Saved in:
Bibliographic Details
Main Authors: Gao, Chen, Zhao, Zixin, Shao, Lv, Liu, Tong
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.11317
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910125964394496
author Gao, Chen
Zhao, Zixin
Shao, Lv
Liu, Tong
author_facet Gao, Chen
Zhao, Zixin
Shao, Lv
Liu, Tong
contents Click-Through Rate (CTR) prediction has long been dominated by discriminative paradigms that optimize local decision boundaries within candidate-specific subspaces. However, these models often fail to capture the global joint distribution and the continuous structural evolution of user intent across all-domain movelines. While generative approaches attempt to model global transition patterns, existing methods suffer from discretization-induced information collapse by remapping nuanced e-commerce signals into discrete linguistic or categorical spaces, failing to preserve the topological fidelity of interest trajectories. To overcome these limitations, we propose a novel generative pre-training paradigm that models user intent as a continuous evolutionary trajectory on a high-dimensional latent interest manifold, termed the Next Interest Flow (NIF). We introduce kinematic constraints to govern this flow: Interest Diversity is achieved via tangent space decomposition, while Evolution Velocity ensures trajectory smoothness through geodesic regularization. To bridge the objective mismatch between generative pre-training and discriminative fine-tuning, we propose a bidirectional alignment strategy to synchronize semantic spaces. Furthermore, we develop a Temporal Sequential Pairwise (TSP) mechanism to instill temporal causality within the discriminative framework. We present the All-domain Moveline Evolution Network (AMEN), a unified framework implementing this pipeline. Extensive experiments on a 6.7-billion instance industrial dataset and online A/B tests on Taobao validate AMEN's superiority, achieving +0.87pt AUC gain and +11.6\% CTCVR lift.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11317
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Next Interest Flow: A Generative Pre-training Paradigm for Recommender Systems by Modeling All-domain Movelines
Gao, Chen
Zhao, Zixin
Shao, Lv
Liu, Tong
Information Retrieval
Click-Through Rate (CTR) prediction has long been dominated by discriminative paradigms that optimize local decision boundaries within candidate-specific subspaces. However, these models often fail to capture the global joint distribution and the continuous structural evolution of user intent across all-domain movelines. While generative approaches attempt to model global transition patterns, existing methods suffer from discretization-induced information collapse by remapping nuanced e-commerce signals into discrete linguistic or categorical spaces, failing to preserve the topological fidelity of interest trajectories. To overcome these limitations, we propose a novel generative pre-training paradigm that models user intent as a continuous evolutionary trajectory on a high-dimensional latent interest manifold, termed the Next Interest Flow (NIF). We introduce kinematic constraints to govern this flow: Interest Diversity is achieved via tangent space decomposition, while Evolution Velocity ensures trajectory smoothness through geodesic regularization. To bridge the objective mismatch between generative pre-training and discriminative fine-tuning, we propose a bidirectional alignment strategy to synchronize semantic spaces. Furthermore, we develop a Temporal Sequential Pairwise (TSP) mechanism to instill temporal causality within the discriminative framework. We present the All-domain Moveline Evolution Network (AMEN), a unified framework implementing this pipeline. Extensive experiments on a 6.7-billion instance industrial dataset and online A/B tests on Taobao validate AMEN's superiority, achieving +0.87pt AUC gain and +11.6\% CTCVR lift.
title Next Interest Flow: A Generative Pre-training Paradigm for Recommender Systems by Modeling All-domain Movelines
topic Information Retrieval
url https://arxiv.org/abs/2510.11317