Generative Sequential Notification Optimization via Multi-Objective Decision Transformers Conference: Preprint. Under review.; ; CCS: Computing methodologies Reinforcement learning CCS: Computing methodologies Supervised learning by regression CCS: Information systems Recommender systems

Borja Ocejo, Ruofan Wang, Ke Liu, Rohit Patra, Haotian Shen, David Liu, Yiwen Yuan, Gokulraj Mohanasundaram, Fedor Borisyuk, Prakruthi Prabhakar · 2025

Read Generative Sequential Notification Optimization via Multi-Objective Decision Transformers Conference: Preprint. Under review.; ; CCS: Computing methodologies Reinforcement learning CCS: Computing methodologies Supervised learning by regression CCS: Information systems Recommender systems by Borja Ocejo, Ruofan Wang, Ke Liu, Rohit Patra, Haotian Shen, David Liu, Yiwen Yuan, Gokulraj Mohanasundaram, Fedor Borisyuk, Prakruthi Prabhakar (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-02458.txt.

Description

Borja Ocejo and nine co-authors wrote a 2025 applied machine learning paper on deciding when to send notifications in a large production system, LinkedIn's. Each message trades user value against fatigue, and choices form a long sequence rather than single steps.

Earlier offline reinforcement learning, such as Conservative Q-Learning, proved unstable and hard to reproduce at scale. The team reframes the learning task with Decision Transformers, which cast policy learning as supervised learning conditioned on target returns. Additions include a multi-reward design for ongoing tasks and quantile-based conditioning on those returns. A production system with circular sequence buffers supports near-real-time decisions. Online tests raised session activity over the older agent while keeping fatigue in check.

Catalog record

CREATORBorja Ocejo, Ruofan Wang, Ke Liu, Rohit Patra, Haotian Shen, David Liu, Yiwen Yuan, Gokulraj Mohanasundaram, Fedor Borisyuk, Prakruthi Prabhakar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-02458

Browse the full library · About the Tower of Babel