对比强化学习与大模型在健康干预中的效果，发现后者更受欢迎但未必更优。
Structured Exploration vs. Generative Flexibility: A Field Study Comparing Bandit and LLM Architectures for Personalised Health Behaviour Interventions
- 用上下文带兵算法和大模型生成不同健康提醒，比较其效果。
- 大模型生成的内容被用户认为更帮助，但带兵算法未带来额外提升。
- 用户更认可回应自身输入的反馈，而非单一技巧重复推送。
行为改变技术（BCTs）是数字健康干预的核心，但如何选择和有效传递仍具挑战。上下文带兵算法可基于统计优化选择BCT，而大语言模型（LLMs）则能生成灵活、情境敏感的信息。我们开展了一项为期4周的物理活动动机研究（N=54；9次术后访谈），比较了五种每日消息策略：随机模板、带兵算法+模板、纯大模型生成、混合带兵+大模型，以及结合用户历史交互的大模型生成。结果显示，基于大模型的方法显著更受用户欢迎，但在不同大模型条件间无显著差异。出人意料的是，带兵算法在BCT选择上的优化并未带来感知帮助性的提升。未经约束的大模型过度聚焦于单一BCT，而带兵系统则强制实现技术间的探索-利用平衡。量化与质性分析表明，对用户输入的上下文回应是感知帮助性的关键驱动因素。本文为反思性人工智能健康行为改变系统的设计提供了建议，应对结构化探索与生成自主之间的权衡。
原文摘要 · Abstract (English)
Behaviour Change Techniques (BCTs) are central to digital health interventions, yet selecting and delivering effective techniques remains challenging. Contextual bandits enable statistically grounded optimisation of BCT selection, while Large Language Models (LLMs) offer flexible, context-sensitive message generation. We conducted a 4-week study on physical activity motivation (N=54; 9 post-study interviews) that compared five daily messaging approaches: random templates, contextual bandit with templates, LLM generation, hybrid bandit+LLM, and LLM with interaction history. LLM-based approaches were rated substantially more helpful than templates, but no significant differences emerged among LLM conditions. Unexpectedly, bandit optimisation for BCTs selection yielded no additional perceived helpfulness compared with LLM-only approaches. Unconstrained LLMs focused heavily on a single BCT, whereas bandit systems enforced systematic exploration-exploitation across techniques. Quantitative and qualitative findings suggest contextual acknowledgement of user input drove perceived helpfulness. We contribute design suggestions for reflective AI health behaviour change systems that address a trade-off between structured exploration and generative autonomy.
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。