arXiv：cs.LG· Dianjing Fan, Yao Li, Kyaw Hpone Myint, Dwipam Katariya, Alexandre G. R. Day, Pranab Mohanty, Giri Iyengar·· 5 小时前
从点击到意图：用 LLM 蒸馏分类体系构建跨平台会话嵌入，用于金融服务推荐
From Clicks to Intent: Cross-Platform Session Embeddings with LLM-Distilled Taxonomy for Financial Services Recommendations
AI 导读
该研究提出一个面向金融服务的双用途意图预测框架，用自监督 Transformer 将原始网页点击流编码为紧凑的会话嵌入，并通过基于 LLM 的分类体系生成与蒸馏流程产出可解释的意图标签。
正文
Abstract:Sequential user behavior modeling is widely adopted in industrial recommender systems; however, significant gaps remain in financial services, where pre-login web interactions and authenticated in-app experiences differ drastically. Specifically, pre-login web users typically explore new products, whereas logged-in app users focus on account servicing. Due to the challenge of cross-channel entity resolution (e.g., matching anonymous web sessions to authenticated mobile accounts), web-based intent signals remain underutilized for post-authentication personalization. Existing methods for capturing web-based intent are often ad-hoc and narrow, lacking the flexibility to support both quantitative downstream recommendations and qualitative understanding at scale. In this work, we propose a scalable and dual-purpose intent prediction framework for web-based interactions and demonstrate its applicability for personalization. Our approach transforms raw web clickstreams into two outputs: a self-supervised Transformer encodes multi-modal clickstreams into a compact session embedding, while an LLM-based taxonomy generation and distillation pipeline produces interpretable intent labels. Our system demonstrates that self-supervised clickstream representations combined with LLM-distilled taxonomies can jointly serve quantitative tasks and qualitative understanding in production: on the mobile homepage tile ranking task, the session embedding improves macro Recall@1 by 1.88% and reduces Log Loss by 13.38% over production baselines. On the user conversion prediction task, the embedding outperforms the LLM labels by 4.3% on micro F1, while the distillation layer delivers interpretable labels at ultra-low latency with only a 7% performance drop.
| Comments: | Dianjing Fan and Yao Li equally contributed to this work. 7 pages, 1 figure |
| Subjects: | Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2606.26277 [cs.IR] |
| (or arXiv:2606.26277v3 [cs.IR] for this version) | |
| https://doi.org/10.48550/arXiv.2606.26277 arXiv-issued DOI via DataCite |
Submission history
From: Kyaw Hpone Myint [view email]
[v1]
Wed, 24 Jun 2026 18:18:56 UTC (267 KB)
[v2]
Tue, 13 Oct 2026 17:32:30 UTC (216 KB)
[v3]
Wed, 14 Oct 2026 17:04:12 UTC (216 KB)
来源：arXiv：cs.LG · arxiv.org