In the complex systems of supply chains, capturing the dynamic nature of multilateral interactions remains a key challenge for link prediction. Traditional approaches often rely on static or dyadic representations, which may overlook evolving higher-order structural patterns. This study proposes a multiplex network link prediction framework that integrates time-evolving multilateral trade patterns explored by network motifs to estimate potential trade flows. Applied to a high-purity quartz supply chain network, which is strategically important to high-tech industries, the method achieves up to 3.6% higher prediction performance compared to baseline models. We demonstrate that motif-based features significantly enhance prediction accuracy by encoding meso-scale evolution of multilateral trade patterns over time and across network layers. Results further reveal distinct roles of countries within the multiplex structure, with the United States, Germany, China, and Japan exhibiting layer-specific dominance. In addition, the United States will lead upstream potential trade, with Eurasia strengthening midstream and downstream trade ties. The proposed approach offers a generalizable methodology for link prediction in evolving multiplex networks, and is helpful for policy guidance of supply chains.

Community detections considering motif-based higher-order features has demonstrated its effectiveness in distinguishing network organizations. However, traditional approaches are susceptible to feature drift resulting from the direct superposition of two-dimensional structures. In order to preserve nodes' structural characteristics while maximize information similarities of adjacent nodes, this study proposes a two-stage community detection framework (Motif-aware Structural Optimization for Community Detection, MotifStruCD) that integrates motif frequency analysis and structural optimization. In the first stage, node neighborhood distributions are extracted through a motif-based transfer probability matrix to preserve both higher-order and lower-order information, and then initial node feature vectors are generated by random walk. In the second stage, node features are first enhanced through graph attention networks, and structural optimization is further accomplished through contrastive learning. Experimental validation through clustering experiments on 19 datasets demonstrates that the MotifStruCD algorithm exhibits high effectiveness with a maximum increase up to 10.87% in the Q-value, and robustness against data noise or structural perturbations are further validated.

IntroductionWith the development of global industrialization, metal minerals have become a global focal point of great power competition. An in-depth investigation into the evolution of global mineral trade structures, alongside an analysis of the relationships between metal minerals trade, resource endowments, industrialization, regional dynamics, and geopolitical competition, is crucial for nations to formulate effective trade policies and enhance the stability of global mineral trade development.MethodsThis study explores the evolution trends of the global metal minerals trade structure from 1990 to 2022 based on complex network and further detect community structure using the Infomap algorithm.Results and discussionThe results show that (1) There is a general upward trend in global metal minerals trade from 1990 to 2022, which can be segmented into phases of slow, rapid, and moderate growth. (2) The two major trading circles formed in the early stage with Japan and Europe as the center have changed, forming a dual pattern with China as the super demand center and Australia as the super supply center. (3) China, Australia, the U.S., Japan, Brazil, and the European Union play key roles in shaping the global trade network, with the structure of the global metal minerals trade network primarily being driven by demand centers. (4) According to the economic trends and the evolving resource demands, the global trade structure will translate from a concentrated bipolar model to a diversified network with multiple trade centers. The conclusion of this research helps specify international policies and maintain supply chain resilience.

IntroductionIn the context of energy transition, the competition for copper resources among countries has intensified, and the global copper trade has become a vitally important trade chain. The global copper ore trade network is influenced by various factors, including resource distribution, supply, demand, prices, transportation costs, etc.MethodsTo understand the evolution process of copper trade network and to predict the trend of supply chain structure evolution in future, in this paper, we construct a spatial weighted complex network evolution model based on complex network theory and gravity model using the import and export data and distance data of countries from 1990 to 2022.Results and discussionSimulation results show that the possibility of establishing copper ore trade between countries follows the spatial weighted complex network evolution model. It is proportional to the expected trade flow between countries and inversely proportional to the distance. The model will support the simulation analysis of the supply chain network structure evolution and help to carry out in-depth research on the forecast of future trade relations between important countries.

Postdoctoral training is a career stage often described as a demanding and anxiety-laden time when many promising PhDs see their academic dreams slip away due to circumstances beyond their control. We use a unique dataset of academic publishing and careers to chart the more or less successful postdoctoral paths. We build a measure of academic success on the citation patterns two to five years into a faculty career. Then, we monitor how students’ postdoc positions—in terms of relocation, change of topic, and early well-cited papers—relate to their early-career success. One key finding is that the postdoc period seems more important than the doctoral training to achieve this form of success. This is especially interesting in light of the many studies of academic faculty hiring that link Ph.D. granting institutions and hires, omitting the postdoc stage. Another group of findings can be summarized as a Goldilocks principle: It seems beneficial to change one’s direction, but not too much.

Temporality, a crucial characteristic in the formation of social relationships, was used to quantify the long-term time effects of networks for link prediction models, ignoring the heterogeneity of time effects on different time scales. In this work, we propose a novel approach to link prediction in temporal networks, extending existing methods with a cognitive mechanism that captures the dynamics of the interactions. Our approach computes the weight of the edges and their change over time, similar to memory traces in the human brain, by simulating the process of forgetting and strengthening connections depending on the intensity of interactions. We utilized five ground-truth datasets, which were used to predict social ties, missing events, and potential links.We found: 1) the cognitive mechanism enables more accurate capture of the heterogeneity of the temporal effect, leading to an average precision improvement of 9% compared to baselines with competitive area under curve (AUC); 2) the local structure and synchronous agent behavior contribute differently to different types of datasets; and 3) appropriately increasing the time intervals, which may reduce the negative impact from noise when dividing time windows to calculate the behavioral synchrony of agents, is effective for link prediction tasks.

IntroductionInnovative energy startups are expediting the energy transition through the adoption of emerging technologies, including blockchain, fintech, artificial intelligence, and crowdfunding. However, existing research primarily focuses on technological capabilities at the startup level and macro-level national applications to explore the current state of innovative energy adoption. In contrast, limited attention has been paid to analyzing the role attributes of innovative energy startups and their correlations with potential success, which are critical for understanding their development trajectories within the energy market.MethodsThis study develops a temporal investment information network for global energy startups, drawing on data from energy enterprises worldwide between 2005 and 2024. The research examines the role attributes of startups and explores the temporal topological characteristics of the network. We propose a success evaluation model based on the features of successful startups to assess the potential of innovative energy startups.Results and DiscussionThe findings indicate that, despite their relatively small market share, innovative energy startups exert significant influence. Notably, successful startups typically exhibit higher betweenness centrality and lower closeness centrality. Moreover, factors such as network degree, centrality, and government administrative capacity play crucial roles in determining the success of innovative energy startups. In the evaluation model constructed using these factors, network structural characteristics contribute the most, achieving an evaluation accuracy of 0.984. This study provides valuable insights for policymakers evaluating innovative energy development trends and for investors assessing the potential of startups.

Finding potential successful startups is always a key issue for industrial innovation and economic development, yet it poses a significant challenge due to the complexity of investments and low success rates. Compared with existing models on knowledge correlations among pairwise startups in a first-order perspective, potential dependencies among sequential investment behaviors reveal knowledge correlations among multiple startups, which requires modeling from a higher order perspective. In this article, a novel higher order network (HON) framework, generated by dependencies among investment behaviors with timestamps, is proposed to identify the pattern of knowledge flows among startups, which has been approved higher accuracy in predicting investment behaviors. Moreover, we introduce a HON-based centrality indicator to measure the importance of startups. Experiments compared with baseline models have shown that the startups identified by proposed indicator are more influential in knowledge propagation and are closer to success. An empirical study conducted by Crunchbase database further reveals that internet-based startups occupy a significant position in investment landscapes, with those associated with finance and commerce not only attracting considerable investments but also facilitating greater success for related startups.

Postdoctoral training is a career stage often described as a demanding and anxiety-laden time when many promising PhDs see their academic dreams slip away due to circumstances beyond their control. We use a unique data set of academic publishing and careers to chart the more or less successful postdoctoral paths. We build a measure of academic success on the citation patterns two to five years into a faculty career. Then, we monitor how students' postdoc positions – in terms of relocation, change of topic, and early well-cited papers – relate to their early-career success. One key finding is that the postdoc period seems more important than the doctoral training to achieve this form of success. This is especially interesting in light of the many studies of academic faculty hiring that link Ph.D. granting institutions and hires, omitting the postdoc stage. Another group of findings can be summarized as a Goldilocks principle: it seems beneficial to change one's direction, but not too much.

Studies have indicated that focusing solely on pairwise interactions between two nodes disregards the associativity among multi-nodes in the network's local structure. This associativity can be seen as dependencies among nodes, where certain edges' presence depends on the path leading to it. Examinations on diverse datasets have approved that the variable order of chained dependencies allows for the preservation of structure information, which enables the reconstruction of the original network into a Higher-Order Network (HON) with improved quality of network representation. This paper proposes a Density-based Higher-Order Network Embedding (DHONE) algorithm, which integrates the concept of higher-order density into the network-embedding process in order to classify the contribution of different orders of dependencies. Through the construction of a novel and effective higher-order adjacency matrix, DHONE steadily improves the accuracy of network representation learning. Experimental results demonstrate DHONEs proficiency in improving embedding accuracy and overall algorithm robustness. Furthermore, grounded in the concept of higher-order density proposed herein, numerous dependencies have been discerned within the network generated from trajectories, potentially indicating the role of multi-node structures in networks.

Link prediction -- to identify potential missing or spurious links in temporal network data -- has typically been based on local structures, ignoring long-term temporal effects. In this chapter, we propose link-prediction methods based on agents' behavioral synchrony. Since synchronous behavior signals similarity and similar agents are known to have a tendency to connect in the future, behavioral synchrony could function as a precursor of contacts and, thus, as a basis for link prediction. We use four data sets of different sizes to test the algorithm's accuracy. We compare the results with traditional link prediction models involving both static and temporal networks. Among our findings, we note that the proposed algorithm is superior to conventional methods, with the average accuracy improved by approximately 2% - 5%. We identify different evolution patterns of four network topologies -- a proximity network, a communication network, transportation data, and a collaboration network. We found that: (1) timescale similarity contributes more to the evolution of the human contact network and the human communication network; (2) such contribution is not observed through a transportation network whose evolution pattern is more dependent on network structure than on the behavior of regional agents; (3) both timescale similarity and local structural similarity contribute to the collaboration network.

In recent years, international energy investment and energy trade activities have developed rapidly. Because energy has commodity and financial product attributes, there is often a correlation between international energy trade and investment. This correlation has regional specificity due to the uneven geographical distribution of energy production and consumption. The existing literature mainly studies the correlation between the international energy “investment–trade” systems from a macroscopic or microscopic perspective. However, the relationship between them among countries from a mesoscopic perspective has not been fully demonstrated. With the development of economic globalization, we need to pay attention to whether the energy trade model of countries can reflect their preference for choosing investment partners and whether the energy investment model can reflect energy trade cooperation. In this paper, by taking the frontier approach of network motifs, we analyzed the correlation between international energy trade and investment from more than 200 economies worldwide from the macroscopic perspective, microscopic perspective, and local structure from a mesoscopic perspective. Meanwhile, we compared the results of the study in 2018 with those in 2022 to obtain the impact of international events on the international energy “investment–trade” networks. We found that 1) the structures of energy trade and investment networks are similar from a macroscopic perspective, which is the basis for exploring the correlation between energy trade and investment. 2) Bilateral cooperation and transaction transmission are important local structures of energy trade and energy investment activities. 3) The formation of an equal and close local structure among economies in energy trade is more likely to be preferred for investment cooperation, and forming a representative local structure with statistical significance among economies in energy investment is more likely to obtain energy trade cooperation. This work innovatively adopts motifs to study the correlation between energy investment and trade, which can help energy investors predict the direction of investment and provide guidance to governments in formulating energy trade policies.

以研究生智能计算课程为例,提出理论、科研与实践3个教学模块间交叉反馈式的教学方法,并从教学难点、教学内容、教学模式、教材建设与教学效果评价方法等方面探讨课程建设,重点探讨教学模块间的交叉反馈机制,从而培养全面发展的人工智能领域人才.

针对学生社会经历少和动手能力薄弱等问题,提出一种基于团队协作的沙盘模拟式项目驱动的软件工程教学模式,分析课程的教学现状,探讨教学内容、教学方法、教学特点和优势、教学评价方法.

结合特色专业探讨课程教学与实践模式创新,对提升高校办学水平并加强办学特色有重要意义.本文提出融合地球科学专业特色的"智能计算"课程教学与实践模式,探讨专业特色在"智能计算"课程中融合的必要性和有效途径.通过提炼典型的地球科学领域问题和智能计算方法,围绕教学目标设计教学内容、教学方法和考核方式并对教学效果进行反思.课程实践将实现地球科学领域人工智能人才综合科研能力的培养.

The innovation and development of emerging technology mostly depend on the way of knowledge convergence defined as the blurring of previously distinct domain-specific knowledge. This paper aims to explore the potential motivation of knowledge convergence and find the law of knowledge convergence, taking the solar energy field as an example. We established Keywords co-occurrence networks of solar energy literature in 2008–2017, and then link prediction is introduced to study the structural mechanism of knowledge convergence. We found that: (1) the common neighbor index better characterizes the knowledge convergence pattern in the knowledge networks among four similarity indicators. (2) The keywords co-occurrence network could effectively mine the structural characteristics of knowledge convergence; (3) the convergence cycle of knowledge in the field of solar energy was about 4 years; (4) keywords with higher betweenness centrality or eigenvector centrality easily generated knowledge convergence; (5) a literature knowledge convergence prediction model is proposed based on these results; and (6) the prediction results showed that scholars should pay attention to six basic issues including energy storage, efficiency, cost, ecological effect, application scenarios, and hybrid photovoltaic systems. This work can provide guidance not only for scholars to grasp the research direction and to generate more innovations but for the government to formulate the policies of government funding.

综合能源系统的脆弱环节辨识对于系统的安全经济运行具有重要意义,复杂网络理论为系统的脆弱环节辨识提供了思路.将应用于电网的带权重的线路介数概念引入到综合能源系统中,并根据综合能源系统特点进行了2点改进.在改进的带权重线路介数的基础上定义了带权重的节点介数.采用基于能源集线器的综合能源系统进行了脆弱性分析,衡量了脆弱环节受到破坏时对系统的影响.通过观察不同攻击策略下系统表现出的脆弱性,验证了所提方法的有效性.

The invention of electric vehicle can make great contribution to the transition to low-carbon city, because of its lower pollution feature. Electric vehicle has become a popular development trend in the automotive industry. This industry has a high degree of technological dependence. Electric vehicle technologies involve multiple disciplines, and study of the technology convergence among various disciplines can explain the interdisciplinary development in this field. Previous studies lack detailed estimation of promising and potential technology convergence relations as well as the detailed topics in this field. This paper combined qualitative and quantitative methods (e.g., network analysis, link prediction and text mining) and proposed a novel framework to explore promising and potential technology convergence relationships and topics in the field. The results are:(1) Eight promising relationships were found. For instance, the convergence relationship B60K–B60L is a promising example related to the convergence of “dashboards or the mounting of one or more propulsion units and related devices” and “the propulsion, operation monitoring or electric safety in electric vehicles”. (2) The topics of the five most promising ones are mainly about battery arrangement and protection, control systems, framework design, and charging connectors. Battery arrangement and protection are the most noteworthy promising topics. (3) Five potential convergence relationships are detected. For example, the convergence of “propulsion, operation monitoring or electric safety in electric vehicles” and “image data processing or generation” might occur in future innovation. (4) The potential topics might be electric vehicle condition display, compressor or pump electric control. The results could offer a reference to scientists, policy-makers and investors in the electric vehicle field, and the novel framework proposed in this paper could be applied to other technology fields to estimate the trends of technology convergence.