The Spanish Society of Statistics and Operations Research (SEIO) and the BBVA Foundation are pleased to announce the 2023 Spanish Society of Statistics and Operations Research – BBVA Foundation Awards for particularly significant scientific contributions in Statistics and Operations Research.
The awards made under this call, to be bestowed annually at national level, recognize originality, innovation and contributions in Statistics and Operations Research, with the dual aim of encouraging research work in SEIO’s focus areas and its transmission to society.
Best methodological contribution in Statistics
Professors Eustasio del Barrio (Professor of Statistics and Operations Research at the University of Valladolid), Marc Hallin (Emeritus Professor of Mathematics at the Université Libre de Bruxelles, Belgium), Juan Cuesta-Albertos (Professor of Statistics and Operations Research at the University of Cantabria) and Carlos Matrán (Professor of Statistics and Operations Research at the University of Valladolid) receive the award for the best methodological contribution in Statistics for “Distribution and Quantile Functions, Ranks, and Signs in Dimension d: A Measure Transportation Approach,” published in Annals of Statistics. This paper puts forward a new technique to compare data obtained under different experimental conditions, a core aspect of the scientific method of cross-cutting utility in all areas of research.
“We could, for instance, compare peas produced under two different conditions, let’s say A and B, by reference to their diameters,” Del Barrio explains. “In his correspondence with Charles Darwin, Francis Galton proposed comparing the smallest type-A pea with the smallest of type B, then the second smallest of type A with the second smallest of type B, etc., and then counting how often type A came out below type B. This simple idea has surprisingly good statistical properties, but is constrained to measurements of just one magnitude, meaning comparisons are inevitably simplistic (peas can differ in many other measurable characteristics). Our work enables a correct comparison of samples where several magnitudes are being measured simultaneously.”
This innovative methodology is already being applied to the analysis of oceanographic data in Pacific waters, to assess the impact of overfishing, and for data “repair” in AI applications, to prevent algorithmic discrimination against certain minorities.
Best methodological contribution in Operations Research
The award for the best methodological contribution in Operations Research has gone to Roi Naveiro (Professor of Quantitative Methods at CUNEF University), Tahir Ekin (Steven R. Gregg Associate Professor of Quantitative Methods at Texas State University, United States), David Ríos Insua (Research Professor in the Institute of Mathematical Sciences, ICMAT, of the Spanish National Research Council) and Alberto Torres Barrán (Chief Technology Officer at Komorebi AI S.L.) for the paper “Augmented Probability Simulation Methods for Sequential Games,” published in the European Journal of Operational Research.
The article is framed in the novel field of adversarial machine learning. Machine learning or AI algorithms are used as a decision-making input in more and more areas of society, from health to finance, and are also a key component of driverless cars. But they are vulnerable to attacks with potentially severe consequences. Adversarial machine learning was developed to find ways to defend algorithms against this kind of hacking.
Its mathematical basis, game theory, has certain issues that get in the way of its application to adversarial machine learning. The award-winning paper proposes a decisive step towards solving these problems, based on adversarial risk analysis. In the view of the committee, it provides a “robust framework” which “constitutes an important contribution in adversarial risk analysis with strong applications in adversarial machine learning approaches.” For Naveiro, its “natural applications lie in the world of security and cyber security,” and it is precisely in this realm that the authors locate a case study illustrating the usefulness of their proposal. The idea going forward is to develop an algorithm that will aid firms in deciding which cyber security tools are right for them, setting their cost against the client’s vulnerability to attacks.
Best applied contribution in Statistics
Pablo Morales-Álvarez (Assistant Professor of Statistics and Operations Research at the University of Granada), Pablo Ruiz (data scientist at Chartboost), Scott Coughlin (computational specialist with Northwestern IT Research Computing Services at Northwestern University, United States), Rafael Molina (Professor of Computer Science and Artificial Intelligence at the University of Granada) and Aggelos K. Katsaggelos (Joseph Cummings Professor of Electrical and Computer Engineering at Northwestern University, United States) win the award for the Best applied contribution in Statistics for their paper “Scalable Variational Gaussian Processes for Crowdsourcing: Glitch Detection in LIGO,” which appeared in IEEE Transactions on Pattern Analysis and Machine Intelligence.
In this paper, described by the committee as “highly innovative” and “published in one of the leading journals in the field,” the authors apply a statistical method to the detection of gravitational waves, perturbations in the space-time fabric caused by violent events like the fusion of two black holes. The systematic detection of such phenomena – predicted by Einstein’s Theory of Relativity in 1915 and observed experimentally for the first time in 2015 –, is “a costly process, since detector technology is highly sensitive to different sources of noise,” Pablo Morales-Álvarez explains. In this research, he says, “we used a statistical technique known as Gaussian Processes to automatically classify the different noise signals obtained by the detector.”
Another feature of the project is that it is partly a product of citizen science, with “over 30,000 volunteers providing more than seven million labels for more than one million signals,” the scientist points out. “This labeling technique using non-experts, and the huge amount of data available, posed significant methodological challenges.”
Best applied contribution in Operations Research
Jessica Rodríguez-Pereira (Ramón y Cajal researcher at the Universitat Politècnica de Catalunya), Burcu Balçık (Professor of Industrial Engineering at the University of Ozyegin, Turkey), Marie-Ève Rancourt (Professor in the Department of Logistics and Operations Management at HEC Montréal, Canada) and Gilbert Laporte (Professor Emeritus in the Department of Decision Sciences at HEC Montréal, Canada) take the award for Best applied contribution in Operations Research for their paper “A Cost-Sharing Mechanism for Multi-Country Partnerships in Disaster Preparedness,” published in Production and Operations Management.
This study, which the committee describes as “an interesting application of operations research in humanitarian partnership,” focuses on multinational cooperation to aid in disaster preparedness in the Caribbean region by means of a joint emergency supplies network. “Given the growing frequency and severity of natural disasters throughout the world, collaborative practices can provide an efficient way for countries to share risks and resources. For this kind of collaboration to succeed, its cost must be equitably shared among members and be to the benefit of each,” says Jessica Rodríguez-Pereira, whose research team enjoyed the collaboration of the Caribbean Disaster Emergency Management Agency (CDEMA).
“Together we have developed a method that equitably allocates the collaboration budget among the participating countries,” she explains. “Our method takes its principles from the world of insurance, and factors in the characteristics of all member countries in terms of risk level and economic standing.” The award-winning mechanism makes it possible to derive and analyze various “fair and transparent” cost-sharing schemes, in the words of the researcher, which can be used to support collaboration and negotiation processes.
Best contribution in Statistics and Operations Research applied to Data Science and Big Data
Andre Groeger (Ramón y Cajal researcher at the Universitat Autònoma de Barcelona), Hannes Mueller (tenured researcher in the Institute for Economic Analysis of the Spanish National Research Council), Jonathan Hersh (Associate Professor of Economics and Management Science at Chapman University, California, United States), Andrea Matranga (Associate Professor of Business at Chapman University, California, United States) and Joan Serrat (Associate Professor of Computer Science at Universitat Autònoma de Barcelona) are recognized in the category Best contribution in Statistics and Operations Research applied to Data Science and Big Data for their paper “Monitoring War Destruction from Space Using Machine Learning,” published in Proceedings of the National Academy of Sciences.
The project in question marks a qualitative advance in the assessment of war destruction. Until now, analysis of building damage in war zones provided only sketchy and biased information, as the work was mostly done by hand. With this in mind, the authors propose a machine learning algorithm that compares satellite images obtained over time to generate data on war destruction of unprecedented accuracy, frequency and coverage.
“The innovation lies in the exploration of machine learning and statistical validation ideas to harness the power of ever-increasing availability of satellite images in the modern era,” said the committee in its citation. The award-winning algorithm was designed with images of the Syrian civil war between 2013 and 2017, and the researchers now aim to apply it to the war in Ukraine. “I like to think my work can contribute positively to the course of human development,” remarks team member Groeger.
The international committee has a membership proposed by SEIO and the BBVA Foundation. Chairing the committee on this occasion was Ana Paula Barbosa Póvoa, Professor of Operations and Logistics and Head of the Engineering and Management Department in the Instituto Superior Técnico at the University of Lisbon (Portugal). Its members were:
; Araceli Garín, Professor in Quantitative Methods for Economics and Business at the University of the Basque Country (Spain); Xuming He, H. C. Carver Professor of Statistics at the University of Michigan (United States); Martine Labbé, Professor in the Graphs and Mathematical Optimization Unit at the Université Libre de Bruxelles (Belgium); and Daniel Peña, Emeritus Professor of Statistics at the Universidad Carlos III de Madrid (UC3M) (Spain).