# Generalized User Representations for Transfer Learning

arXiv: https://arxiv.org/abs/2403.00584 · HTML: https://mlsanity.com/paper/2403.00584/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 43 references that are in the map.
- First posted on arXiv 1 Mar 2024.

## Builds on (most cited first, 8 of 8)

- [Variational Autoencoders for Collaborative Filtering](https://mlsanity.com/paper/1802.05814/) (2018)
- [Contrastive Learning for Cold-Start Recommendation](https://mlsanity.com/paper/2107.05315/) (2021)
- [Current Challenges and Visions in Music Recommender Systems Research](https://mlsanity.com/paper/1710.03208/) (2017)
- [Sequential Variational Autoencoders for Collaborative Filtering](https://mlsanity.com/paper/1811.09975/) (2018)
- [One Person, One Model, One World: Learning Continual User Representation without Forgetting](https://mlsanity.com/paper/2009.13724/) (2020)
- [Metadata Embeddings for User and Item Cold-start Recommendations](https://mlsanity.com/paper/1507.08439/) (2015)
- [Exploring Adapter-based Transfer Learning for Recommender Systems: Empirical Studies and Practical Insights](https://mlsanity.com/paper/2305.15036/) (2023)
- [Hybrid Recommender Systems: A Systematic Literature Review](https://mlsanity.com/paper/1901.03888/) (2019)

## Built on by (most cited first, 2 of 2)

- [Embedding-to-Prefix: Parameter-Efficient Personalization for Pre-Trained Large Language Models](https://mlsanity.com/paper/2505.17051/) (2025)
- [REPREC: Representation Driven Parameter-Efficient Recommendation System](https://mlsanity.com/paper/2607.24845/) (2026)

## Abstract (arXiv)

We present a novel framework for user representation in large-scale recommender systems, aiming at effectively representing diverse user taste in a generalized manner. Our approach employs a two-stage methodology combining representation learning and transfer learning. The representation learning model uses an autoencoder that compresses various user features into a representation space. In the second stage, downstream task-specific models leverage user representations via transfer learning instead of curating user features individually. We further augment this methodology on the representation's input features to increase flexibility and enable reaction to user events, including new user experiences, in Near-Real Time. Additionally, we propose a novel solution to manage deployment of this framework in production models, allowing downstream models to work independently. We validate the performance of our framework through rigorous offline and online experiments within a large-scale system, showcasing its remarkable efficacy across multiple evaluation tasks. Finally, we show how the proposed framework can significantly reduce infrastructure costs compared to alternative approaches.