{"slug": "truthful-online-preference-aggregation-for-llm-fine-tuning-in-mobile", "title": "Truthful Online Preference Aggregation for LLM Fine-Tuning in Mobile Crowdsourcing", "summary": "Researchers have developed a new online weighted aggregation mechanism for mobile crowdsourcing platforms that ensures truthful feedback from workers during LLM fine-tuning. The mechanism, which dynamically adjusts worker weights based on feedback accuracy, achieves sublinear regret of \\(\\mathcal{O}(\\sqrt{T})\\) over \\(T\\) time slots, outperforming existing methods that suffer linear regret. This advancement addresses the problem of strategic workers misreporting preferences to maximize influence or payment, improving the reliability of AI-generated content in mobile applications like navigation.", "body_md": "arXiv:2605.24052v1 Announce Type: new\nAbstract: To better serve users' demands in mobile applications (e.g., navigation), mobile crowdsourcing platforms can iteratively align large language model (LLM)-generated content (e.g., AI-generated traffic condition predictions) with human feedback collected from crowdsourcing workers (e.g., mobile users). However, workers may strategically misreport their online preference feedback to maximize their influence or payment. Existing pipelines in mobile crowdsourcing (e.g., EM-based weight estimation) fail to identify the most accurate worker in this online setting, resulting in a linear regret $\\mathcal{O}(T)$ over $T$ time slots. In this paper, we study truthful online preference aggregation for LLM fine-tuning in mobile crowdsourcing. We formulate a new dynamic Bayesian game to model the multi-agent online learning process between the platform and strategic mobile workers. We propose a novel online weighted aggregation mechanism that dynamically adjusts each worker's weight in the preference aggregation according to their feedback accuracy. We prove that our mechanism ensures truthful feedback from strategic workers and achieves a sublinear regret $\\mathcal{O}(\\sqrt{T})$ over $T$ time slots. We further extend our mechanism to a challenging scenario with limited worker feedback per time slot, still guaranteeing a sublinear regret $\\mathcal{O}(\\sqrt{T})$. Experiments on LLM fine-tuning with real-world datasets further demonstrate significant performance gains of our mechanisms over benchmark schemes.", "url": "https://wpnews.pro/news/truthful-online-preference-aggregation-for-llm-fine-tuning-in-mobile", "canonical_source": "https://arxiv.org/abs/2605.24052", "published_at": "2026-05-26 04:00:00+00:00", "updated_at": "2026-05-26 04:07:22.431754+00:00", "lang": "en", "topics": ["machine-learning", "large-language-models", "artificial-intelligence", "ai-research"], "entities": [], "alternates": {"html": "https://wpnews.pro/news/truthful-online-preference-aggregation-for-llm-fine-tuning-in-mobile", "markdown": "https://wpnews.pro/news/truthful-online-preference-aggregation-for-llm-fine-tuning-in-mobile.md", "text": "https://wpnews.pro/news/truthful-online-preference-aggregation-for-llm-fine-tuning-in-mobile.txt", "jsonld": "https://wpnews.pro/news/truthful-online-preference-aggregation-for-llm-fine-tuning-in-mobile.jsonld"}}