I observed that the most popular restaurants and dishes on food‑delivery platforms such as Swiggy, Zomato, and DoorDash seem to be getting even more popular over time.
I hypothesize that this popularity amplification may be driven by "degenerating feedback loops" in the platforms’ machine‑learning driven recommendation systems.
- Popularity amplification observed on Swiggy, Zomato, DoorDash.
- Open question: Is there empirical evidence that popularity is increasing due to algorithmic feedback loops?
- Open question: How do these loops quantitatively affect recommendation diversity?
- Open question: What mitigations could reduce popularity bias while maintaining user satisfaction?
- Open question: Do dynamics differ between Swiggy, Zomato, and DoorDash given their distinct market contexts?
- "Degenerating feedback loops" refer to a situation where the system repeatedly reinforces already‑popular items, causing them to dominate recommendations further.
- The claim is speculative, intended to spark research or discussion about fairness and equity in recommendation algorithms.