# Don't Contrast the Impossible: Region-Constrained Batching for Contrastive User Modeling on a Local Community Platform

arXiv: https://arxiv.org/abs/2607.28971 · HTML: https://mlsanity.com/paper/2607.28971/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 27 references that are in the map.
- First posted on arXiv 31 Jul 2026.

## Builds on (most cited first, 20 of 20)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [SimCSE: Simple Contrastive Learning of Sentence Embeddings](https://mlsanity.com/paper/2104.08821/) (2021)
- [On Layer Normalization in the Transformer Architecture](https://mlsanity.com/paper/2002.04745/) (2020)
- [Contrastive Learning with Hard Negative Samples](https://mlsanity.com/paper/2010.04592/) (2020)
- [RocketQA: An Optimized Training Approach to Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2010.08191/) (2020)
- [Hard Negative Mixing for Contrastive Learning](https://mlsanity.com/paper/2010.01028/) (2020)
- [Debiased Contrastive Learning](https://mlsanity.com/paper/2007.00224/) (2020)
- [DCN V2: Improved Deep & Cross Network and Practical Lessons for Web-scale Learning to Rank Systems](https://mlsanity.com/paper/2008.13535/) (2020)
- [Recommendations as Treatments: Debiasing Learning and Evaluation](https://mlsanity.com/paper/1602.05352/) (2016)
- [Unbiased Recommender Learning from Missing-Not-At-Random Implicit Feedback](https://mlsanity.com/paper/1909.03601/) (2019)
- [PinnerFormer: Sequence Modeling for User Representation at Pinterest](https://mlsanity.com/paper/2205.04507/) (2022)
- [Training Sparse Mixture Of Experts Text Embedding Models](https://mlsanity.com/paper/2502.07972/) (2025)
- [Scaling Law for Recommendation Models: Towards General-purpose User Representations](https://mlsanity.com/paper/2111.11294/) (2021)
- [Cross-Batch Negative Sampling for Training Two-Tower Recommenders](https://mlsanity.com/paper/2110.15154/) (2021)
- [Scaling User Modeling: Large-scale Online User Representations for Ads Personalization in Meta](https://mlsanity.com/paper/2311.09544/) (2023)
- [Nomic Embed Vision: Expanding the Latent Space](https://mlsanity.com/paper/2406.18587/) (2024)
- [On the Theories Behind Hard Negative Sampling for Recommendation](https://mlsanity.com/paper/2302.03472/) (2023)
- [UserBERT: Contrastive User Model Pre-training](https://mlsanity.com/paper/2109.01274/) (2021)
- [uCTRL: Unbiased Contrastive Representation Learning via Alignment and Uniformity for Collaborative Filtering](https://mlsanity.com/paper/2305.12768/) (2023)

## Abstract (arXiv)

Contrastive learning is widely used for user modeling in large-scale recommender systems, where standard in-batch negatives implicitly assume universal exposure that any user can be shown any item. On local community platforms such as Karrot, however, exposure is geographically constrained; many user-item pairs are impossible by design yet still treated as negatives during training, diluting the contrastive learning signal. We address this impossible negatives problem and propose Region-Constrained Batch Sampling (RCBS), a simple yet effective batching method that constructs region-homogeneous mini-batches so that users are contrasted primarily against items they could feasibly see. By replacing impossible negatives with feasible ones, RCBS naturally introduces harder and more informative negatives under realistic exposure constraints. With offline evaluations and online A/B tests, we show that RCBS consistently improves user representation quality and consequently enhances home feed ranking, retrieval, and display ads ranking. The resulting user embeddings have been deployed in production across various applications.