眼睛将闭合：一种基于视觉地图反馈强化学习的视觉下一位置预测模型
Eyes Will Shut: A Vision-Based Next GPS Location Prediction Model by Reinforcement Learning from Visual Map Feed Back
AI总结：
针对现有下一位置预测模型缺乏人类式地图推理的问题，提出基于视觉地图反馈强化学习的VLMLocPredictor，通过两阶段训练实现SOTA性能与跨城市泛化。
摘要：
下一位置预测是人类移动性研究中的一项基础任务，在交通规划、城市治理和流行病预测等领域有着广泛的应用。在实践中，当人类试图预测轨迹中的下一个位置时，他们通常会在脑海中将轨迹可视化在地图上，并基于道路连通性和移动趋势进行推理。然而，现有的大多数下一位置预测模型并不像人类那样在地图上进行推理。幸运的是，近期视觉语言模型（VLMs）的发展已展现出强大的视觉感知甚至视觉推理能力。这开辟了一种新的可能性：通过将道路网络和轨迹渲染到图像上，并利用VLM的推理能力，我们可以使模型以类人的方式进行轨迹推理。为了探索这一想法，我们首先提出了一种名为视觉引导位置搜索（VGLS）的方法，该方法在不修改任何内部参数的情况下评估通用VLM是否具备基于轨迹的推理能力。基于VGLS结果的洞察，我们进一步提出了我们的主要方法：VLMLocPredictor，它由两个阶段组成：在第一阶段，我们设计了两个监督微调（SFT）任务，帮助VLM理解道路网络和轨迹结构，并在此类视觉输入上获得基本推理能力。在第二阶段，我们引入了基于视觉地图反馈的强化学习，使模型能够通过与环境的交互来自我提升其下一位置预测能力。在来自四个不同城市的数据集上进行的实验表明，我们的方法达到了最先进的（SOTA）性能，并且与其他基于LLM的方法相比，展现出优越的跨城市泛化能力。
英文摘要：
Next Location Prediction is a fundamental task in the study of human mobility, with wide-ranging applications in transportation planning, urban governance, and epidemic forecasting. In practice, when humans attempt to predict the next location in a trajectory, they often visualize the trajectory on a map and reason based on road connectivity and movement trends. However, the vast majority of existing next-location prediction models do not reason over maps \textbf{in the way that humans do}. Fortunately, the recent development of Vision-Language Models (VLMs) has demonstrated strong capabilities in visual perception and even visual reasoning. This opens up a new possibility: by rendering both the road network and trajectory onto an image and leveraging the reasoning abilities of VLMs, we can enable models to perform trajectory inference in a human-like manner. To explore this idea, we first propose a method called Vision-Guided Location Search (VGLS), which evaluates whether a general-purpose VLM is capable of trajectory-based reasoning without modifying any of its internal parameters. Based on insights from the VGLS results, we further propose our main approach: VLMLocPredictor, which is composed of two stages: In the first stage, we design two Supervised Fine-Tuning (SFT) tasks that help the VLM understand road network and trajectory structures and acquire basic reasoning ability on such visual inputs. In the second stage, we introduce Reinforcement Learning from Visual Map Feedback, enabling the model to self-improve its next-location prediction ability through interaction with the environment. Experiments conducted on datasets from four different cities show that our method achieves state-of-the-art (SOTA) performance and exhibits superior cross-city generalization compared to other LLM-based approaches.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
