• Graduate Program
    • Why study Business Data Science?
    • Program outline
    • Courses
    • Course registration
    • Admissions
    • Facilities
  • Research
  • News
  • Events
    • Summer School
      • Summer school
      • Deep Learning
      • Econometrics and Data Science Methods for Business, Economics and Finance
      • Foundations of Machine Learning with Applications in Python
      • Networks in Micro- and Macroeconomics
      • Tinbergen Institute Summer School Program
    • Events Calendar
    • Events archive
    • Conference: Consumer Search and Markets
    • Tinbergen Institute Lectures
  • Summer School
  • Alumni
Home | News | Hong Deng, Bas Donkers and Dennis Fok received Amazon Research Award
News | February 10, 2023

Hong Deng, Bas Donkers and Dennis Fok received Amazon Research Award

Tinbergen Institute PhD student Hong Deng and faculty members Bas Donkers and Dennis Fok recently received an Amazon Research Award, for Deng's PhD project ‘Real-Time Personalization in Dynamic Environments’.

Hong Deng, Bas Donkers and Dennis Fok received Amazon Research Award

The award includes $20,000.00 in cash funding and an additional $20,000.00 in AWS credits. Awardees also have access to more than 300 Amazon public datasets, along with AWS AI/ML services and tools. The awarded project is funded under the Amazon Advertising call for proposals – spring 2022. See all award recipients on the website of Amazon. 

The abstract reads: Real-time personalization engines help select the best company action for each customer. Such actions can be product recommendations, but also concrete marketing actions such as (personalized) emails or online targeting. These engines are important enablers of customization in E-commerce. Yet, the development of such engines that can be applied in real-time is not trivial. It is especially challenging when there are many actions to optimize over, when relevant contextual information is available, or when the environment is dynamic: both the available actions and rewards change over time. We aim to develop an easy-to-implement personalization engine to support fast adaptive decisions-making in such settings. We formalize the personalization problem under the multi-armed bandit framework and propose a new contextual bandit algorithm based on the particle-filtering technique. Our method allows firms to introduce new personalized options, calibrate their impact using prior knowledge from historical data, rapidly update these prior beliefs as new observations arrive sequentially and learn adaptively with potential time trend. With an application to news-article recommendation, we show that the proposed method lifts click-through-rate by 60% compared to the random policy and is computationally efficient.


Photo: Hong Deng, Bas Donkers (left), and Dennis Fok (right).