Prototype-Aligned Federated Soft-Prompts for Continual Web Personalization

TL;DR

This paper introduces ProtoFed-SP, a privacy-preserving, prompt-based framework for continual web personalization that effectively balances stability and plasticity, improving performance across multiple benchmarks.

Contribution

It presents a novel prototype-anchored, dual-timescale soft prompt method that enhances continual personalization while maintaining privacy and reducing forgetting.

Findings

ProtoFed-SP improves NDCG@10 by +2.9% and HR@10 by +2.0% over baselines.

It achieves notable gains on Amazon-Books, H&M, and Taobao datasets.

The method lowers forgetting and preserves accuracy under differential privacy budgets.

Abstract

Continual web personalization is essential for engagement, yet real-world non-stationarity and privacy constraints make it hard to adapt quickly without forgetting long-term preferences. We target this gap by seeking a privacy-conscious, parameter-efficient interface that controls stability-plasticity at the user/session level while tying user memory to a shared semantic prior. We propose ProtoFed-SP, a prompt-based framework that injects dual-timescale soft prompts into a frozen backbone: a fast, sparse short-term prompt tracks session intent, while a slow long-term prompt is anchored to a small server-side prototype library that is continually refreshed via differentially private federated aggregation. Queries are routed to Top-M prototypes to compose a personalized prompt. Across eight benchmarks, ProtoFed-SP improves NDCG@10 by +2.9% and HR@10 by +2.0% over the strongest baselines, with notable gains on Amazon-Books (+5.0% NDCG vs. INFER), H&M (+2.5% vs. Dual-LoRA), and Taobao (+2.2% vs. FedRAP). It also lowers forgetting (AF) and Steps-to-95% and preserves accuracy under practical DP budgets. Our contribution is a unifying, privacy-aware prompting interface with prototype anchoring that delivers robust continual personalization and offers a transparent, controllable mechanism to balance stability and plasticity in deployment.