MoveFM-R: Advancing Mobility Foundation Models via Language-driven Semantic Reasoning
On the map: LLM reasoning → Spatial and embodied reasoning
Abstract (arXiv)
Mobility Foundation Models (MFMs) have advanced the modeling of human movement patterns, yet they face a ceiling due to limitations in data scale and semantic understanding. While Large Language Models (LLMs) offer powerful semantic reasoning, they lack the innate understanding of spatio-temporal statistics required for generating physically plausible mobility trajectories. To address these gaps, we propose MoveFM-R, a novel framework that unlocks the full potential of mobility foundation models by leveraging language-driven semantic reasoning capabilities. It tackles two key challenges: the vocabulary mismatch between continuous geographic coordinates and discrete language tokens, and the representation gap between the latent vectors of MFMs and the semantic world of LLMs. MoveFM-R is built on three core innovations: a semantically enhanced location encoding to bridge the geography-language gap, a progressive curriculum to align the LLM's reasoning with mobility patterns, and an interactive self-reflection mechanism for conditional trajectory generation. Extensive experiments demonstrate that MoveFM-R significantly outperforms existing MFM-based and LLM-based baselines. It also shows robust generalization in zero-shot settings and excels at generating realistic trajectories from natural language instructions. By synthesizing the statistical power of MFMs with the deep semantic understanding of LLMs, MoveFM-R pioneers a new paradigm that enables a more comprehensive, interpretable, and powerful modeling of human mobility. The implementation of MoveFM-R is available online at https://anonymous.4open.science/r/MoveFM-R-CDE7/.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 35 of its 57 references that are in the map. First posted on arXiv 26 Sep 2025.
Builds on
35 of 57 references.
Closest ancestors · 2
Same direction, newest first
- 2025LocationReasoner: Evaluating LLMs on Real-World Site Selection Reasoning16 Jun
- 2024Chain-of-Planned-Behaviour Workflow Elicits Few-Shot Mobility Generation in LLMs15 Feb
Other references · 33
- 2017Proximal Policy Optimization Algorithmscited by 15,215
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
- 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
- 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
- 2025Qwen3 Technical Reportcited by 8,820
- 2015Fast R-CNNcited by 2,933
- 2025Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Modelscited by 1,261
- 2024Qwen2.5-Coder Technical Reportcited by 1,203
- 2016Generative Adversarial Text to Image Synthesiscited by 735
- 2022Autoregressive Image Generation using Residual Quantizationcited by 722
- 2021Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigmcited by 606
- 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Expertscited by 287
21 more
- 2024Rethinking Interpretability in the Era of Large Language Modelscited by 111
- 2024Large Language Models as Urban Residents: An LLM Agent Framework for Personal Mobility Generationcited by 100
- 2024Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Expertscited by 95
- 2020A Survey on Deep Learning for Human Mobilitycited by 88
- 2023DiffTraj: Generating GPS Trajectory with Diffusion Probabilistic Modelcited by 65
- 2023GETNext: Trajectory Flow Map Enhanced Transformer for Next POI Recommendationcited by 58
- 2024ControlTraj: Controllable Trajectory Generation with Topology-Constrained Diffusion Modelcited by 49
- 2024Simulating Human-like Daily Activities with Desire-driven Autonomycited by 30
- 2024LG-Traj: LLM Guided Pedestrian Trajectory Predictioncited by 25
- 2024Mobility-LLM: Learning Visiting Intentions and Travel Preferences from Human Mobility Data with Large Language Modelscited by 22
- 2025TrajLLM: A Modular LLM-Enhanced Agent-Based Framework for Realistic Human Trajectory Simulationcited by 22
- 2023Learning to Simulate Daily Activities via Modeling Dynamic Human Needscited by 20
- 2024UniTraj: Learning a Universal Trajectory Foundation Model from Billion-Scale Worldwide Tracescited by 18
- 2024TrajFM: A Vehicle Trajectory Foundation Model for Region and Task Transferabilitycited by 16
- 2025Generative Next POI Recommendation with Semantic IDcited by 13
- 2024Pretrained Mobility Transformer: A Foundation Model for Human Mobilitycited by 8
- 2025Enhancing Large Language Models for Mobility Analytics with Semantic Location Tokenizationcited by 8
- 2025One Fits All: General Mobility Trajectory Modeling via Masked Conditional Diffusioncited by 5
- 2024Deciphering Human Mobility: Inferring Semantics of Trajectories with Large Language Modelscited by 4
- 2025TransferTraj: A Vehicle Trajectory Learning Model for Region and Task Transferabilitycited by 4
- 2025From Points to Places: Towards Human Mobility-Driven Spatiotemporal Foundation Models via Understanding Placescited by 3
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2025MoveGPT: Scaling Mobility Foundation Models with Spatially-Aware Mixture of Experts
- 2025RHYTHM: Reasoning with Hierarchical Temporal Tokenization for Human Mobility
- 2023Where Would I Go Next? Large Language Models as Human Mobility Predictors
- 2024LIMP: Large Language Model Enhanced Intent-aware Mobility Prediction
- 2025Using LLMs for Analyzing AIS Data
- 2025TrajSceneLLM: A Multimodal Perspective on Semantic GPS Trajectory Analysis
- 2025Understanding the Geospatial Reasoning Capabilities of LLMs: A Trajectory Recovery Perspective
- 2025Zero-Shot Human Mobility Forecasting via Large Language Model with Hierarchical Reasoning