Optimal Transport From Theory to Real-World: Insights From swissQuant Scientific Board Member

Share
At swissQuant, we're committed to harnessing the transformative power of mathematics in finance and technology. As we approach International Day of Mathematics on March 14, we're excited to highlight the profound impact of mathematical concepts through the insights of Professor Alessio Figalli, a distinguished member of our Scientific Advisory Board and Fields Medal Laureate.

Professor Figalli’s expertise in optimal transport not only highlights the historical significance of mathematical concepts, but also their modern applications that continue to revolutionize various industries. Read more about Prof. Figalli’s background and work at ETH Zurich below.

Understanding Optimal Transport

“Optimal transport is a very old topic; it’s about moving material from one place to another in the most efficient way possible,” explains Professor Figalli.

Key Takeaways:

  1. Historical Significance: Optimal transport dates back to the French Revolution, pioneered by mathematician Gaspard Monge, who sought efficient ways to move materials for military fortifications. 
  1. Economic and Mathematical Development: Initially viewed as an economic problem, optimal transport has evolved into a complex mathematical challenge, attracting interest for its depth and the variety of problems it can solve. 
  1. Practical Applications in Modern Technology: Optimal transport finds practical applications in artificial intelligence and machine learning, especially in image comparison and optimization problems. 
  1. Interdisciplinary Connections: The principle of optimal transport extends beyond mathematics, influencing fields such as fluid dynamics, probability theory, and atmospheric phenomena. 
  1. Future Potential: Its adaptability in machine learning algorithms promises more stable and efficient outcomes in optimization problems.

The Journey of Optimal Transport

From Historical Roots to Modern Applications

Optimal transport, initially a logistical strategy during the French Revolution, has transcended its military origins to become a cornerstone of modern computer science.


“It started during the French Revolution and then during the Napoleonic campaign, when Gaspard Monge, a French mathematician, wanted to understand the most efficient way to take materials, extract them, and then move them to places where they were needed to build fortifications,” reflects Professor Figalli.

This evolution from a tactical military tool to a multifaceted mathematical concept exemplifies the adaptive nature of mathematical theories.

A Quantum Leap: From Logistics to Economics

The 20th century ushered in a new chapter for optimal transport, as it transitioned from the battlefield to the economic arena. 

The visionary Russian mathematician Kantorovich reconceived optimal transport as an economic model, demonstrating how it could streamline the distribution of goods, thereby reducing costs.  

His pioneering work not only earned him the Nobel Prize in Economics but also expanded the horizons of optimal transport into new domains.

Exploring the Diverse Applications of Optimal Transport

Optimal transport, a field that has intrigued mathematicians for decades, is recognized for its broad applicability in various fields of mathematics and science. Its profound impact is evident in areas ranging from fluid dynamics to machine learning, showcasing the versatility of this mathematical concept.

  • Fluid Dynamics: The relationship between optimal transport and fluid dynamics proved to be a pivotal discovery. Techniques from optimal transport have been instrumental in studying the complex equations that govern fluid movement.

  • Probability and Differential Equations: Subsequent explorations revealed the relevance of optimal transport in probability theory and the study of partial differential equations, further expanding its utility across mathematical disciplines. 

Revolutionizing AI and Machine Learning

In today’s digital world, optimal transport plays an important role in artificial intelligence and machine learning, especially in image comparison.

Professor Figalli elaborates on this application: 

 “This means that I can think of optimal transport as a way to compare pictures because you give me two pictures made of pixels, and then I ask myself how expensive it is for me to transport all the pixels present in the first image to the pixels in the second image.”

This approach not only enriches our understanding of data comparison, but also advances the ability of AI technologies to recognize and process complex patterns, underscoring their utility in the rapidly evolving technological landscape.

Optimization: Applications in Natural Sciences & Engineering

In meteorology, the movement of clouds and atmospheric fronts is described by complex equations. Optimal transport theory underlies this process, as clouds, composed of tiny water particles, strive for efficiency in their movement through the air.

Clouds transition from one configuration to another over time, with particles being optimally transported from one point to another within the cloud. This hidden optimization process reflects nature’s tendency to seek optimal solutions in the movement of clouds.

In the technological realm, optimal transport has found a novel application as a loss function in machine learning algorithms. By facilitating the comparison of data distributions, it offers a fresh perspective on optimization challenges, promising improved stability and convergence of algorithmic solutions.

Envisioning the Future: The Infinite Potential of Mathematics

As we anticipate new discoveries, the journey of optimal transport continues to inspire innovation. Its evolution from a practical solution to a profound mathematical concept and a tool for modern technology reflects the transformative power of mathematics, promising endless possibilities in quantitative finance and beyond.

The journey of optimal transport from its origins to its role in cutting-edge technologies is a powerful reminder of the endless possibilities of mathematical exploration.

At swissQuant, we are continually inspired by the ways in which mathematical principles can be harnessed to innovate and solve real-world challenges.

Profile Prof. Dr. Alessio Figalli

Image of Professor Alessio Figalli, Fields Medal Winner for his contributions to the theory of optimal transport and its applications.

Scientific Board Member swissQuant

Alessio Figalli is a Professor in Mathematics at ETH Zurich, and the Director of the FIM (Institute for Mathematical Research) at ETH. After a joint Phd in Mathematics at SNS Pisa (Italy) and ENS Lyon (France) in 2007, he was a professor in France and the US, before moving to ETH Zurich in 2016. He works on several topics in mathematics, including Calculus of Variations, Optimal Transportation, and Partial Differential Equations. For his achievements, he was awarded the Fields Medal in 2018.

DO YOU WANT TO KNOW MORE?

Get to Know Us And Our Scientific Board

Our Fields Medalists Reflect on The Power of Mathematics

Table of Contents

Reading Progress

Master complexity, amplify impact:

Let's work together

MACHINE LEARNING PREDICTION

Download our SNB Case Study

The SNB Case Study will be sent to your email. Check your spam folder too.

THE FUTURE OF RETAIL BANKING

Download our LLB Case Study

The LLB Case Study will be sent to your email. Check your spam folder too.