# Item Level Exploration Traffic Allocation in Large-scale Recommendation Systems

arXiv: https://arxiv.org/abs/2505.09033 · HTML: https://mlsanity.com/paper/2505.09033/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 May 2025.

## Built on by (most cited first, 1 of 1)

- [Content Exploration Beyond the Feed: Creator Supply and the Shared Corpus](https://mlsanity.com/paper/2608.29430/) (2026)

## Abstract (arXiv)

This paper contributes to addressing the item cold start problem in large-scale recommender systems, focusing on how to efficiently gain initial visibility for newly ingested content. We propose an exploration system designed to efficiently allocate impressions to these fresh items. Our approach leverages a learned probabilistic model to predict an item's discoverability, which then informs a scalable and adaptive traffic allocation strategy. This system intelligently distributes exploration budgets, optimizing for the long-term benefit of the recommendation platform. The impact is a demonstrably more efficient cold-start process, leading to a significant increase in the discoverability of new content and ultimately enriching the item corpus available for exploitation, as evidenced by its successful deployment in a large-scale production environment.