本地私有上下文与云端推理协同，提升个性化生成质量与速度
SpecSteer: Synergizing Local Context and Global Reasoning for Efficient Personalized Generation
- 用贝叶斯融合思想实现设备端与云端的智能协作
- 在不访问用户原始数据前提下验证逻辑正确性，过滤错误输出
- 适合注重隐私保护且需高质量个性化生成的场景
实现个性化智能面临核心困境：将用户历史传至中心化大模型存在隐私风险，而本地小模型缺乏高质量生成所需的推理能力。初步研究表明，仅靠本地增强无法可靠弥补这一差距。为此，我们提出SpecSteer，一种非对称协同推理框架，融合私有设备端上下文与云规模推理能力。SpecSteer将协作建模为贝叶斯知识融合，并复用推测解码作为分布式对齐协议，形成草稿-验证-恢复三阶段流程：设备端模型生成个性化序列；云端通过基于比例的机制验证，解耦推理验证与私有上下文，无需访问原始用户数据即可过滤逻辑缺陷；若被拒绝，则通过引导恢复注入本地意图完成修正。实验表明，SpecSteer有效弥合理论差距，实现更优个性化生成性能，同时相较标准基线提速2.36倍。
原文摘要 · Abstract (English)
Realizing personalized intelligence faces a core dilemma: sending user history to centralized large language models raises privacy concerns, while on-device small language models lack the reasoning capacity required for high-quality generation. Our pilot study shows that purely local enhancements remain insufficient to reliably bridge this gap. We therefore propose SpecSteer, an asymmetric collaborative inference framework that synergizes private on-device context with cloud-scale reasoning. SpecSteer casts collaboration as Bayesian knowledge fusion and repurposes speculative decoding as a distributed alignment protocol, yielding a Draft--Verify--Recover pipeline: the on-device model drafts personalized sequences; the cloud validates via a ratio-based mechanism that decouples reasoning verification from private context, filtering logical flaws without accessing raw user context; upon rejection, a steering recovery injects local intent during correction. Experiments demonstrate that SpecSteer successfully closes the reasoning gap and achieves superior personalized generation performance, while delivering a 2.36x speedup over standard baselines.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。