Showing papers by "HEC Montréal published in 2019"

Understanding digital transformation: A review and a research agenda

Abstract: Extant literature has increased our understanding of specific aspects of digital transformation, however we lack a comprehensive portrait of its nature and implications. Through a review of 282 works, we inductively build a framework of digital transformation articulated across eight building blocks. Our framework foregrounds digital transformation as a process where digital technologies create disruptions triggering strategic responses from organizations that seek to alter their value creation paths while managing the structural changes and organizational barriers that affect the positive and negative outcomes of this process. Building on this framework, we elaborate a research agenda that proposes [1] examining the role of dynamic capabilities, and [2] accounting for ethical issues as important avenues for future strategic IS research on digital transformation.

AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks

Abstract: Click-through rate (CTR) prediction, which aims to predict the probability of a user clicking on an ad or an item, is critical to many online applications such as online advertising and recommender systems. The problem is very challenging since (1) the input features (e.g., the user id, user age, item id, item category) are usually sparse and high-dimensional, and (2) an effective prediction relies on high-order combinatorial features (a.k.a. cross features), which are very time-consuming to hand-craft by domain experts and are impossible to be enumerated. Therefore, there have been efforts in finding low-dimensional representations of the sparse and high-dimensional raw features and their meaningful combinations. In this paper, we propose an effective and efficient method called the AutoInt to automatically learn the high-order feature interactions of input features. Our proposed algorithm is very general, which can be applied to both numerical and categorical input features. Specifically, we map both the numerical and categorical features into the same low-dimensional space. Afterwards, a multi-head self-attentive neural network with residual connections is proposed to explicitly model the feature interactions in the low-dimensional space. With different layers of the multi-head self-attentive neural networks, different orders of feature combinations of input features can be modeled. The whole model can be efficiently fit on large-scale raw data in an end-to-end fashion. Experimental results on four real-world datasets show that our proposed approach not only outperforms existing state-of-the-art approaches for prediction but also offers good explainability. Code is available at: \urlhttps://github.com/DeepGraphLearning/RecommenderSystems.

InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximization

Abstract: This paper studies learning the representations of whole graphs in both unsupervised and semi-supervised scenarios. Graph-level representations are critical in a variety of real-world applications such as predicting the properties of molecules and community analysis in social networks. Traditional graph kernel based methods are simple, yet effective for obtaining fixed-length representations for graphs but they suffer from poor generalization due to hand-crafted designs. There are also some recent methods based on language models (e.g. graph2vec) but they tend to only consider certain substructures (e.g. subtrees) as graph representatives. Inspired by recent progress of unsupervised representation learning, in this paper we proposed a novel method called InfoGraph for learning graph-level representations. We maximize the mutual information between the graph-level representation and the representations of substructures of different scales (e.g., nodes, edges, triangles). By doing so, the graph-level representations encode aspects of the data that are shared across different scales of substructures. Furthermore, we further propose InfoGraph*, an extension of InfoGraph for semi-supervised scenarios. InfoGraph* maximizes the mutual information between unsupervised graph representations learned by InfoGraph and the representations learned by existing supervised methods. As a result, the supervised encoder learns from unlabeled data while preserving the latent semantic space favored by the current supervised task. Experimental results on the tasks of graph classification and molecular property prediction show that InfoGraph is superior to state-of-the-art baselines and InfoGraph* can achieve performance competitive with state-of-the-art semi-supervised models.

Session-Based Social Recommendation via Dynamic Graph Attention Networks

Abstract: Online communities such as Facebook and Twitter are enormously popular and have become an essential part of the daily life of many of their users. Through these platforms, users can discover and create information that others will then consume. In that context, recommending relevant information to users becomes critical for viability. However, recommendation in online communities is a challenging problem: 1) users' interests are dynamic, and 2) users are influenced by their friends. Moreover, the influencers may be context-dependent. That is, different friends may be relied upon for different topics. Modeling both signals is therefore essential for recommendations. We propose a recommender system for online communities based on a dynamic-graph-attention neural network. We model dynamic user behaviors with a recurrent neural network, and context-dependent social influence with a graph-attention neural network, which dynamically infers the influencers based on users' current interests. The whole model can be efficiently fit on large-scale data. Experimental results on several real-world data sets demonstrate the effectiveness of our proposed approach over several competitive baselines including state-of-the-art models.

KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation

Abstract: Pre-trained language representation models (PLMs) cannot well capture factual knowledge from text. In contrast, knowledge embedding (KE) methods can effectively represent the relational facts in knowledge graphs (KGs) with informative entity embeddings, but conventional KE models cannot take full advantage of the abundant textual information. In this paper, we propose a unified model for Knowledge Embedding and Pre-trained LanguagE Representation (KEPLER), which can not only better integrate factual knowledge into PLMs but also produce effective text-enhanced KE with the strong PLMs. In KEPLER, we encode textual entity descriptions with a PLM as their embeddings, and then jointly optimize the KE and language modeling objectives. Experimental results show that KEPLER achieves state-of-the-art performances on various NLP tasks, and also works remarkably well as an inductive KE model on KG link prediction. Furthermore, for pre-training and evaluating KEPLER, we construct Wikidata5M, a large-scale KG dataset with aligned entity descriptions, and benchmark state-of-the-art KE methods on it. It shall serve as a new KE benchmark and facilitate the research on large KG, inductive KE, and KG with text. The source code can be obtained from this https URL.

Online Continual Learning with Maximal Interfered Retrieval

Abstract: Continual learning, the setting where a learning agent is faced with a never-ending stream of data, continues to be a great challenge for modern machine learning systems. In particular the online or "single-pass through the data" setting has gained attention recently as a natural setting that is difficult to tackle. Methods based on replay, either generative or from a stored memory, have been shown to be effective approaches for continual learning, matching or exceeding the state of the art in a number of standard benchmarks. These approaches typically rely on randomly selecting samples from the replay memory or from a generative model, which is suboptimal. In this work, we consider a controlled sampling of memories for replay. We retrieve the samples which are most interfered, i.e. whose prediction will be most negatively impacted by the foreseen parameters update. We show a formulation for this sampling criterion in both the generative replay and the experience replay setting, producing consistent gains in performance and greatly reduced forgetting. We release an implementation of our method at https://github.com/optimass/Maximally_Interfered_Retrieval

Exact Combinatorial Optimization with Graph Convolutional Neural Networks

Abstract: Combinatorial optimization problems are typically tackled by the branch-and-bound paradigm. We propose a new graph convolutional neural network model for learning branch-and-bound variable selection policies, which leverages the natural variable-constraint bipartite graph representation of mixed-integer linear programs. We train our model via imitation learning from the strong branching expert rule, and demonstrate on a series of hard problems that our approach produces policies that improve upon state-of-the-art machine-learning methods for branching and generalize to instances significantly larger than seen during training. Moreover, we improve for the first time over expert-designed branching rules implemented in a state-of-the-art solver on large problems. Code for reproducing all the experiments can be found at https://github.com/ds4dm/learn2branch.

Boundary Work among Groups, Occupations, and Organizations: From Cartography to Process

Abstract: This article reviews scholarship dealing with the notion of “boundary work,” defined as purposeful individual and collective effort to influence the social, symbolic, material, or temporal boundari...

A concise guide to existing and emerging vehicle routing problem variants.

Abstract: Vehicle routing problems have been the focus of extensive research over the past sixty years, driven by their economic importance and their theoretical interest. The diversity of applications has motivated the study of a myriad of problem variants with different attributes. In this article, we provide a concise overview of existing and emerging problem variants. Models are typically refined along three lines: considering more relevant objectives and performance metrics, integrating vehicle routing evaluations with other tactical decisions, and capturing fine-grained yet essential aspects of modern supply chains. We organize the main problem attributes within this structured framework. We discuss recent research directions and pinpoint current shortcomings, recent successes, and emerging challenges.

Exact Branch-Price-and-Cut Algorithms for Vehicle Routing

Abstract: Vehicle routing problems (VRPs) are among the most studied problems in operations research. Nowadays, the leading exact algorithms for solving many classes of VRPs are branch-price-and-cut algorith...

The electric vehicle routing problem with energy consumption uncertainty

Abstract: Compared with conventional freight vehicles, electric freight vehicles create less local pollution and are thus generally perceived as a more sustainable means of goods distribution. In urban areas, such vehicles must often perform the entirety of their delivery routes without recharging. However, their energy consumption is subject to a fair amount of uncertainty, which is due to exogenous factors such as the weather and road conditions, endogenous factors such as driver behaviour, and several energy consumption parameters that are difficult to measure precisely. Hence we propose a robust optimization framework to take into account these energy consumption uncertainties in the context of an electric vehicle routing problem. The objective is to determine minimum cost delivery routes capable of providing strong guarantees that a given vehicle will not run out of charge during its route. We formulate the problem as a robust mixed integer linear program and solve small instances to optimality using robust optimization techniques. Furthermore, we develop a two-phase heuristic method based on large neighbourhood search to solve larger instances of the problem, and we conduct several numerical tests to assess the quality of the methodology. The computational experiments illustrate the trade-off between cost and risk, and demonstrate the influence of several parameters on best found solutions. Furthermore, our heuristic identifies 42 new best solutions when tested on instances of the closely related robust capacitated vehicle routing problem.

Vehicle Routing and Location Routing with Intermediate Stops: A Review

Abstract: This paper reviews the literature on vehicle routing problems and location routing problems with intermediate stops. We classify publications into different categories from both an application-base...

Probabilistic Logic Neural Networks for Reasoning

Abstract: Knowledge graph reasoning, which aims at predicting missing facts through reasoning with observed facts, is critical for many applications. Such a problem has been widely explored by traditional logic rule-based approaches and recent knowledge graph embedding methods. A principled logic rule-based approach is the Markov Logic Network (MLN), which is able to leverage domain knowledge with first-order logic and meanwhile handle uncertainty. However, the inference in MLNs is usually very difficult due to the complicated graph structures. Different from MLNs, knowledge graph embedding methods (e.g. TransE, DistMult) learn effective entity and relation embeddings for reasoning, which are much more effective and efficient. However, they are unable to leverage domain knowledge. In this paper, we propose the probabilistic Logic Neural Network (pLogicNet), which combines the advantages of both methods. A pLogicNet defines the joint distribution of all possible triplets by using a Markov logic network with first-order logic, which can be efficiently optimized with the variational EM algorithm. Specifically, in the E-step, a knowledge graph embedding model is used for inferring the missing triplets, while in the M-step, the weights of the logic rules are updated according to both the observed and predicted triplets. Experiments on multiple knowledge graphs prove the effectiveness of pLogicNet over many competitive baselines.

Empowering Women: The Role of Emancipative Forces in Board Gender Diversity

Abstract: This study investigates the effect of country-level emancipative forces on corporate gender diversity around the world. Based on Welzel’s (Freedom rising: human empowerment and the quest for emancipation. Cambridge University Press, New York, 2013) theory of emancipation, we develop an emancipatory framework of board gender diversity that explains how action resources, emancipative values and civic entitlements enable, motivate and encourage women to take leadership roles on corporate boards. Using a sample of 6390 firms operating in 30 countries around the world, our results show positive single and combined effects of the framework components on board gender diversity. Our research adds to the existing literature in a twofold manner. First, our integrated framework offers a more encompassing, complete and theoretically richer picture of the key drivers of board gender diversity. Second, by testing the framework empirically, we extend the evidence on national drivers of board gender diversity.

To What Extent Do Gender Diverse Boards Enhance Corporate Social Performance

Abstract: The inconclusiveness of previous research on the association between gender diverse boards (GDB) and corporate social performance (CSP) has led us to revisit the question in light of stakeholder management and institutional theories. Given that corporate social responsibility (CSR) is a multidimensional concept, we test the influence of GDB on various groups of stakeholders. By considering the interaction between stakeholders’ power and directors’ personal motivations toward the prioritization of stakeholders’ claims, we find that GDB are positively related to CSR dimensions that are related to less powerful stakeholders such as the environment, contractors, and the community. However, GDB do not appear to have a significant impact on CSR dimensions that are associated with stakeholders who benefit from more institutionalized power, such as employees and customers.

Session-based Social Recommendation via Dynamic Graph Attention Networks

Abstract: Online communities such as Facebook and Twitter are enormously popular and have become an essential part of the daily life of many of their users. Through these platforms, users can discover and create information that others will then consume. In that context, recommending relevant information to users becomes critical for viability. However, recommendation in online communities is a challenging problem: 1) users' interests are dynamic, and 2) users are influenced by their friends. Moreover, the influencers may be context-dependent. That is, different friends may be relied upon for different topics. Modeling both signals is therefore essential for recommendations. We propose a recommender system for online communities based on a dynamic-graph-attention neural network. We model dynamic user behaviors with a recurrent neural network, and context-dependent social influence with a graph-attention neural network, which dynamically infers the influencers based on users' current interests. The whole model can be efficiently fit on large-scale data. Experimental results on several real-world data sets demonstrate the effectiveness of our proposed approach over several competitive baselines including state-of-the-art models.

The electric bus fleet transition problem

Abstract: The use of electric bus fleets has become a topical issue in recent years. Several companies and municipalities, either voluntarily or to comply with legal requirements, will transition to greener bus fleets in the next decades. Such transitions are often established by fleet electrification targets, which dictate the number of electric buses that should be in the fleet by a given time period. In this paper we introduce a comprehensive optimization-based decision making tool to support such transitions. More precisely, we present a fleet replacement problem which allows organizations to determine bus replacement plans that will meet their fleet electrification targets in a cost-effective way, namely considering purchase costs, salvage revenues, operating costs, charging infrastructure investments, and demand charges. We account for several charging infrastructure options, such as slow and fast plug-in stations, overhead pantograph chargers, and inductive (wireless) chargers. We refer to this problem as the electric bus fleet transition problem, and we model it as an integer linear program. We apply our model to conduct computational experiments based on several scenarios. We use real data provided by a public transit agency in order to draw insights into optimal transition plans.

A Unified Decomposition Matheuristic for Assembly, Production, and Inventory Routing

Abstract: While the joint optimization of production and outbound distribution decisions in a manufacturing context have been intensively studied in the past decade, the integration of production, inventory,...

Multi-scale Information Diffusion Prediction with Reinforced Recurrent Networks.

Abstract: Information diffusion prediction is an important task which studies how information items spread among users. With the success of deep learning techniques, recurrent neural networks (RNNs) have shown their powerful capability in modeling information diffusion as sequential data. However, previous works focused on either microscopic diffusion prediction which aims at guessing the next influenced user or macroscopic diffusion prediction which estimates the total numbers of influenced users during the diffusion process. To the best of our knowledge, no previous works have suggested a unified model for both microscopic and macroscopic scales. In this paper, we propose a novel multi-scale diffusion prediction model based on reinforcement learning (RL). RL incorporates the macroscopic diffusion size information into the RNN-based microscopic diffusion model by addressing the non-differentiable problem. We also employ an effective structural context extraction strategy to utilize the underlying social graph information. Experimental results show that our proposed model outperforms state-of-the-art baseline models on both microscopic and macroscopic diffusion predictions on three real-world datasets.

Solving a large multi-product production-routing problem with delivery time windows

Abstract: Even though the joint optimization of sequential activities in supply chains is known to yield significant cost savings, the literature concerning optimization approaches that handle the real-life features of industrial problems is scant. The problem addressed in this work is inspired by industrial contexts where vendor-managed inventory policies are applied. In particular, our study is motivated by a meat producer whose supply chain comprises a single meat processing centre with several production lines and a fleet of vehicles that is used to deliver different products to meat stores spread across the country. A considerable set of characteristics, such as product family setups, perishable products, and delivery time windows, needs to be considered in order to obtain feasible integrated plans. However, the dimensions of the problem make it impossible to be solved exactly by current solution methods. We propose a novel three-phase methodology to tackle a large Production-Routing Problem (PRP) combining realistic features for the first time. In the first phase, we attempt to reduce the size of the original problem by simplifying some dimensions such as the number of products, locations and possible routes. In the second phase, an initial PRP solution is constructed through a problem decomposition comprising several inventory-routing problems and one lot-sizing problem. In the third phase, the initial solution is improved by different mixed-integer programming models which focus on small parts of the original problem and search for improvements in the production, inventory management and transportation costs. Our solution approach is tested both on simpler instances available in the literature and on real-world instances containing additional details, specifically developed for a European company’s case study. By considering an integrated approach, we achieve global cost savings of 21.73% compared to the company’s solution.

Integrated Railway Rapid Transit Network Design and Line Planning problem with maximum profit

Abstract: We solve the Integrated Network Design and Line Planning Problem in Railway Rapid Transit systems with the objective of maximizing the net profit over a planning horizon, in the presence of a competing transportation mode. Since the profitability of the designed network is closely related with passengers’ demand and line operation decisions, for a given demand, a transit assignment is required to compute the profit, calculating simultaneously the frequencies of lines and selecting the most convenient train units. The proposed iterative solving procedure is governed by an adaptive large neighborhood search metaheuristic which, at each iteration, calls a branch-and-cut algorithm implemented in Gurobi in order to solve the assignment and network operation problems. We provide an illustration on a real-size scenario.