# CRM: Retrieval Model with Controllable Condition

arXiv: https://arxiv.org/abs/2412.13844 · HTML: https://mlsanity.com/paper/2412.13844/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 21 references that are in the map.
- First posted on arXiv 18 Dec 2024.

## Builds on (most cited first, 14 of 14)

- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [Sequence to Sequence Learning with Neural Networks](https://mlsanity.com/paper/1409.3215/) (2014)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems](https://mlsanity.com/paper/2005.01643/) (2020)
- [Recommender Systems with Generative Retrieval](https://mlsanity.com/paper/2305.05065/) (2023)
- [Controllable Multi-Interest Framework for Recommendation](https://mlsanity.com/paper/2005.09347/) (2020)
- [Joint Optimization of Tree-based Index and Deep Model for Recommender Systems](https://mlsanity.com/paper/1902.07565/) (2019)
- [Large Scale Product Graph Construction for Recommendation in E-commerce](https://mlsanity.com/paper/2010.05525/) (2020)
- [DimeRec: A Unified Framework for Enhanced Sequential Recommendation via Generative Diffusion Models](https://mlsanity.com/paper/2408.12153/) (2024)
- [A Comprehensive Survey on Retrieval Methods in Recommender Systems](https://mlsanity.com/paper/2407.21022/) (2024)
- [HoME: Hierarchy of Multi-Gate Experts for Multi-Task Learning at Kuaishou](https://mlsanity.com/paper/2408.05430/) (2024)
- [Interest Clock: Time Perception in Real-Time Streaming Recommendation System](https://mlsanity.com/paper/2404.19357/) (2024)
- [Full Stage Learning to Rank: A Unified Framework for Multi-Stage Systems](https://mlsanity.com/paper/2405.04844/) (2024)
- [KuaiFormer: Transformer-Based Retrieval at Kuaishou](https://mlsanity.com/paper/2411.10057/) (2024)

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

- [PROMISE: Process Reward Models Unlock Test-Time Scaling Laws in Generative Recommendations](https://mlsanity.com/paper/2601.04674/) (2026)

## Abstract (arXiv)

Retrieval models for recommendation need to account for outcomes such as watch time while searching large item corpora efficiently. We present the Controllable Retrieval Model (CRM), an outcome-conditioned retrieval framework that adds watch time as an input to the user tower. During training, CRM learns to predict the next item given its observed watch time; at inference, we supply a desired watch time to steer retrieval without retraining. The item tower is independent of the condition, so item embeddings can still be precomputed and indexed for approximate nearest neighbor (ANN) search. We implement CRM as both a DNN-based two-tower model and a Decision Transformer. The Transformer interleaves item tokens with watch-time-to-go condition tokens, using the target watch time as a numerical prompt for next-item prediction. For each online request, we choose between the average and maximum watch times in the user's recent history as the condition. CRM is deployed in Kuaishou's short-video recommendation system, which serves over 400 million users. Online A/B tests show video watch time improvements of 0.372% in Kuaishou and 0.457% in Kuaishou Lite.