Sparse Personalized Text Generation with Multi-Trajectory Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ni, Bo, Fu, Haowei, Ge, Qinwen, Dernoncourt, Franck, Basu, Samyadeep, Lipka, Nedim, Yoon, Seunghyun, Wang, Yu, Ahmed, Nesreen K., Mukherjee, Subhojyoti, Mathur, Puneet, Rossi, Ryan A., Derr, Tyler
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908996987781120
author Ni, Bo
Fu, Haowei
Ge, Qinwen
Dernoncourt, Franck
Basu, Samyadeep
Lipka, Nedim
Yoon, Seunghyun
Wang, Yu
Ahmed, Nesreen K.
Mukherjee, Subhojyoti
Mathur, Puneet
Rossi, Ryan A.
Derr, Tyler
author_facet Ni, Bo
Fu, Haowei
Ge, Qinwen
Dernoncourt, Franck
Basu, Samyadeep
Lipka, Nedim
Yoon, Seunghyun
Wang, Yu
Ahmed, Nesreen K.
Mukherjee, Subhojyoti
Mathur, Puneet
Rossi, Ryan A.
Derr, Tyler
contents As Large Language Models (LLMs) advance, personalization has become a key mechanism for tailoring outputs to individual user needs. However, most existing methods rely heavily on dense interaction histories, making them ineffective in cold-start scenarios where such data is sparse or unavailable. While external signals (e.g., content of similar users) can offer a potential remedy, leveraging them effectively remains challenging: raw context is often noisy, and existing methods struggle to reason over heterogeneous data sources. To address these issues, we introduce PAT (Personalization with Aligned Trajectories), a reasoning framework for cold-start LLM personalization. PAT first retrieves information along two complementary trajectories: writing-style cues from stylistically similar users and topic-specific context from preference-aligned users. It then employs a reinforcement learning-based, iterative dual-reasoning mechanism that enables the LLM to jointly refine and integrate these signals. Experimental results across real-world personalization benchmarks show that PAT consistently improves generation quality and alignment under sparse-data conditions, establishing a strong solution to the cold-start personalization problem.
format Preprint
id arxiv_https___arxiv_org_abs_2604_24996
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Sparse Personalized Text Generation with Multi-Trajectory Reasoning
Ni, Bo
Fu, Haowei
Ge, Qinwen
Dernoncourt, Franck
Basu, Samyadeep
Lipka, Nedim
Yoon, Seunghyun
Wang, Yu
Ahmed, Nesreen K.
Mukherjee, Subhojyoti
Mathur, Puneet
Rossi, Ryan A.
Derr, Tyler
Artificial Intelligence
As Large Language Models (LLMs) advance, personalization has become a key mechanism for tailoring outputs to individual user needs. However, most existing methods rely heavily on dense interaction histories, making them ineffective in cold-start scenarios where such data is sparse or unavailable. While external signals (e.g., content of similar users) can offer a potential remedy, leveraging them effectively remains challenging: raw context is often noisy, and existing methods struggle to reason over heterogeneous data sources. To address these issues, we introduce PAT (Personalization with Aligned Trajectories), a reasoning framework for cold-start LLM personalization. PAT first retrieves information along two complementary trajectories: writing-style cues from stylistically similar users and topic-specific context from preference-aligned users. It then employs a reinforcement learning-based, iterative dual-reasoning mechanism that enables the LLM to jointly refine and integrate these signals. Experimental results across real-world personalization benchmarks show that PAT consistently improves generation quality and alignment under sparse-data conditions, establishing a strong solution to the cold-start personalization problem.
title Sparse Personalized Text Generation with Multi-Trajectory Reasoning
topic Artificial Intelligence
url https://arxiv.org/abs/2604.24996