Project Metrics Research Articles

Semi-automatic data annotation based on feature-space projection and local quality metrics: An application to cerebral emboli characterization.

We propose a semi-supervised learning approach to annotate a dataset with reduced requirements for manual annotation and with controlled annotation error. The method is based on feature-space projection and label propagation using local quality metrics. First, an auto-encoder extracts the features of the samples in an unsupervised manner. Then, the extracted features are projected by a t-distributed stochastic neighbor embedding algorithm into a two-dimensional (2D) space. A selection of the best 2D projection is introduced based on the silhouette score. The expert annotator uses the obtained 2D representation to manually label samples. Finally, the labels of the labeled samples are propagated to the unlabeled samples using a K-nearest neighbor strategy and local quality metrics. We compare our method against semi-supervised optimum-path forest and K-nearest neighbor label propagation (without considering local quality metrics). Our method achieves state-of-the-art results on three different datasets by labeling more than 96% of the samples with an annotation error from 7% to 17%. Additionally, our method allows to control the trade-off between annotation error and number of labeled samples. Moreover, we combine our method with robust loss functions to compensate for the label noise introduced by automatic label propagation. Our method allows to achieve similar, and even better, classification performances compared to those obtained using a fully manually labeled dataset, with up to 6% in terms of classification accuracy.

Discovering the effect metrics for innovation projects

Working life-integrated innovation projects have become a common part of education. An innovation project is a social phenomenon of learning that brings the competence of several individuals together in a social process through which a novel idea is turned into practical reality to convey value to its users. University–industry innovation collaboration research has mainly understood the role of the university as a partner responsible for research. However, these projects are often based on student teamwork, an open challenge from real working life and a predefined intention to produce a concrete, novel, and innovative product, service, or new operating model into use. The outcomes are meant to be taken into use to convey value for business and society. A paucity of empirical research exists on the development and assessment of these innovation education programs based on their effects on the surrounding society. The variables that are important in studying the effects of innovation education projects on surrounding society, participating companies, and organisations must be identified. In this study, a set of effect measurement scale variables were discovered through a document analysis and focus group interviews. The findings imply the importance of recognising the effects of the project both on the clients' and their organisations' goals (project output, benefits of participation, and what survives) and the wider impact of the projects on the surrounding society (impacts on the daily lives of citizens; well-being and health; economy, ecology, digitalisation; and social matters). Thus, a preliminary impact assessment model is suggested. Keywords: Innovation project, university-industry collaboration, impact, higher education, innovation pedagogy, document analysis, focus group

Boundary-Rigidity of Projective Metrics and the Geodesic X-Ray Transform

We prove that given a compact convex non-empty domain {mathcal {M}} in the plane, a function delta :partial mathcal Mtimes partial {mathcal {M}}rightarrow {mathbb {R}}_+ can be extended to a projective metric d on {mathcal {M}} if and only if delta (P,R)+delta (Q,S)-delta (P,S)-delta (Q,R)>0 for any convex quadrangle Box (PQRS) inscribed in partial {mathcal {M}}. Moreover, this extension is unique.

Finslerian Projective Metrics with Small Quadratic Spheres

Finslerian Projective Metrics with Small Quadratic Spheres

An optimisation model for incentivising the development of marginal oil and gas fields amidst increasingly complex ownership patterns: UKCS case study

Many recent discoveries in the UKCS have been economically marginal and comparatively small such that they are unable to be developed on a standalone basis. Nevertheless, there is a proliferation of production facilities and transport infrastructure that can, on accepting a tariff, process hydrocarbons from these smaller-sized oil and gas fields, thereby contributing to the UK Government's strategy of maximising economic recovery (MER). However, the empirical dimensions of cost-sharing arrangements between a field owner (host field or hub) and tie-in (satellite) fields are not well examined in the literature although they influence long-term economics in matured oil provinces. In this paper, we develop a mixed integer programming optimisation model to analyse how these third-party access to infrastructure issues impact MER in the UKCS. A first-best situation with only one regional operator is empirically compared with the actual situation of multiple ownership. We then assess the impact of different infrastructure unbundling provisions via pipeline tariff choices and their interaction with the fiscal regime on the net present value (NPV) and other project metrics such as timings of hub and field shutdowns. We find that differences in field ownership, tariff choice and changes to the fiscal regime impact the overall NPV of field developments. Our results also show that having a progressive fiscal regime in a mature basin such as the UKCS is important to support continued operations under low oil prices while increasing the government's take at high prices. The cost-sharing effect shows that imposing strict cost-sharing rules does not change overall project valuations, even in a low oil price scenario. An additional contribution to the extant literature is that tariff determination could be based on cost-sharing rules that enjoin each field tieback to a hub to pay a split tariff, comprised of a fixed cost of service (access charge) and variable (marginal) costs. • Presents MIP model for analysing economic dependencies of infrastructure assets. • Ownership style, choice of tariffs, and fiscal regime impact overall field value. • Reducing marginal tax rates improves profitability, reflecting neutrality and progressivity. • Tariffs are decisive as to whether satellite or user fields are developed. • Tariff determination to maximise MER should be based on two-part cost-sharing.

Projective Metric Geometry and Clifford Algebras

Each vector space that is endowed with a quadratic form determines its Clifford algebra. This algebra, in turn, contains a distinguished group, known as the Lipschitz group. We show that only a quotient of this group remains meaningful in the context of projective metric geometry. This quotient of the Lipschitz group can be viewed as a point set in the projective space on the Clifford algebra and, under certain restrictions, leads to an algebraic description of so-called kinematic mappings.

Genetics and the question of purity in cutthroat trout restoration

As molecular techniques become more advanced, scientists and practitioners are calling for restoration to leverage genetic and genomic approaches. We address the role of genetics in the restoration and conservation of cutthroat trout in the western United States, where new genetic insights have upended previous assumptions about trout diversity and distribution. Drawing on a series of examples, we examine how genetically pure trout populations are identified, protected, and produced through restoration practices. In landscapes that have been profoundly impacted by human activities, genetics can offer seemingly objective metrics for restoration projects. Our case studies, however, indicate that (1) genetic purity is fragile and contingent, with notions of what genetics are “pure” for a given species or subspecies continually changing, and (2) restoration focused on achieving “genetically pure” native populations can deliberately or inadvertently obscure the socioecological histories of particular sites and species, even as (3) many “genetically pure” trout populations have endured on the landscape as a result of human modifications such as roads and dams. In addition to raising conceptual questions, designations of genetic purity influence policy. These include tensions between restoring connectivity and restoring genetic purity, influencing Wild and Scenic River Act designations, and the securing of water rights. Cutthroat trout restoration would benefit from adopting a broader, more holistic framework rather than fixating exclusively or primarily on genetic purity and hybridization threats.

Sustainability assessment framework and methodology with trans-disciplinary numerical simulation model for analytical floatovoltaic energy system planning assessments

Floatovoltaics is rapidly emerging as a novel type of sustainable energy technology, in which solar photovoltaic installations are sited directly on open-water spaces. As an agro-renewable energy-generation technology, it makes dual use of water to generate revenue from under-utilised irrigation water surfaces while also offering mutually beneficial layers of land-saving, environmental conservation and water-preservation benefits. Standardised metrics for ground-mounted photovoltaic projects, however, do not properly account for the technology’s extended range of resource-use-efficiencies and impact-effect-positives. Such knowledge gaps hinder evidence-based scientific assessments in regulatory project permissions mandated by law. Technology planning and impact assessment practices can benefit from a computer-aided technique to characterise floatovoltaic performance profiles. This paper introduces a conceptual empirical modelling framework, a holistic system dynamics-thinking methodology and a computer synthesis model to empirically predict the performance and sustainability profiles of prospective floatovoltaic installations. By inherently exploring the techno-economic and techno-environmental externalities of floatovoltaic enterprises, it translates performance profiles into sustainability indicators, articulated as WELF-nexus parameters. The paper details the integrated analytical framework, mathematical modelling formulation and digital computer synthesis model towards quantitative floatovoltaic energy system planning and sustainability assessments. The study’s main finding is that an integrated techno-enviro-economic floatovoltaic assessment methodology can be successfully modelled as a context-sensitive synthesis technique in a system dynamics modelling environment. The proposed technique can find utility in solving real-world problems with assessments in efficiency, feasibility and sustainability for agricultural floatovoltaics.

Friendly Calls to Seniors: an Interprofessional Student Volunteer Program

<h3>Introduction</h3> Older adults compose a segment of the population who are especially vulnerable to loneliness and isolation and have been disproportionately impacted by the effects of the COVID-19 pandemic.^{[1],[2],[3],[4]} Evidence shows traditional social support programs enhance their health and well-being; the current pandemic provides an opportunity to understand the potential benefits of virtual social support.^[5],[6] Friendly Calls to Seniors (FCS) is an interprofessional student volunteer program that is part of Columbia University's COVID-19 Student Service Corps (CSSC), founded at the outset of the pandemic. Utilizing a script developed by student leaders and faculty advisors from Columbia University Irving Medical Center (CUIMC), volunteers regularly call older adults with underlying mental health conditions. The calls provide social support and opportunities to identify and refer health concerns to care providers, thereby addressing both medical and social needs this group faces, particularly during the COVID-19 pandemic. <h3>Methods</h3> FCS is a phone call-based social support program in which student volunteers provide companionship and conduct health risk assessments for interested clients, who are older adults with underlying psychiatric conditions referred by clinical providers in the Washington Heights region of Manhattan. Student volunteers are on-boarded, HIPAA-trained, and supervised by two student leaders, as well as two faculty members of CUIMC, a geriatric psychiatrist and an occupational therapist. Volunteers use a HIPAA-compliant platform to securely call clients, using a translator if necessary. Calls are conducted using a script, which includes a suggested introduction, sample conversational topics, and a risk assessment for health and functioning that incorporates psychiatric and occupational considerations. The script also includes an escalation flowchart to refer client issues to the attention of the faculty supervisors to be connected with social workers, medical professionals, or community resources. Calls are tracked through secure survey forms and project metrics are tabulated weekly. The volunteer team convenes in weekly Zoom meetings to review volunteer-client interactions and pose questions that may have arisen. The project has integrated educational workshops from other organizations such as Columbia's Narrative Medicine Program. The student and faculty leadership also collaborates with community-based organizations and naturally occurring retirement communities (NORCs) in the Washington Heights region. <h3>Results</h3> At the time of writing, a total of 400 calls have been made by FCS student volunteers to 72 unique clients, ranging in age from 61 to 96. Of these, 38 are English-speaking and 34 are Spanish speaking; 49 clients elected to continue receiving calls. All clients have existing mental and physical health conditions, including depression, anxiety, neurocognitive disorders, bipolar disorder, and PTSD. Medical comorbidities, which were highly prevalent, existing in 45% of clients, included diabetes mellitus and hypertension. In total, 48 escalations have been made since the start of the program. Table 1 displays a breakdown of the escalation reasons: There have been 41 total volunteers since the program's inception, of whom 18 are currently active. The volunteers represent a wide range of health professions and undergraduate programs at Columbia, as seen in Table 2. Qualitative volunteer feedback about FCS was obtained through call log responses and reflection sessions. Volunteers cited increased knowledge of community resources and improved skills in relationship-building and addressing geriatric-specific concerns, such as discussing mental health concerns in an age-attuned and culturally sensitive manner. Additional skills gained included the ability to properly coordinate with an interpreter and communicate across language barriers; many students had never worked with an interpreter prior to being a part of FCS. <h3>Conclusions</h3> The FCS program exhibits promising qualitative results supporting vulnerable older adults. Using a script with a systematic risk assessment, volunteers have performed regular outreach to seniors and referred them to appropriate social services and/or healthcare professionals as needed. Clients and the FCS team have mutually benefited from the interprofessional perspectives informing this program, which have led to improved understanding of professional roles, available community resources, and opportunities for comprehensive care. Looking ahead, FCS will continue to be relevant as the COVID-19 pandemic draws on and many seniors remain socially isolated. The team will develop quantitative outcome metrics to better understand the impact of its work. The interprofessional, community-based education model, exemplified by the FCS interprofessional project, will be developed as a generalizable model for ensuring quality of care and best practice. <h3>Funding</h3> none

ECONOMIC VALUATION OF THE Eucalyptus urograndis PRODUCTION IN TWO TYPES OF CONTAINERS

The systematic treatment of uncertainties associated with forest production processes allows the establishment of technical and managerial strategies to mitigate the risk inherent in the application of capital demanded in the production of cuttings. Therefore, the aim of this research was to ascertain the unit risk to produce Eucalyptus cuttings using the Monte Carlo simulation method. Based on the estimate of the production cost of Eucalyptus cuttings and in the expected revenues of six productive systems, characterized as forestry investment projects, it was possible to construct cash flows followed by the metrics used for the decision of mutually exclusive investment projects. The incorporation of risk analysis had as propositions the uncertainties associated with the productive factors, in which probability distributions were attributed. Then, pseudorandom numbers were generated by the Monte Carlo simulation method, with the purpose of measuring the stochastic economic values and the probabilities of occurrences for the project metrics. The stochastic results showed that the probability of economic viability for the analyzed conditions is null. The most critical stochastic variables of forestry projects are investments and the cycle production of cuttings.

ISCOUT: Science-task planning and formation maneuver design for the IRASSI space interferometer

Abstract IRASSI is an astronomy mission aimed at observing stellar disks and protoplanetary regions using a constellation of five free-flying telescopes operating around the Sun-Earth/Moon L2 point. In these regions of the cosmos, important chemical/physical phenomena can be observed in the far-infrared, which may advance our current understanding of how pre-biotic conditions in Earth-like planets are formed. The distinguishing aspect of the IRASSI interferometer, however, is its free-flying concept. Rather than relying on a fixed formation configuration and relative position control, IRASSI relies on continuously drifting baselines (e.g., inter-satellite distances) and the knowledge of these baselines during the science observations. However, not all IRASSI targets require the same formation strategy and therefore the planning of the sequence of observation tasks and of the corresponding reconfiguration and science maneuvers has been investigated. The manuscript provides thus an overview of the planning architecture devised for IRASSI, called iSCOUT, composed of different modules: 1) Target Planner Module (TPM), 2) Reconfiguration Module (ReM), 3) Collision- and invisibility Avoidance Module (CAM), 4) Baseline Pattern Module (BPM). The first module (TPM), generates an optimized sequence of targets within a specified window period, taking into account variables such as visibility of the targets at each time step, retargeting time based on the current state, target priority or number of previously observed targets of the same type. The second module, ReM, assigns positions to the telescopes based on the target sequence generated by the TPM. Each translation maneuver is optimized to achieve a specified initial formation configuration, which is both perpendicular to the target direction and three-dimensional. Parameters such as fuel use and fuel-balance as well as ΔV and ΔV-balance are used for optimizing the reconfiguration of the formation. The third module, CAM, ensures that the telescopes following the ReM trajectories do not violate relative proximity restrictions or intervisibility requirements during the reconfiguration. In such an event, the CAM computes a new path in order to avoid collisions and invisibility instances. The last module, BPM, calculates the drifting trajectories that achieve the required patterns in the so-called ‘(U, V)-plane’ during the science observations. (U, V)-plane projection metrics and several operational constraints are used for the optimization of the trajectories. Each module has been individually optimized and the paper describes the main outcomes thereof. Results concerning and end-to-end iSCOUT simulation for the IRASSI mission are presented and future work goals complete the paper.

Freelancing Models for Fostering Innovation and Problem Solving in Software Startups: An Empirical Comparative Study

Context: freelancers and startups could provide each other with promising opportunities that lead to mutual growth, by improving software development metrics, such as cost, time, and quality. Niche skills processed by freelancers could help startups reduce uncertainties associated with developments and markets, with the ability to quickly address market issues (and with higher quality). This requires the associations between freelancers and startup to be long-term, based on trust, and promising agreements driven by motivations (leading to the growth of both parties). Freelancers could help startups foster innovations and undertake software development tasks in better ways than conducted in-house, if they are selected using informed decision-making. Objectives: the paper has three objectives, (1) to explore the strategies of startups to outsource software development tasks to freelancers (termed as freelancing association strategies); (2) to identify challenges in such outsourcings; and (3) to identify the impacts of outsourcing tasks to freelancers on overall project metrics. The overall objective is to understand the strategies for involving freelancers in the software development process, throughout the startup lifecycle, and the associated challenges and the impacts that help to foster innovation (to maintain competitive advantages). Method: this paper performs empirical studies through case studies of three software startups located in Italy, France, and India, followed by a survey of 54 freelancers. The results are analyzed and compared in the identification of association models, issues, challenges, and reported results arising because of such associations. The case study results are validated using members checking with the research participants, which shows a higher level of result agreements. Results: the results indicate that the freelancer association strategy is task based, panel based, or a hybrid. The associations are constrained by issues such as deciding pricing, setting deadlines, difficulty in getting good freelancers, quality issues with software artefacts, and efforts to access freelancer work submissions for reward. The associations have a positive impact on software development if there is availability of good freelancers (which lasts long for various tasks). The paper finally provides a freelancing model framework and recommends activities that could result in making the situation beneficial to both parties, and streamline such associations. Fostering innovation in startups is, thus, a trade-off situation, which is limited and supported by many conflicting parameters.

Métricas basadas en un enfoque guiado por riesgos para evaluar la comunicación, cooperación y coordinación en equipos de desarrollo global de software

El éxito del desarrollo global de software (DGS) depende, en gran parte, de la forma como se comunican, cooperan y coordinan (3Cs) los esfuerzos en la gestión de los proyectos y los equipos de trabajo geográficamente dispersos. En la literatura, se identifican varios trabajos que proponen riesgos y métricas para proyectos de DGS, pero algunos de ellos no se describen con detalle, son ambiguos o complicados de entender, lo que dificulta ubicarlos en alguna categoría precisa. Además, es necesario clasificar los riesgos de una manera menos subjetiva, que mejore su comprensión; facilite la definición y aplicación de herramientas para medir su impacto, y genere algunos indicadores que apoyen la toma de decisiones en los equipos de DGS. En este artículo, se proponen 37 riesgos para la comunicación, cooperación y coordinación en el DGS, como resultado de la comparación e integración de riesgos identificados en la literatura. También se presenta un conjunto de métricas definidas con la aplicación del paradigma Goal Question Metric (GQM), organizadas con estas interrogantes: ¿qué mide?, donde se establece el atributo por calibrar, y ¿cómo lo mide?, donde se indica la métrica a través de los siguientes elementos: descripción, tipo, escala y función de cálculo. Mediante la técnica de grupo focal, se evaluó la completitud, pertinencia, claridad, precisión y facilidad de uso de las métricas, que podrían resultar útiles para las organizaciones, equipos, proyectos y consultores que operan en contextos de DGS.

Revising Computer Science Networking Hands-On Courses in the Context of the Future Internet

Contribution: In a context where hands-on courses are biased toward specific technologies, a novel creativity-provoking instructional approach for networking undergraduate courses is successfully applied following action research principles and active and creative learning techniques. Background: Extensive engineering-oriented networking courses have been proposed with a strong focus on specific protocol solutions. At the same time, the amount and complexity of techniques is notably increasing with the advent of the Future Internet. As a result, the curricula looses focus on the fundamentals of networking algorithms. Intended Outcomes: We address algorithmic learning in networking for computer sciences, where students are expected to 1) create; 2) develop; 3) analyze; and 4) compare algorithms and processes regardless of protocol-specific technologies. At least 70% of the students are expected to meet this goal while enhancing their engagement and motivation in a time-constrained course schedule. Application Design: To achieve 1) and 2), we instrument an active experimental strategy, while objectives 3) and 4) are tackled with creative learning techniques, both applied in an action research framework. The approach is supported by state-of-the-art networking application interfaces and simulators. Furthermore, a blended and game learning component favors the engagement via comparison and competition of students' project metrics. Findings: The experiment is carried out by professors of the Computer Science Bachelor's degree taught in FAMAF. Results show that the applied methodology met the intended outcomes, and improved by 7% in a two-year cycle. Furthermore, the approach was very well received based on student's feedback.

Estimated View of Renewable Resources as a Sustainable Electrical Energy Source, Case Study

Renewable energy resources for energy generation in the Gaza Strip (GS) emerge as a complementary resource, not only to meet the need of society for energy in various forms, but also to find available, sustainably, and less costly resources for institutions as with individuals. These resources could provide safety in use, increase energy security, and address environmental concerns. This work aimed to assess the renewable energy potential in the GS, which has few natural resources and a real crisis in energy supply. Therefore, an integrated evaluation methodology is assumed for renewable energy potential; the methodology comprises mathematical as well as software programming to assess the renewable energy. The results reveal the assessment of the existing four renewable energy sources: solar, wind, wave, and biomass. The estimations exhibit that the solar and biomass resources are dominant. The methodology consists of mathematical models for calculating the energy outputs from different resources in the study area. ArcGIS 10.1 software used for calculating solar radiation, WindSim 9.0.0 software for estimating wind energy and OpenWEC software tool for evaluating the force stored in coastal waves. The potential of each energy and the integration of all are depicted in individual maps by ArcGIS. It was found that if anaerobic digestion technology is applied, 65.15 MW electric power could be produced from about one-and a third million tons/year of biomass. Depending on the available statistical data due to the specified annual time period and the household electric demand that is about 120 MW at the corresponding time, this means that the energy from biomass can cover more than half of the demand. In addition, solar energy potentials were estimated to reach 1195 GWh/year. Therefore, the existing power may be replaced by biomass and solar energy. It also demonstrates that the energy potential maps can be used as metrics for possible sustainability projects in the GS.

New Statistical Metrics for Multisite Replication Projects

SummaryIncreasingly, researchers are attempting to replicate published original studies by using large, multisite replication projects, at least 134 of which have been completed or are on going. These designs are promising to assess whether the original study is statistically consistent with the replications and to reassess the strength of evidence for the scientific effect of interest. However, existing analyses generally focus on single replications; when applied to multisite designs, they provide an incomplete view of aggregate evidence and can lead to misleading conclusions about replication success. We propose new statistical metrics representing firstly the probability that the original study's point estimate would be at least as extreme as it actually was, if in fact the original study were statistically consistent with the replications, and secondly the estimated proportion of population effects agreeing in direction with the original study. Generalized versions of the second metric enable consideration of only meaningfully strong population effects that agree in direction, or alternatively that disagree in direction, with the original study. These metrics apply when there are at least 10 replications (unless the heterogeneity estimate τ^=0, in which case the metrics apply regardless of the number of replications). The first metric assumes normal population effects but appears robust to violations in simulations; the second is distribution free. We provide R packages (Replicate and MetaUtility).

Utilizing patient advocates in Parkinson's disease: A proposed framework for patient engagement and the modern metrics that can determine its success.

The wide application of patient engagement and its associated benefits has increased across government, academic and pharmaceutical research. However, neither an identified standard practice for the process of engagement, nor utilization of common metrics to assess associated outcomes, exists. Parkinson's Foundation developed a patient engagement framework and metrics to assess engagement within the academic research and drug development sectors. This approach was developed over the course of several years through assessing the literature, acquiring feedback from researchers and people with Parkinson's disease and adapting practices to be relevant and generalizable across patient engagement projects. This framework includes the: 1) creation of a scope of work, 2) establishment of guiding principles, 3) selection and training of participants, 4) co‐determination of project metrics, 5) execution of the project and 6) dissemination of project findings. Parkinson's Foundation has also worked with academic, government and pharmaceutical stakeholders to identify metrics that assess both the quality of patient engagement and outcomes associated with patient engagement on projects. By improving patient engagement project methodologies and metrics, global clinical trials can have access to evidence‐based patient engagement practices to more efficiently capture the needs of, and potentially benefit, the patient community.

Incentives and Gaming in Collaborative Projects Under Risk-Sharing Partnerships

Problem definition: This article explores the incentive issues and gaming behaviors of firms under risk-sharing partnerships in a project management setting motivated by real-life examples. Academic/practical relevance: Collaboration prevails in projects within diverse industries. The risk-sharing partnership, in which each partner pays for its own cost and shares the outcome (either reward or loss) on project completion, is one of the most popular ways to manage collaborations in practice. However, the risk-sharing partnership may lead to project failure in the forms of excessive delays and cost overruns, but the driving forces (for example, incentives) and mechanisms (for example, gaming behaviors) in project management settings are not yet fully understood. Methodology: Relative to the one-firm-does-all strategy, we studied how risk-sharing partnerships may affect firms’ incentives in project execution and thus, project metrics (duration and cost) for various project networks (serial versus parallel), risk levels (deterministic versus stochastic duration), and information status (symmetry versus asymmetry). Results: We found that risk-sharing partnerships may encourage deliberate delays and cost overruns through various mechanisms, such as the Prisoner’s Dilemma, the Supplier’s Dilemma, and the Coauthor’s Dilemma. Counterintuitively, information asymmetry may outperform information symmetry on project metrics for both deterministic and stochastic duration, contingent on the network structure, cost parameters, and partners’ beliefs. Managerial implications: By connecting theory to practice, we provide insights into the incentive issues of some real-life projects and justifications for several mitigation strategies to avoid such gaming behaviors in practice.

A game theory approach to the existence and uniqueness of nonlinear Perron-Frobenius eigenvectors

We establish a generalized Perron-Frobenius theorem, based on a combinatorial criterion which entails the existence of an eigenvector for any nonlinear order-preserving and positively homogeneous map $f$ acting on the open orthant $\mathbb{R}_{\scriptscriptstyle >0}^n$. This criterion involves dominions, i.e., sets of states that can be made invariant by one player in a two-person game that only depends on the behavior of $f$ "at infinity". In this way, we characterize the situation in which for all $\alpha, \beta > 0$, the "slice space" $\mathcal{S}_\alpha^\beta := \{ x \in \mathbb{R}_{\scriptscriptstyle >0}^n \mid \alpha x \leq f(x) \leq \beta x \}$ is bounded in Hilbert's projective metric, or, equivalently, for all uniform perturbations $g$ of $f$, all the orbits of $g$ are bounded in Hilbert's projective metric. This solves a problem raised by Gaubert and Gunawardena (Trans. AMS, 2004). We also show that the uniqueness of an eigenvector is characterized by a dominion condition, involving a different game depending now on the local behavior of $f$ near an eigenvector. We show that the dominion conditions can be verified by directed hypergraph methods. We finally illustrate these results by considering specific classes of nonlinear maps, including Shapley operators, generalized means and nonnegative tensors.

ADPE: Adaptive Dynamic Projected Embedding on Heterogeneous Information Networks

Network embedding (NE) aims to represent network information appropriately by learning low-dimensional and dense vectors for the nodes and edges of information network. Actually, the real world is almost full of heterogeneous information networks, which stimulates the emergence of heterogeneous information networks (HINs) embedding models. However, parts of existing HIN embedding models like meta-path-based methods only capture limited and aggregated information of relations, whereas some models based on metric or distance learning are usually of high computational complexity and slow training speed. In this paper, we present a novel heterogeneous information network embedding model, which applies dynamic projection metrics and translation mechanisms to the complicated heterogeneous information networks including multiple nodes and different relations. In order to overcome the imbalance of the distribution of relations in HIN and optimize the training process, we introduce an adaptive loss function for model optimization. Further more, we propose a hybrid model with baseline method as the initialization of the model. Experiments have been implemented on some real-world HIN datasets. And empirical results show that our model significantly outperforms the state-of-the-art representation learning models.