Summary of Disco: An End-to-end Bandit Framework For Personalised Discount Allocation, by Jason Shuo Zhang et al.
DISCO: An End-to-End Bandit Framework for Personalised Discount Allocation
by Jason Shuo Zhang, Benjamin Howson, Panayiota Savva, Eleanor Loh
First submitted to arxiv on: 10 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper introduces DISCO, an end-to-end contextual bandit framework for personalised discount code allocation at ASOS. The framework combines Thompson Sampling with integer programming to control operational costs. To address high-dimensional actions and preserve relationships between price and sales, radial basis functions are used to represent the continuous action space, in combination with context embeddings from a neural network. These features enable pooled learning across similar actions, high accuracy, and preservation of negative price elasticity. Offline analysis shows that DISCO improves its performance over time despite global constraints. The framework is then tested online through an A/B test, achieving a significant improvement of >1% in average basket value compared to legacy systems. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary DISCO is a new way for ASOS to decide which discounts to offer customers and when. This helps them save money while making sure customers are happy. The system uses special math to figure out what works best and makes adjustments as needed. It also tries to understand how prices affect sales, which can help make more accurate decisions. By testing this new way with real customers, the researchers found that it did better than the old system at getting people to buy more stuff. |
Keywords
* Artificial intelligence * Neural network
