Large language model method improves personalized headlines and clicks
Generate to Explore, Select to Exploit: Aligning LLM-based Headline Generation with Personalized Recommendation
Information RetrievalArtificial Intelligence
Summary
Showing the same headline for an online item doesn’t always fit what different people want to see. The authors created a two-step method using large language models to first suggest many varied headline options and then pick the best one based on user reactions. This helps capture more specific user interests instead of settling for a generic choice. When tested on a real platform with millions of users, this approach increased clicks and the time users spent looking at content.
What this means in practice
- •For e-commerce product teams: Improve product listings by dynamically generating personalized headlines that better attract different customer segments.$Commercial implications: Enables e-commerce platforms to offer tailored product headlines that increase clicks and sales, directly supporting revenue growth.
- •For news app developers: Create diverse headline options from AI and select top-performing ones in real time to boost reader attention and engagement.
Authors
Yi Chen, Rufeng Cheng, Qiang Xie, Tao Li
Abstract
In industrial recommendation feeds, presenting a static headline for an item often fails to satisfy the diverse, multimodal interests of the user population, particularly suppressing the needs of long-tail audiences. While Large Language Models (LLMs) have been integrated into recommendation for content understanding or ranking, directly optimizing them to output a single best headline typically leads to mode collapse---converging to generic patterns that satisfy average tastes but miss specific latent intents. To bridge this gap, we introduce GESE (Generate to Explore, Select to Exploit), a framework operating at the system's presentation layer that decouples personalization into generative exploration and selective exploitation. First, we treat the LLM as a probabilistic explorer, utilizing Group Sequence Policy Optimization (GSPO) with a hierarchical reward mechanism to generate a candidate set that maximizes the semantic coverage of potential user interests. Subsequently, a lightweight, real-time feedback-aware selector acts as the exploiter, identifying the optimal realization from the candidate pool based on instant contextual signals. Extensive deployment on a commercial platform with over 100 million daily active users demonstrates that GESE significantly outperforms state-of-the-art baselines, achieving a 2.57% lift in CTR and 0.87% in dwell time. These results validate that decoupling diversity-oriented generation from precision-oriented selection offers a robust blueprint for aligning generative AI with dynamic user utility.