource overhead introduced by blockchain, we propose the Directed Acyclic Graph-based Asynchronous Federated Learning (DAG-AFL) framework. We develop a tip selection algorithm that considers temporal freshness, node reachability and model accuracy, with a DAG-based trusted verification strategy. Extensive experiments on 3 benchmarking datasets against eight state-of-the-art approaches demonstrate that DAG-AFL significantly improves training efficiency and model accuracy by 22.7% and 6.5% on average, respectively.
Open AccessDAG-AFL:Directed Acyclic Graph-based Asynchronous Federated Learning
Due to the distributed nature of federated learning (FL), the vulnerability of the global model and the need for coordination among many client devices pose significant challenges. As a promising decentralized, scalable and secure solution, blockchain-based FL methods have attracted widespread attention in recent years. However, traditional consensus mechanisms designed for Proof of Work (PoW) similar to blockchain incur substantial resource consumption and compromise the efficiency of FL, particularly when participating devices are wireless and resource-limited. To address asynchronous client participation and data heterogeneity in FL, while limiting the additional resource overhead introduced by blockchain, we propose the Directed Acyclic Graph-based Asynchronous Federated Learning (DAG-AFL) framework. We develop a tip selection algorithm that considers temporal freshness, node reachability and model accuracy, with a DAG-based trusted verification strategy. Extensive experiments on 3 benchmarking datasets against eight state-of-the-art approaches demonstrate that DAG-AFL significantly improves training efficiency and model accuracy by 22.7% and 6.5% on average, respectively.
- Open Access
- 2 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-20
Numerical simulation and governance framework for multi stakeholder symbiotic evolution in digital innovation ecosystems
The symbiotic evolution of stakeholders within digital innovation ecosystems (DIES) is crucial for achieving ecological sustainability. Thus, we integrate the Lotka-Volterra symbiotic evolution model into digital DIES research by constructing both three-actor and four-actor dynamic interaction frameworks and simulating multiple symbiotic evolution scenarios. This study indicates that: (1) The principal actors in DIES are core digital enterprises, innovation partners, digital product users, and digital platforms, with their interdependencies governed by symbiotic coefficients. (2) Introducing the digital platform into the four-actor symbiotic model significantly increases the system’s maximum carrying capacity and equilibrium scale. Parasitic symbiosis and commensal symbiosis may temporarily enhance individual actors’ competitiveness, but they undermine the overall stability of cooperation, whereas mutually beneficial symbiosis provides the optimal pattern for system evolution. (3) Building on these insights, we propose a governance framework centered on resource balancing, value sharing, and trust cooperation, and demonstrate its practical applicability through an Alibaba Cloud case study. This work not only broadens the scope of symbiosis theory within a digital-economy context but also provides actionable guidance for designing efficient, resilient DIES.
- 2 July 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-20
Open AccessNumerical simulation and governance framework for multi stakeholder symbiotic evolution in digital innovation ecosystems
The symbiotic evolution of stakeholders within digital innovation ecosystems (DIES) is crucial for achieving ecological sustainability. Thus, we integrate the Lotka-Volterra symbiotic evolution model into digital DIES research by constructing both three-actor and four-actor dynamic interaction frameworks and simulating multiple symbiotic evolution scenarios. This study indicates that: (1) The principal actors in DIES are core digital enterprises, innovation partners, digital product users, and digital platforms, with their interdependencies governed by symbiotic coefficients. (2) Introducing the digital platform into the four-actor symbiotic model significantly increases the system’s maximum carrying capacity and equilibrium scale. Parasitic symbiosis and commensal symbiosis may temporarily enhance individual actors’ competitiveness, but they undermine the overall stability of cooperation, whereas mutually beneficial symbiosis provides the optimal pattern for system evolution. (3) Building on these insights, we propose a governance framework centered on resource balancing, value sharing, and trust cooperation, and demonstrate its practical applicability through an Alibaba Cloud case study. This work not only broadens the scope of symbiosis theory within a digital-economy context but also provides actionable guidance for designing efficient, resilient DIES.
- Open Access
- 19 June 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
Interaction Privacy Vulnerability in Federated Recommendation and Lossless Countermeasure
Federated Recommendation (FedRec) systems are recognized as privacy-preserving solutions for collaboratively training recommender models without sharing users’ private data. However, recent studies have revealed that FedRec systems are vulnerable to interaction-level membership inference attacks. In such attacks, a semi-honest server can employ crafted methods to infer users’ interacted items. In this paper, we identify that user preference information is predominantly stored in the user-uploaded parameters rather than in the local parameters after local training. Leveraging this insight, we expose a new interaction vulnerability and introduce the PubPara attack. Our experiments show that PubPara improves the inference performance by at least 40% over existing attacks, while requiring the minimal inference time and remaining robust against current defense methods. To safeguard user privacy without compromising recommender performance, we propose MultiVerse, a novel countermeasure. MultiVerse utilizes untrained items outside the user’s local training data to obfuscate the server’s inference of interacted items. It includes a four-step strategy (training, optimization, refinement, and denoising) to achieve robust defense. Extensive experiments on three representative FedRec models (F-NCF, F-LightGCN, and FedRAP) across three real-world datasets validate that MultiVerse significantly degrades the attack’s inference performance to near the level of random guess while maintaining lossless recommender performance.
- 19 June 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Information Systems
Open AccessInteraction Privacy Vulnerability in Federated Recommendation and Lossless Countermeasure
Federated Recommendation (FedRec) systems are recognized as privacy-preserving solutions for collaboratively training recommender models without sharing users’ private data. However, recent studies have revealed that FedRec systems are vulnerable to interaction-level membership inference attacks. In such attacks, a semi-honest server can employ crafted methods to infer users’ interacted items. In this paper, we identify that user preference information is predominantly stored in the user-uploaded parameters rather than in the local parameters after local training. Leveraging this insight, we expose a new interaction vulnerability and introduce the PubPara attack. Our experiments show that PubPara improves the inference performance by at least 40% over existing attacks, while requiring the minimal inference time and remaining robust against current defense methods. To safeguard user privacy without compromising recommender performance, we propose MultiVerse, a novel countermeasure. MultiVerse utilizes untrained items outside the user’s local training data to obfuscate the server’s inference of interacted items. It includes a four-step strategy (training, optimization, refinement, and denoising) to achieve robust defense. Extensive experiments on three representative FedRec models (F-NCF, F-LightGCN, and FedRAP) across three real-world datasets validate that MultiVerse significantly degrades the attack’s inference performance to near the level of random guess while maintaining lossless recommender performance.
- Open Access
- 11 April 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 39 (18) , 19198-19206
Emergence-Inspired Multi-Granularity Causal Learning
Existing causal learning algorithms focus on micro-level causal discovery, confronting significant challenges in identifying the influence of macro systems, composed of micro-level variables, on other variables. This difficulty arises because the causal relationships in macro systems are often mediated through micro-level causal interactions, which can lead to erroneous causal discovery or omission when dispersed. To address this issue, we propose the Emergence-inspired Multi-granularity Causal learning (EMCausal) method. Inspired by the emerging phenomena of aggregating micro level variables into macro level representations, EMCausal introduces a progressive mapping encoder to simulate the process, thus capturing the causal relationships driven by these macro entities. Next, it introduces a causal consistency constraint to collaboratively reconstruct micro variables using macro-level representations, enabling the learning of a multi-granular causal structure. Experimental results on both synthetic and real datasets demonstrate that EMCausal can identify causal graphs under the influence of causal emergence, outperforming competitive baselines in term of accuracy and robustness.
- 11 April 2025
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 39 (18) , 19198-19206
Open AccessEmergence-Inspired Multi-Granularity Causal Learning
Existing causal learning algorithms focus on micro-level causal discovery, confronting significant challenges in identifying the influence of macro systems, composed of micro-level variables, on other variables. This difficulty arises because the causal relationships in macro systems are often mediated through micro-level causal interactions, which can lead to erroneous causal discovery or omission when dispersed. To address this issue, we propose the Emergence-inspired Multi-granularity Causal learning (EMCausal) method. Inspired by the emerging phenomena of aggregating micro level variables into macro level representations, EMCausal introduces a progressive mapping encoder to simulate the process, thus capturing the causal relationships driven by these macro entities. Next, it introduces a causal consistency constraint to collaboratively reconstruct micro variables using macro-level representations, enabling the learning of a multi-granular causal structure. Experimental results on both synthetic and real datasets demonstrate that EMCausal can identify causal graphs under the influence of causal emergence, outperforming competitive baselines in term of accuracy and robustness.
- Open Access
- 26 March 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (3) , e0315275
Optimal presale strategy under the deposit inflation mode: Ordinary or sinking strategy
In the context of the rapid development of e-commerce, the advanced selling strategy has become an important means widely used by online retailers, and the innovation of this advanced selling strategy has been extended to the logistics service level. Based on the consumer utility theory, combined with the strategic behavior of consumers, this paper compares the advantages and disadvantages of ordinary pre-sale strategy (OPS) and sinking pre-sale strategy (SPS) under the deposit inflation mode. The research results show that when consumers have a high level of awareness of logistics services and a moderate proportion of tail payment, the sunken pre-sale strategy has more advantages, and the improvement of consumers’ perception coefficient of logistics, proportion of tail payment, and SPS logistics speed is conducive to the increase of retailers’ profits. In addition, retailers need to be cautious when setting pre-sale prices and deposits to avoid excessive deposit settings leading to the loss of consumers. Regardless of the inflation factor of the deposit, the amount of payment remains unchanged, but the higher coefficient helps to increase demand and profit. The research not only provides theoretical support for retailers in formulating advanced selling strategies but also reveals the close relationship between consumer behavior and the market.
- 26 March 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (3) , e0315275
Open AccessOptimal presale strategy under the deposit inflation mode: Ordinary or sinking strategy
In the context of the rapid development of e-commerce, the advanced selling strategy has become an important means widely used by online retailers, and the innovation of this advanced selling strategy has been extended to the logistics service level. Based on the consumer utility theory, combined with the strategic behavior of consumers, this paper compares the advantages and disadvantages of ordinary pre-sale strategy (OPS) and sinking pre-sale strategy (SPS) under the deposit inflation mode. The research results show that when consumers have a high level of awareness of logistics services and a moderate proportion of tail payment, the sunken pre-sale strategy has more advantages, and the improvement of consumers’ perception coefficient of logistics, proportion of tail payment, and SPS logistics speed is conducive to the increase of retailers’ profits. In addition, retailers need to be cautious when setting pre-sale prices and deposits to avoid excessive deposit settings leading to the loss of consumers. Regardless of the inflation factor of the deposit, the amount of payment remains unchanged, but the higher coefficient helps to increase demand and profit. The research not only provides theoretical support for retailers in formulating advanced selling strategies but also reveals the close relationship between consumer behavior and the market.
- Open Access
- 21 May 2024
- Published by Springer Nature in Discover Computing
- Vol. 27 (1) , 1-20
Fast computation of General SimRank on heterogeneous information network
- 21 May 2024
- Published by Springer Nature in Discover Computing
- Vol. 27 (1) , 1-20
Open AccessFast computation of General SimRank on heterogeneous information network
- Open Access
- 12 January 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Knowledge Discovery From Data
- Vol. 18 (3) , 1-23
Causality-Based Fair Multiple Decision by Response Functions
A recent trend of fair machine learning is to build a decision model subjected to causality-based fairness requirements, which concern with the causality between sensitive attributes and decisions. Almost all (if not all) solutions focus on a single fair decision model and assume no hidden confounder to model causal effects in a too simplified way. However, multiple interdependent decision models are actually used and discrimination may transmit among them. The hidden confounder is another inescapable fact and causal effects cannot be computed from observational data in the unidentifiable situation. To address these problems, we propose a method called CMFL (Causality-based Multiple Fairness Learning). CMFL parameterizes the causal model by response-function variables, whose distributions capture the randomness of causal models. CMFL treats each classifier as a soft intervention to infer the post-intervention distribution, and combines the fairness constraints with the classification loss to train multiple decision classifiers. In this way, all classifiers can make approximately fair decisions. Experiments on synthetic and benchmark datasets confirm its effectiveness, the response-function variables can deal with the unidentifiable issue and hidden confounders.
- 12 January 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Knowledge Discovery From Data
- Vol. 18 (3) , 1-23
Open AccessCausality-Based Fair Multiple Decision by Response Functions
A recent trend of fair machine learning is to build a decision model subjected to causality-based fairness requirements, which concern with the causality between sensitive attributes and decisions. Almost all (if not all) solutions focus on a single fair decision model and assume no hidden confounder to model causal effects in a too simplified way. However, multiple interdependent decision models are actually used and discrimination may transmit among them. The hidden confounder is another inescapable fact and causal effects cannot be computed from observational data in the unidentifiable situation. To address these problems, we propose a method called CMFL (Causality-based Multiple Fairness Learning). CMFL parameterizes the causal model by response-function variables, whose distributions capture the randomness of causal models. CMFL treats each classifier as a soft intervention to infer the post-intervention distribution, and combines the fairness constraints with the classification loss to train multiple decision classifiers. In this way, all classifiers can make approximately fair decisions. Experiments on synthetic and benchmark datasets confirm its effectiveness, the response-function variables can deal with the unidentifiable issue and hidden confounders.
- RetractedOpen Access
- 19 September 2022
- Published by Hindawi Limited in Journal of Sensors
- Vol. 2022, 1-12
5G Smart Sensor Network-Based e-Commerce Investment Risk Management System for SMEs
5G Smart Sensor Network-Based e-Commerce Investment Risk Management System for SMEs: The rapid development of Internet technology in the new era has led to the prosperous development of the e-commerce industry, and more and more traditional retail enterprises are transforming into e-commerce enterprises in line with the development of the times to survive. Based on the relevant financial risk management theories, this paper designs an e-commerce investment risk management system for small and medium-sized enterprises based on the 5G intelligent sensor network, identify the financial risks arising from the e-commerce model, and takes appropriate countermeasures to manage the identified risks. For the optimal coverage problem of homogeneous terminals in the subnetwork of a smart sensing network, this paper introduces the concept of area coverage contribution, transforms the problem of selecting the optimal coverage subset in a large area into the problem of selecting the substrate particles, and organically combines the particle swarm algorithm in the bionic algorithm to consider the network survival period, area sensing coverage, and iteration cost for the hierarchical distributed large-scale IoT architecture. A minimum coverage subset dormancy scheduling algorithm based on homogeneous terminals is proposed. The simulation results are compared with the greedy algorithm and the multioptimized target task assignment algorithm, and it is demonstrated that this algorithm can consider multiple optimization objectives such as energy balance, scheduling cycle, timeout rate, network survival cycle, and system reliability and perform well in the case of large-scale deployment. Finally, an appropriate financial risk evaluation model is constructed to objectively evaluate the identified financial risks, and targeted financial risk control measures and suggestions are proposed based on four aspects: financing, investment, operation, and others.
- 19 September 2022
- Published by Hindawi Limited in Journal of Sensors
- Vol. 2022, 1-12
RetractedOpen Access5G Smart Sensor Network-Based e-Commerce Investment Risk Management System for SMEs
5G Smart Sensor Network-Based e-Commerce Investment Risk Management System for SMEs: The rapid development of Internet technology in the new era has led to the prosperous development of the e-commerce industry, and more and more traditional retail enterprises are transforming into e-commerce enterprises in line with the development of the times to survive. Based on the relevant financial risk management theories, this paper designs an e-commerce investment risk management system for small and medium-sized enterprises based on the 5G intelligent sensor network, identify the financial risks arising from the e-commerce model, and takes appropriate countermeasures to manage the identified risks. For the optimal coverage problem of homogeneous terminals in the subnetwork of a smart sensing network, this paper introduces the concept of area coverage contribution, transforms the problem of selecting the optimal coverage subset in a large area into the problem of selecting the substrate particles, and organically combines the particle swarm algorithm in the bionic algorithm to consider the network survival period, area sensing coverage, and iteration cost for the hierarchical distributed large-scale IoT architecture. A minimum coverage subset dormancy scheduling algorithm based on homogeneous terminals is proposed. The simulation results are compared with the greedy algorithm and the multioptimized target task assignment algorithm, and it is demonstrated that this algorithm can consider multiple optimization objectives such as energy balance, scheduling cycle, timeout rate, network survival cycle, and system reliability and perform well in the case of large-scale deployment. Finally, an appropriate financial risk evaluation model is constructed to objectively evaluate the identified financial risks, and targeted financial risk control measures and suggestions are proposed based on four aspects: financing, investment, operation, and others.