American University of Beirut

Talks Abstracts

​​​​​​​

April 27-28, 2023

​Day 1 - April 27, 2023 | OSB Maamari Auditorium 

​Track 1: Data Science for Disaster Planning 

  • Priya Singh On Beyond crisis response: How WFP Supply Chain is using data and analytics to drive impactful change.

    Talk abstract: 
    The World Food Programme is the largest humanitarian agency in the world, working in over 120 countries, bringing life-saving food to people displaced by conflict, and made destitute by disasters, and helping individuals and communities find life-changing solutions to the multiple challenges they face in building better futures. 
    As the global context has evolved, the Supply Chain division has been redefining its role in WFP’s dual mandate of saving and changing lives.  Central to this is the use of research and evidence to dispel the notion that WFP buys food internationally and distributes in developing and fragile countries, to the detriment of local economies.    

    In 2022, WFP’s purchasing has in markets (4.2 million metric tons of food,50% from local and regional markets, 123,000 metric tons from smallholder farmers).
    Recognizing the power in helping to build and sustain livelihoods for people and communities that we interact with, this talk will focus on how the Supply Division is embedding analytics throughout its operational value chain, leveraging innovation and partnerships to bring greater efficiency to our work across the globe.​

  • Joline Uichanco On Using data for disaster preparedness in the Pacific typhoon belt: A case study of the Philippines

    T
    alk abstract: The importance of prepositioning was a hard lesson learned from Super Typhoon Haiyan that devastated the Philippines in 2013, when many affected by the typhoon did not have immediate access to food and water. In a typhoon-prone country, it is important to build resilience through an effective prepositioning model. In this talk, I will discuss how I engaged with the Philippine government and used data to develop optimization models to effectively pre-position relief items before a typhoon.

  • Hiba Baroud On Breaking Down Silos: The Interdisciplinary Power of Data Science in Disaster Resilience

    T
    alk abstract: 
    ​- Disaster planning is complex and requires the integration of multiple disciplines and input from diverse stakeholders.
    - Communication and collaboration are key to achieve effective disaster response and require that critical information be available and shared in an organized and timely manner.
    - Predicting the potential outcome of disasters provides valuable insights to inform decisions which hinges on the integration of data from diverse sources with varying levels of accessibility, accuracy, and reliability.

    - A
    dvances in data science have enabled interdisciplinary research that can help improve disaster resilience and achieve coordinated crisis management.
    - Case studies will be featured demonstrating how data science methods are used to combine information from different sources on the built, natural, and social systems and inform decision-making under uncertainty.

  • ​​​Suzana Maranhão Moreno On SKAI - Innovative technical framework for damaging assessment.

    ​Talk abstract:
    This talk will be about SKAI, an open-source project from WFP and Google Research that uses Artificial Intelligence (AI) and satellite imagery to reduce the amount of time needed to understand the impact of disasters.

  • Yessika Montiel On Incorporating real-time data analysis and predictive modeling into crisis management plans for Amazon Logistics

    Talk abstract: ​Data Analysis and developing predictive modeling can help Amazon and any company in last mile logistics industry to action plans for increasing capacity during a crisis, such as a pandemic. By understanding data and setting realistic objectives, Amazon can optimize their delivery network, identify potential bottlenecks and disruptions, and prioritize deliveries based on urgency. Data analysis plays a critical role in crisis management for helping logistics companies navigate challenging situations efficiently.​

​Track 2​​: Data Science for Public Policy 

  • ​Rima Turk Ariss On Using Data Analytics to Support Economic and Financial Policies and Infom Policies

    Talk abstract: The presentation will provide an overview of how the IMF uses data to assess the economic and financial health of its membership, analyze trends, and inform policies. We will discuss how the IMF uses data for the analysis of macroeconomic and financial sector issues to inform policies. We will also shed light on some elements of the IMF’s risk assessment framework for country vulnerabilities.

  • Carole Al Sharabaty On Transforming the Public Sector with Technology and Artificial Intelligence.

    Talk abstract: The speaker will explore how technology and data-driven decision-making can propel public sector reform​. She will share real-world examples of how data modeling has transformed administrations and institutions in the security, governance, and media sectors in Lebanon. Through these case studies, she will demonstrate how technology and data modeling can help to foster a new social contract, one that prioritizes accountability and transparency in the public sector.

​Track 3​​: Data Science for Smart Cities​

  • Elsa Arcaute On Socio-hydrological resilience in Mexico City.

    Talk abstract: Water security is a critical problem that many cities are facing today, and unfortunately, many will face in the future. The problem is highly complex, bringing together aspects related to water scarcity, ageing infrastructure, and water management to name a few, and it is being further exacerbated by climate change and population growth. The more vulnerable communities are commonly the most affected by these problems. In this talk, we present a framework to assess the socio-hydrological resilience of cities, and look at mitigation strategies for the particular case of Mexico City. 
  • Polly Hudson On The Colouring Cities Research Programme- developing  a global network of open data platforms on building stocks.

    Talk abstract: The Colouring Cities Research Programme, managed by the Alan Turing Institute, develops open-source code for knowledge sharing mapping platforms that provide open spatial data on the composition, performance and dynamics of building stocks. The aim is to create open tools that help improve stock quality, sustainability, efficiency and resilience, in line with the United Nations’ New Urban Agenda & its Sustainable Development Goals, and that increase insights into stocks as complex dynamic systems. This talk describes how a global network of Colouring Cities platforms is currently being developed by a consortium of academic institutions.

  • Eileen Martin On How Data Science is Transforming Urban Geophysics

    Talk abstract: Geophysicists map the properties of the Earth's subsurface, in part, to provide estimates of how much the ground will shake during any potential future earthquake scenarios. Geophysicists gather data indirectly: by using many sensors to record data on seismic waves propagating through the Earth, then carrying out high-dimensional numerical optimization to image (in other words, to map estimates of) the material properties of the ground. Today, machine learning methods are being integrated into geophysical imaging. In recent years, we have had access to an exciting new data source that allows us to use our fiber optic telecommunications networks to collect more seismic data close to the urban areas that need improved earthquake hazard tools. However, these data are collected at such a fast rate and with such high resolution that humans cannot inspect it manually. This talk will discuss several examples of how we are using machine learning techniques to understand and assist humans in processing these big data, and how we are using machine learning to ensure people's privacy is protected in areas where these data are collected. 

​Track 4​​: Data Science for Social Good​

  • ​Elodie Ngoc-thy and Yara El Moussaoui On Putting Social Listening Data into Action: The Role of the Africa Infodemic Response Alliance (AIRA) in Fighting Infodemics in Africa.

    Talk abstract:
     The rise of infodemics, as defined as the overabundance of information during a health crisis, has become a significant challenge to public health interventions in Africa, with dis/misinformation spreading faster than ever before in the age of social media. In Africa, the challenges posed by infodemics are further compounded by a complex context characterized by frequent disease outbreaks, such as Ebola, cholera, and yellow fever, which have all been exacerbated by the spread of dis/misinformation, lower vaccination rates for both COVID-19 and routine immunizations compared to other continents, increasing social media penetration with more and more people accessing information through social media platforms leading to the proliferation of dis/misinformation, a diverse and often challenging media landscape with a range of diverse and often competing voices and perspectives, and ongoing conflicts in some regions. The Africa Infodemic Response Alliance (AIRA) is a WHO-hosted network that was formed to combat this growing threat by bringing together diverse stakeholders from across the continent to coordinate a comprehensive response. One of the critical tools that AIRA uses to combat infodemics is social listening, which involves analyzing online and offline social data to identify trends in public opinion, prioritize the most concerning trends and provide operational recommendations to address them. In this talk, we will explore how AIRA has put social listening data into action to fight infodemics in Africa. We will also examine examples that illustrate how social listening has been used to shape communication campaigns, strengthen governmental systems, and support emergency humanitarian responses in different countries in Africa. Finally, we will discuss the challenges and successes of social listening as a tool for infodemic response, and explore opportunities for future research and collaboration in this critical area.

  • ​Brigitte Khair-Mountain On How can Data Science revolutionize humanitarian crises response and sustainable human development?

    Talk abstract: A historic overview of the use of data in humanitarian and development response around the world. What has worked and what hasn’t, how fast were the UN and the humanitarian agencies able to play catch up with the exponential advances in technology and data science? What can be done and how? What are the bureaucratic hurdles within the international aid system delaying the good use of this tool in prospecting, contingency planning, responding, and evaluating global response strategies.

  • ​Angela Alzir Assi & Samia Melhem On Leveraging Data Science to Inform Programs and Policy Design

    Talk abstract: Leveraging big data for the analysis of skills is now more important than ever. In this fast pace changing world of work, the set of skills required by employers has been constantly changing, and will continue to do so. This session will show the importance of leveraging such data, to inform policy recommendation and dialogue as well as program design.

​Track 5​​: Data Science in Healthcare

  • ​Caroline Buckee On Data Science in Crisis: Leveraging Data for Humanitarian Disaster Response
  • Francesca Dominici On Data Science for Cleaner Air: Leveraging machine learning and causal inference to tackle air pollution and climate disasters​

Day 2 - April 28, 2023 | OSB Maamari Auditorium​

  • Recruitment Information Session by Joline Uichanco to promote the PhD program at University of Michigan.
  • Data Science Research Talks, Hiba Baroud, On Who Contributes to Disaster Preparedness? Predicting Decision Making in Social Dilemmas of Disaster Resilience.

    Talk abstract: While disasters are a pervasive threat to society, people and critical infrastructure play a significant role in shaping the risks associated with these events. For instance, investing in public infrastructure development (e.g., sea walls, public hospitals, early warning systems) strengthens disaster resilience. However, these projects are typically viewed as public goods which engenders collective action problems associated with their funding. In other words, individuals will either help by investing in public infrastructure projects that benefit the entire community or be tempted to reap benefits of resilient infrastructure development without giving up their resources (i.e., “free ride”). Who is investing in these projects and who is a free rider? How can the actions of individual decision makers ripple through complex systems and impact disaster resilience? To answer these questions, a data-driven approach is developed that integrates game theory, statistical learning, agent-based models, Bayesian methods, and economic models to predict the responses of individuals participating in common-pool resource games and use the outcome to encode data-driven behaviors in agent-based models that simulate disaster scenarios. To overcome the complexity of parametrizing simulation models and the lack of data on historical disasters, Approximate Bayesian Computation is used to calibrate models and quantify parameter uncertainties. This talk will demonstrate how collective decision-making is enhanced by utilizing the proposed interdisciplinary approach using theoretical and real use cases that analyze the economic risks of supply chain disruptions and the return on investment in resilient port infrastructure.
  • Joline Uichanco On Overcoming the profitability problem of e-commerce retail using data-driven product framing

    Talk abstract:
    Commerce retail is known to suffer from thin profit margins. Using the data from a major U.S. retailer, we show that jointly planning product framing and order fulfillment can have a significant impact on online retailers’ profitability. We study a joint product framing and order fulfillment problem. In each period, the retailer needs to decide how to “frame” (i.e., display, rank, price) each product on her website as well as how to fulfill a new demand. We use techniques such as randomized algorithms and graph-based algorithms to provide a tractable solution heuristic. Our proposed policy significantly reduces shipping costs by using product framing to manage demand so that it occurs close to the location of the inventory.

  • Elsa Arcaute On Constructing spatial communities for interventions

    Talk abstract: 
    Innovations spread more easily within communities, nevertheless, these are unknown most of the time. In this talk we will look at different ways to create proxies using the principle of homophily and taking into consideration the spatial embeddedness of the networks. 

  • Eileen Martin On New Algorithmic Approaches for Large-scale Seismology

    Talk abstract: The field of seismology has undergone two major advances in the past decade: (i) the widespread adoption of ambient seismic noise interferometry techniques that extract signals from long recordings of randomly distributed seismic vibrations, and (ii) rapid advances in low-cost sensor technologies that enable scientists to collect data from orders of magnitude more sensors than ever before. In particular, for environmental and safety applications (e.g. permafrost thaw monitoring, mine safety monitoring, or earthquake hazard mapping), these techniques have enabled seismology to become an affordable technique for monitoring the Earth's surface and subsurface properties. This talk will explore several new algorithmic strategies which have enabled scientists to greatly reduce the computational cost and time of processing these large-scale ambient seismic noise data, with the algorithm designs focused on reducing communication. These new techniques have allowed us to enable more interactive workflows for seismologists, and have reduced the size of some computations from cluster-scale to laptop-scale.  ​

  • Karianne Bergen, PhD (virtual); Assistant Professor of Earth, Environmental, & Planetary Sciences; Assistant Professor of Computer Science; Data Science Initiative; Department of Earth, Environmental, & Planetary Sciences Brown University, RI, USA

  • Ghina AlAtat ​On Double Sided Platforms with Stochastic Supply and Demand.

    Talk abstract: ​Double sided platforms offer convenience for customers and flexibility for workers. For instance, in the context of ride hailing the convenience translates in customers requesting rides anytime from any location to any destination (within some boundaries). Moreover, drivers are flexible to decide when to be available to accept requests and when to be unavailable. Therefore, by design, supply and demand are stochastic. We model these features following a queueing theoretic approach. In the single class case, the problem can be viewed as a typical queueing system with a stochastically changing number of available servers. We analyze the stability of such system and obtain diffusion approximations to quantify its performance. We discuss how our results can be generalized to the context of a closed queueing network.

Contact Us

For various questions, please try contacting us via social media first!
read more

Privacy Statement

We take data privacy seriously and adhere to all applicable data privacy laws and regulations.
read more

Copyright and Disclaimer

Written permission is needed to copy or disseminate all or part of the materials on the AUB website.
read more

Title IX, Non-Discrimination, and Anti-Discriminatory Harassment

AUB is committed to providing a safe, respectful, and inclusive environment to all members of its community.
read more