Showing papers by "HEC Montréal published in 2017"

Graph Attention Networks

Abstract: We present graph attention networks (GATs), novel neural network architectures that operate on graph-structured data, leveraging masked self-attentional layers to address the shortcomings of prior methods based on graph convolutions or their approximations. By stacking layers in which nodes are able to attend over their neighborhoods' features, we enable (implicitly) specifying different weights to different nodes in a neighborhood, without requiring any kind of costly matrix operation (such as inversion) or depending on knowing the graph structure upfront. In this way, we address several key challenges of spectral-based graph neural networks simultaneously, and make our model readily applicable to inductive as well as transductive problems. Our GAT models have achieved or matched state-of-the-art results across four established transductive and inductive graph benchmarks: the Cora, Citeseer and Pubmed citation network datasets, as well as a protein-protein interaction dataset (wherein test graphs remain unseen during training).

Battery degradation and behaviour for electric vehicles: Review and numerical analyses of several models

Abstract: The use of electric vehicles for goods distribution opens up a wide range of research problems. Battery electric vehicles (BEVs) operate on batteries that have a limited life, as well as specific charging and discharging patterns which need to be considered in the context of their use for goods distribution. While many transportation problems associated with the integration of freight electric vehicles in distribution management problems have been investigated, there is room for further research on specifically how to model battery degradation and behaviour in such problems. The aim of this paper is to provide tractable models for transportation scientists that will allow predicting the lifetime degradation and instantaneous charging and discharging behaviour of BEV batteries.

Effectiveness of mHealth interventions for patients with diabetes: An overview of systematic reviews

Abstract: Background Diabetes is a common chronic disease that places an unprecedented strain on health care systems worldwide. Mobile health technologies such as smartphones, mobile applications, and wearable devices, known as mHealth, offer significant and innovative opportunities for improving patient to provider communication and self-management of diabetes. Objective The purpose of this overview is to critically appraise and consolidate evidence from multiple systematic reviews on the effectiveness of mHealth interventions for patients with diabetes to inform policy makers, practitioners, and researchers. Methods A comprehensive search on multiple databases was performed to identify relevant systematic reviews published between January 1996 and December 2015. Two authors independently selected reviews, extracted data, and assessed the methodological quality of included reviews using AMSTAR. Results Fifteen systematic reviews published between 2008 and 2014 were eligible for inclusion. The quality of the reviews varied considerably and most of them had important methodological limitations. Focusing on systematic reviews that offered the most direct evidence, this overview demonstrates that on average, mHealth interventions improve glycemic control (HbA1c) compared to standard care or other non-mHealth approaches by as much as 0.8% for patients with type 2 diabetes and 0.3% for patients with type 1 diabetes, at least in the short-term (≤12 months). However, limitations in the overall quality of evidence suggest that further research will likely have an important impact in these estimates of effect. Conclusions Findings are consistent with clinically relevant improvements, particularly with respect to patients with type 2 diabetes. Similar to home telemonitoring, mHealth interventions represent a promising approach for self-management of diabetes.

Variable neighborhood search: basics and variants

Abstract: Variable neighborhood search (VNS) is a framework for building heuristics, based upon systematic changes of neighborhoods both in a descent phase, to find a local minimum, and in a perturbation phase to escape from the corresponding valley. In this paper, we present some of VNS basic schemes as well as several VNS variants deduced from these basic schemes. In addition, the paper includes parallel implementations and hybrids with other metaheuristics.

A Matter of Life or Death: How Extreme Context Research Matters for Management and Organization Studies

Abstract: Organization scholarship has seen an escalation of interest in research into extremes. Comprising several interconnected domains, this growing body of research is decidedly fragmented. This fragmen ...

Adversarial Network Embedding

Abstract: Learning low-dimensional representations of networks has proved effective in a variety of tasks such as node classification, link prediction and network visualization. Existing methods can effectively encode different structural properties into the representations, such as neighborhood connectivity patterns, global structural role similarities and other high-order proximities. However, except for objectives to capture network structural properties, most of them suffer from lack of additional constraints for enhancing the robustness of representations. In this paper, we aim to exploit the strengths of generative adversarial networks in capturing latent features, and investigate its contribution in learning stable and robust graph representations. Specifically, we propose an Adversarial Network Embedding (ANE) framework, which leverages the adversarial learning principle to regularize the representation learning. It consists of two components, i.e., a structure preserving component and an adversarial learning component. The former component aims to capture network structural properties, while the latter contributes to learning robust representations by matching the posterior distribution of the latent representations to given priors. As shown by the empirical results, our method is competitive with or superior to state-of-the-art approaches on benchmark network embedding tasks.

Underdog Entrepreneurs: A Model of Challenge–Based Entrepreneurship:

Abstract: Although there has been abundant research on the positive personality and environmental qualities that stimulate entrepreneurship, we argue that negative personal circumstances of an economic, sociocultural, cognitive, and physical/ emotional nature may have an equally powerful role to play in getting people to become effective entrepreneurs. These challenges create conditions and experiences that motivate particular adaptive requirements which in turn foster outcomes such as work discipline, risk tolerance, social and network skills, and creativity.

An Attention-based Collaboration Framework for Multi-View Network Representation Learning

Abstract: Learning distributed node representations in networks has been attracting increasing attention recently due to its effectiveness in a variety of applications. Existing approaches usually study networks with a single type of proximity between nodes, which defines a single view of a network. However, in reality there usually exists multiple types of proximities between nodes, yielding networks with multiple views. This paper studies learning node representations for networks with multiple views, which aims to infer robust node representations across different views. We propose a multi-view representation learning approach, which promotes the collaboration of different views and lets them vote for the robust representations. During the voting process, an attention mechanism is introduced, which enables each node to focus on the most informative views. Experimental results on real-world networks show that the proposed approach outperforms existing state-of-the-art approaches for network representation learning with a single view and other competitive approaches with multiple views.

Hidden in Plain Sight: The Importance of Scale in Organizations’ Attention to Issues

Abstract: The organizational attention literature has an epistemological bias, in that it explains how and why organizations notice issues. The ontological or real attributes of the issues are largely ignored, subordinated, or confounded with this epistemological orientation. In this article we argue that organizations sometimes miss issues, not only because of attentional failures but also because of the temporal and spatial scale of the underlying processes related to the issues. Some processes are of such large or small scale they escape organizational attention. We argue that large-scale processes, such as those related to climate change, require broad attentional extent, whereas small-scale processes, such as those related to local variations in poverty, require fine attentional grain. This work aims to shed light on the relatively underexplored question of why some issues are not noticed, with important implications for both theory and practice.

Adversarial Network Embedding

Abstract: Learning low-dimensional representations of networks has proved effective in a variety of tasks such as node classification, link prediction and network visualization. Existing methods can effectively encode different structural properties into the representations, such as neighborhood connectivity patterns, global structural role similarities and other high-order proximities. However, except for objectives to capture network structural properties, most of them suffer from lack of additional constraints for enhancing the robustness of representations. In this paper, we aim to exploit the strengths of generative adversarial networks in capturing latent features, and investigate its contribution in learning stable and robust graph representations. Specifically, we propose an Adversarial Network Embedding (ANE) framework, which leverages the adversarial learning principle to regularize the representation learning. It consists of two components, i.e., a structure preserving component and an adversarial learning component. The former component aims to capture network structural properties, while the latter contributes to learning robust representations by matching the posterior distribution of the latent representations to given priors. As shown by the empirical results, our method is competitive with or superior to state-of-the-art approaches on benchmark network embedding tasks.

End-Of-Life Medical Spending In Last Twelve Months Of Life Is Lower Than Previously Reported.

Abstract: Although end-of-life medical spending is often viewed as a major component of aggregate medical expenditure, accurate measures of this type of medical spending are scarce. We used detailed health care data for the period 2009–11 from Denmark, England, France, Germany, Japan, the Netherlands, Taiwan, the United States, and the Canadian province of Quebec to measure the composition and magnitude of medical spending in the three years before death. In all nine countries, medical spending at the end of life was high relative to spending at other ages. Spending during the last twelve months of life made up a modest share of aggregate spending, ranging from 8.5 percent in the United States to 11.2 percent in Taiwan, but spending in the last three calendar years of life reached 24.5 percent in Taiwan. This suggests that high aggregate medical spending is due not to last-ditch efforts to save lives but to spending on people with chronic conditions, which are associated with shorter life expectancies.

Transparency in literature reviews: an assessment of reporting practices across review types and genres in top IS journals

Abstract: The central role of information systems review articles has been recognised in a recent explosion of interest in editorials, research articles, and opinion papers investigating methods and ...

Explicating the role of innovation intermediaries in the “unknown”: a contingency approach

Abstract: Purpose – Innovation intermediaries have become key actors in open innovation (OI) contexts. Research has improved the understanding of the managerial challenges inherent to intermediation in situations in which problems are rather well defined. Yet, in some OI situations, the relevant actor networks may not be known, there may be no clear common interest, or severe problems may exist with no legitimate common place where they can be discussed. The purpose of this paper is to contribute to the research on innovation intermediaries by showing how intermediaries address managerial challenges related to a high degree of unknown. Design/methodology/approach – The authors draw upon the extant literature to highlight the common core functions of different types of intermediaries. The authors then introduce the " degree of unknown " as a new contingency variable for the analysis of the role of intermediaries for each of these core functions. The authors illustrate the importance of this new variable with four empirical case studies in different industries and countries in which intermediaries are experiencing situations of high level of unknown. Findings – The authors highlight the specific managerial principles that the four intermediaries applied in creating an environment for collective innovation. Originality/value – Thereby, the authors clarify what intermediation in the unknown may entail.

New Enhancements for the Exact Solution of the Vehicle Routing Problem with Time Windows

Abstract: The vehicle routing problem with time windows VRPTW consists of finding least-cost vehicle routes to satisfy the demands of customers that can be visited within specific time windows. We introduce two enhancements for the exact solution of the VRPTW by branch-price-and-cut BPC. First, we develop a sharper form of the limited-memory subset-row inequalities by representing the memory as an arc subset rather than a node subset. Second, from the elementary inequalities introduced by Balas in 1977, we derive a family of inequalities that dominate them. These enhancements are embedded into an exact BPC algorithm that includes state-of-the-art features such as bidirectional labeling, decremental state-space relaxation, completion bounds, variable fixing, and route enumeration. Computational results show that these enhancements are particularly effective for the most difficult instances and that our BPC algorithm can solve all 56 Solomon instances with 100 customers and 51 of 60 Gehring and Homberger instances with 200 customers. The online appendix is available at https://doi.org/10.1287/ijoc.2016.0744 .

Using eHealth Technologies: Interests, Preferences, and Concerns of Older Adults

Abstract: Background: The Internet and eHealth technologies represent new opportunities for managing health. Age, sex, socioeconomic status, and current technology use are some of the known factors that influence individuals’ uptake of eHealth; however, relatively little is known about facilitators and barriers to eHealth uptake specific to older adults, particularly as they relate to their experiences in accessing health care. Objective: The aim of our study was to explore the interests, preferences, and concerns of older adults in using the Internet and eHealth technologies for managing their health in relation to their experiences with the current health care system. Methods: Two focus groups (n=15) were conducted with adults aged 50+ years. Pragmatic thematic analysis using an inductive approach was conducted to identify the interests, preferences, and concerns of using the Internet and eHealth technologies. Results: Five themes emerged that include (1) Difficulty in identifying credible and relevant sources of information on the Web; (2) Ownership, access, and responsibility for medical information; (3) Peer communication and support; (4) Opportunities to enhance health care interactions; and (5) Privacy concerns. These findings support the potential value older adults perceive in eHealth technologies, particularly in their ability to provide access to personal health information and facilitate communication between providers and peers living with similar conditions. However, in order to foster acceptance, these technologies will need to provide personal and general health information that is secure, readily accessible, and easily understood. Conclusions: Older adults have diverse needs and preferences that, in part, are driven by their experiences and frustrations with the health care system. Results can help inform the design and implementation of technologies to address gaps in care and access to health information for older adults with chronic conditions who may benefit the most from this approach. [Interact J Med Res 2017;6(1):e3]

Growing Productivity Without Growing Wages: The Micro-Level Anatomy of the Aggregate Labor Share Decline

Abstract: The aggregate labor share in U.S. manufacturing declined dramatically over the last three decades: Since the mid-1980’s, the compensation for labor declined from 67% to 47% of value added which is unseen in any other sector of the U.S. economy. The labor share of the typical U.S. manufacturing plants, in contrast, rose by over 5 percentage points. We reconcile these two facts by documenting: (1) an important reallocation of production towards “hyper-productive plants” and, (2) a downward adjustment of the labor share of those same plants over time. These two related forces account for almost all the change in the trend of aggregate labor share in the manufacturing sector, with only a small role for exit of high-labor-share plants. Relative to their peers, plants that account for the majority of production by the late 2000's arrive at a low labor share by gradually increasing value added by a factor of three while keeping employment and compensation unchanged.

Tenure in Office and Public Procurement

Abstract: We study the impact of politicians' tenure in office on the outcomes of public procurement. To this purpose, we match a data set on the politics of Italian municipal governments to a data set on the procurement auctions they administered. In order to identify a causal relation, we apply two different identification strategies. First, we compare elections where the incumbent mayor barely won another term, with elections where the incumbent mayor barely lost and a new mayor took over. Second, we cross-validate these estimates using a unique quasi-experiment determined by the introduction of a two-term limit on the mayoral office in March 1993. This reform granted one potential extra term to mayors appointed before the reform. The main result is that an increase in the mayor's tenure is associated with ``worse'' outcomes: fewer bidders per auction, a higher cost of procurement, a higher probability that the winner is local and that the same firm is awarded repeated auctions. Taken together, our estimates are informative of the possibility that time in office progressively leads to collusion between government officials and a few favored local bidders. Other interpretations receive less support in the data

Capitalization: A Cultural Guide

Abstract: What does it mean to turn something into capital? What does considering things as assets entail? What does the prevalence of an investor’s viewpoint require? What is this culture of valuation that asks that we capitalize on everything? How can we make sense of the traits, necessities and upshots of this pervasive cultural condition? This book takes the reader to an ethnographic stroll down the trail of capitalization. Start-up companies, research centers, consulting firms, state enterprises, investment banks, public administrations: the territory can certainly prove strange and disorienting at first sight, with its blurred boundaries between private appropriation and public interest, economic sanity and moral breakdown, the literal and the metaphorical, the practical and the ideological. The traveler certainly requires a resolutely pragmatist attitude, and a taste for the meanders of signification. But in all the sites in which we set foot in this inquiry we recognize a recurring semiotic complex: a scenario of valuation in which things signify by virtue of their capacity to become assets in the eye of an imagined investor. A ground-breaking anthropological investigation on the culture of contemporary capitalism, this work directs attention to the largely unexplored problem of capitalization and offers a critical resource for current debates on neoliberalism and financialization.