[2603.01590] IDProxy: Cold-Start CTR Prediction for Ads and Recommendation at Xiaohongshu with Multimodal LLMs

Computer Science > Information Retrieval arXiv:2603.01590 (cs) [Submitted on 2 Mar 2026] Title:IDProxy: Cold-Start CTR Prediction for Ads and Recommendation at Xiaohongshu with Multimodal LLMs Authors:Yubin Zhang, Haiming Xu, Guillaume Salha-Galvan, Ruiyan Han, Feiyang Xiao, Yanhua Huang, Li Lin, Yang Luo, Yao Hu View a PDF of the paper titled IDProxy: Cold-Start CTR Prediction for Ads and Recommendation at Xiaohongshu with Multimodal LLMs, by Yubin Zhang and Haiming Xu and Guillaume Salha-Galvan and Ruiyan Han and Feiyang Xiao and Yanhua Huang and Li Lin and Yang Luo and Yao Hu View PDF HTML (experimental) Abstract:Click-through rate (CTR) models in advertising and recommendation systems rely heavily on item ID embeddings, which struggle in item cold-start settings. We present IDProxy, a solution that leverages multimodal large language models (MLLMs) to generate proxy embeddings from rich content signals, enabling effective CTR prediction for new items without usage data. These proxies are explicitly aligned with the existing ID embedding space and are optimized end-to-end under CTR objectives together with the ranking model, allowing seamless integration into existing large-scale ranking pipelines. Offline experiments and online A/B tests demonstrate the effectiveness of IDProxy, which has been successfully deployed in both Content Feed and Display Ads features of Xiaohongshu's Explore Feed, serving hundreds of millions of users daily. Subjects: Information Retrieval (c...