Aligning LLM with Humans for Travel Choices: A Persona-Based Embedding Learning Approach

References

• Achiam J, Adler S, Agarwal S, Ahmad L, Akkaya I, Aleman FL, Almeida D, et al. (2023) Gpt-4 technical report. Preprint, submitted March 15, https://arxiv.org/abs/2303.08774.Google Scholar
• Anthis JR, Liu R, Richardson SM, Kozlowski AC, Koch B, Evans J, Brynjolfsson E, Bernstein M (2025) LLM social simulations are a promising research method. Preprint, submitted April 3, https://arxiv.org/abs/2504.02234.Google Scholar
• Argyle LP, Busby EC, Fulda N, Gubler JR, Rytting C, Wingate D (2023) Out of one, many: Using language models to simulate human samples. Political Anal. 31(3):337–351.Crossref, Google Scholar
• Arkoudi I, Krueger R, Azevedo CL, Pereira FC (2023) Combining discrete choice models and neural networks through embeddings: Formulation, interpretability and performance. Transportation Res. Part B Methodological. 175:102783.Crossref, Google Scholar
• Bierlaire M, Axhausen K, Abay G (2001) The acceptance of modal innovation: The case of Swissmetro. 1st Swiss Transport Res. Conf. Session Choices (Monte Verità/Ascona, Switzerland).Google Scholar
• Binz M, Akata E, Bethge M, Brändle F, Callaway F, Coda-Forno J, Dayan P., et al. (2025) A foundation model to predict and capture human cognition. Nature 644(8078):1002–1009.Crossref, Google Scholar
• Brown T, Mann B, Ryder N, Subbiah M, Kaplan JD, Dhariwal P, Neelakantan A, et al. (2020) Language models are few-shot learners. Larochelle H, Ranzato M, Hadsell R, Balcan MF, Lin H, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 33 (Curran Associates, Red Hook, NY), 1877–1901.Google Scholar
• Chamieh I, Zesch T, Giebermann K (2024) LLMs in short answer scoring: Limitations and promise of zero-shot and few-shot approaches. Kochmar E, Bexte M, Burstein J, Horbach A, Laarmann-Quante R, Tack A, Yaneva V, Yuan Z, eds. Proc. 19th Workshop Innovative Use NLP Building Ed. Appl. (BEA 2024) (Association for Computational Linguistics, Mexico City), 309–315.Google Scholar
• Chen Y, Liu TX, Shan Y, Zhong S (2023) The emergence of economic rationality of GPT. Proc. Natl. Acad. Sci. USA 120(51):e2316205120.Crossref, Google Scholar
• Chen C, He Y, Wang H, Chen J, Luo Q (2024a) DelayPTC-LLM: Metro passenger travel choice prediction under train delays with large language models. Preprint, submitted September 28, https://arxiv.org/abs/2410.00052.Google Scholar
• Chen R, Wang C, Sun Y, Zhao X, Xu S (2025a) From perceptions to decisions: Wildfire evacuation decision prediction with behavioral theory-informed LLMs. Che W, Nabende J, Shutova E, Pilehvar MT, eds. Proc. 63rd Annual Meeting Assoc. Comput. Linguistics, Long Papers, vol. 1 (Association for Computational Linguistics, Vienna, Austria), 29754–29778.Google Scholar
• Chen Y, Chi B, Li C, Zhang Y, Liao C, Chen X, Xie N (2025b) Toward interactive next location prediction driven by large language models. IEEE Trans. Comput. Soc. Systems 12(5):2696–2712.Crossref, Google Scholar
• Chen J, Wang X, Xu R, Yuan S, Zhang Y, Shi W, Xie J, et al. (2024b) From persona to personalization: A survey on role-playing language agents. Preprint, submitted April 28, https://arxiv.org/abs/2404.18231.Google Scholar
• Chuang YS, Nirunwiroj K, Studdiford Z, Goyal A, Frigo V, Yang S, Shah D, Hu J, Rogers T (2024) Beyond demographics: Aligning role-playing LLM-based agents using human belief networks. Al-Onaizan Y, Bansal M, Chen YN, eds. Findings Assoc. Comput. Linguistics: EMNLP 2024 (Association for Computational Linguistics, Miami), 14010–14026.Google Scholar
• Dargay JM (2001) The effect of income on car ownership: Evidence of asymmetry. Transportation Res. Part A Policy Practice 35(9):807–821.Crossref, Google Scholar
• Fan C, Chen J, Jin Y, He H (2024) Can large language models serve as rational players in game theory? A systematic analysis. Wooldridge M, Dy J, Natarajan S, eds. Proc. Thirty-Eighth AAAI Conf. Artificial Intelligence (AAAI Press, Washington, DC), 17960–17967.Google Scholar
• Fernando C, Banarse D, Michalewski H, Osindero S, Rocktäschel T (2023) Promptbreeder: Self-referential self-improvement via prompt evolution. Preprint, submitted September 28, https://arxiv.org/abs/2309.16797.Google Scholar
• Gao C, Lan X, Li N, Yuan Y, Ding J, Zhou Z, Xu F, Li Y (2024) Large language models empowered agent-based modeling and simulation: A survey and perspectives. Humanities Soc. Sci. Comm. 11(1):1–24.Google Scholar
• Gemini Team Google (2023) Gemini: A family of highly capable multimodal models. Preprint, submitted December 19, https://arxiv.org/abs/2312.11805.Google Scholar
• Gong L, Lin Y, Lu Y, Han X, Liu Y, Guo S, Lin Y, Wan H (2025) Mobility-LLM: Learning visiting intentions and travel preference from human mobility data with large language models. Globerson A, Mackey L, Belgrave D, Fan A, Paquet U, Tomczak JM, Zhang C, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 37 (Curran Associates, Red Hook, NY), 36185–36217.Google Scholar
• Grossmann I, Feinberg M, Parker DC, Christakis NA, Tetlock PE, Cunningham WA (2023) AI and the transformation of social science research. Science 380(6650):1108–1109.Crossref, Google Scholar
• Haj-Yahia S, Mansour O, Toledo T (2025) Incorporating domain knowledge in deep neural networks for discrete choice models. Transportation Res. Part C Emerging Tech. 171:105014.Crossref, Google Scholar
• Han Y, Pereira FC, Ben-Akiva M, Zegras C (2022) A neural-embedded discrete choice model: Learning taste representation with strengthened interpretability. Transportation Res. Part B Methodological 163:166–186.Crossref, Google Scholar
• Hansen AL, Horton JJ, Kazinnik S, Puzzello D, Zarifhonarvar A (2024) Simulating the survey of professional forecasters. Preprint, submitted February 10, https://doi.org/10.2139/ssrn.5066286.Google Scholar
• Hensher DA (2001) Measurement of the valuation of travel time savings. J. Transport Econom. Policy 35(1):71–98.Crossref, Google Scholar
• Jansen BJ, Jung S, Salminen J (2023) Employing large language models in survey research. Natural Language Processing J. 4:100020.Crossref, Google Scholar
• Johansson MV, Heldt T, Johansson P (2006) The effects of attitudes and personality traits on mode choice. Transportation Res. Part A Policy Practice 40(6):507–525.Crossref, Google Scholar
• Ke G, Meng Q, Finley T, Wang T, Chen W, Ma W, Ye Q, Liu TY (2017) LightGBM: A highly efficient gradient boosting decision tree. Guyon I, von Luxburg U, Bengio S, Wallach H, Fergus R, Vishwanathan SVN, Garnett R, eds. 31st Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 30 (Curran Associates, Red Hook, NY), 3146–3154.Google Scholar
• Khattab O, Singhvi A, Maheshwari P, Zhang Z, Santhanam K, Vardhamanan S, Haq S, et al. (2023) DSPy: Compiling declarative language model calls into self-improving pipelines. Preprint, submitted October 5, https://arxiv.org/abs/2310.03714.Google Scholar
• Kim J, Kovach M, Lee KM, Shin E, Tzavellas H (2024) Learning to be homo economicus: Can an LLM learn preferences from choice. Preprint, submitted January 14, https://arxiv.org/abs/2401.07345.Google Scholar
• Kojima T, Gu SS, Reid M, Matsuo Y, Iwasawa Y (2022) Large language models are zero-shot reasoners. Koyejo S, Mohamed S, Agarwal A, Belgrave D, Cho K, Oh A, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 35 (Curran Associates, Red Hook, NY), 22199–22213.Google Scholar
• Lahoz LT, Pereira FC, Sfeir G, Arkoudi I, Monteiro MM, Azevedo CL (2023) Attitudes and latent class choice models using machine learning. J. Choice Model. 49:100452.Crossref, Google Scholar
• Li A, Chen H, Namkoong H, Peng T (2025) LLM generated persona is a promise with a catch. Preprint, submitted March 18, https://arxiv.org/abs/2503.16527.Google Scholar
• Li X, Huang F, Lv J, Xiao Z, Li G, Yue Y (2024) Be more real: Travel diary generation using LLM agents and individual profiles. Preprint, submitted July 10, https://arxiv.org/abs/2407.18932.Google Scholar
• Liu T, Li M, Yin Y (2024) Can large language models capture human travel behavior? Evidence and insights on mode choice. Preprint, submitted September 19, https://doi.org/10.2139/ssrn.4937575.Google Scholar
• Liu T, Yang J, Yin Y (2025) Toward LLM-agent-based modeling of transportation systems: A conceptual framework. Artificial Intelligence Transportation 1:100001.Crossref, Google Scholar
• Liu S, Cao N, Chen Y, Jiang Y, Cong G (2024) NextLocLLM: Next location prediction using LLMs. Preprint, submitted October 11, https://arxiv.org/abs/2410.09129.Google Scholar
• Lu Y, Huang J, Han Y, Bei B, Xie Y, Wang D, Wang J, He Q (2025) Beyond believability: Accurate human behavior simulation with fine-tuned LLMs. Preprint, submitted March 26, https://arxiv.org/abs/2503.20749.Google Scholar
• Ma Y, Seo T (2025) Incorporating graph neural network into route choice model. Preprint, submitted March 4, https://arxiv.org/abs/2503.02315.Google Scholar
• Mikolov T, Chen K, Corrado G, Dean J (2013) Efficient estimation of word representations in vector space. Preprint, submitted January 16, https://arxiv.org/abs/1301.3781.Google Scholar
• Mokhtarian PL (2024) Pursuing the impossible (?) dream: Incorporating attitudes into practice-ready travel demand forecasting models. Transportation Res. Part A Policy Practice 190:104254.Crossref, Google Scholar
• Ouyang L, Wu J, Jiang X, Almeida D, Wainwright C, Mishkin P, Zhang C, et al. (2022) Training language models to follow instructions with human feedback. Koyejo S, Mohamed S, Agarwal A, Belgrave D, Cho K, Oh A, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 35 (Curran Associates, Red Hook, NY), 27730–27744.Google Scholar
• Park JS, O’Brien J, Cai CJ, Morris MR, Liang P, Bernstein MS (2023) Generative agents: Interactive simulacra of human behavior. Follmer S, Han J, Steimle J, Henry Riche N, eds. Proc. 36th Annual ACM Sympos. User Interface Software Tech. (Association for Computing Machinery, New York), Article 2, 1–22.Google Scholar
• Park JS, Zou CQ, Shaw A, Hill BM, Cai C, Morris MR, Willer R, Liang P, Bernstein MS (2024) Generative agent simulations of 1,000 people. Preprint, submitted November 15, https://arxiv.org/abs/2411.10109.Google Scholar
• Paulssen M, Temme D, Vij A, Walker JL (2014) Values, attitudes and travel behavior: A hierarchical latent variable mixed logit model of travel mode choice. Transportation 41(4):873–888.Crossref, Google Scholar
• Pennington J, Socher R, Manning CD (2014) GloVe: Global vectors for word representation. Moschitti A, Pang B, Daelemans W, eds. Proc. 2014 Conf. Empirical Methods Natural Language Processing (EMNLP) (Association for Computational Linguistics, Doha, Qatar), 1532–1543.Google Scholar
• Pereira FC (2019) Rethinking travel behavior modeling representations through embeddings. Preprint, submitted August 31, https://arxiv.org/abs/1909.00154.Google Scholar
• Rafailov R, Sharma A, Mitchell E, Manning CD, Ermon S, Finn C (2023) Direct preference optimization: Your language model is secretly a reward model. Oh A, Naumann T, Globerson A, Saenko K, Hardt M, Levine S, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 36 (Curran Associates, Red Hook, NY), 53728–53741.Google Scholar
• Schwanen T, Mokhtarian PL (2005) What affects commute mode choice: Neighborhood physical structure or preferences toward neighborhoods? J. Transport Geography 13(1):83–99.Crossref, Google Scholar
• Sifringer B, Lurkin V, Alahi A (2020) Enhancing discrete choice models with representation learning. Transportation Res. Part B Methodological 140:236–261.Crossref, Google Scholar
• Small KA, Verhoef ET, Lindsey R (2024) The Economics of Urban Transportation (Routledge, London).Crossref, Google Scholar
• Song Y, Wang T, Cai P, Mondal SK, Sahoo JP (2023) A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities. ACM Comput. Surveys 55(13s):1–40.Crossref, Google Scholar
• Sun C, Yang K, Reddy RG, Fung Y, Chan HP, Small K, Zhai C, Ji H (2025) Persona-DB: Efficient large language model personalization for response prediction with collaborative data refinement. Rambow O, Wanner L, Apidianaki M, Al-Khalifa H, Di Eugenio B, Schockaert S, eds. Proc. 31st Internat. Conf. Comput. Linguistics (Association for Computational Linguistics, Abu Dhabi, UAE), 281–296.Google Scholar
• Tseng YM, Huang YC, Hsiao TY, Chen WL, Huang CW, Meng Y, Chen YN (2024) Two tales of persona in LLMs: A survey of role-playing and personalization. Al-Onaizan Y, Bansal M, Chen YN, eds. Findings Assoc. Comput. Linguistics (Association for Computational Linguistics, Miami), 16612–16631.Google Scholar
• Vij A, Carrel A, Walker JL (2013) Incorporating the influence of latent modal preferences on travel mode choice behavior. Transportation Res. Part A Policy Practice 54:164–178.Crossref, Google Scholar
• Wang S, Mo B, Zhao J (2020) Deep neural networks for choice analysis: Architecture design with alternative-specific utility functions. Transportation Res. Part C Emerging Tech. 112:234–251.Crossref, Google Scholar
• Wang S, Mo B, Zhao J (2021) Theory-based residual neural networks: A synergy of discrete choice models and deep neural networks. Transportation Res. Part B Methodological 146:333–358.Crossref, Google Scholar
• Wang S, Wang Q, Bailey N, Zhao J (2021) Deep neural networks for choice analysis: A statistical learning theory perspective. Transportation Res. Part B Methodological 148:60–81.Crossref, Google Scholar
• Wang Q, Wang S, Zheng Y, Lin H, Zhang X, Zhao J, Walker J (2024b) Deep hybrid model with satellite imagery: How to combine demand modeling and computer vision for travel behavior analysis? Transportation Res. Part B Methodological 179:102869.Crossref, Google Scholar
• Wang L, Jiang Z, Hu C, Zhao J, Zhu Z, Chen X, Wang Z. , et al. (2025) Comparing AI and human decision-making mechanisms in daily collaborative experiments. iScience 28(6):112711.Crossref, Google Scholar
• Wang J, Jiang R, Yang C, Wu Z, Shibasaki R, Koshizuka N, Xiao C, et al. (2024a) Large language models as urban residents: An LLM agent framework for personal mobility generation. Globerson A, Mackey L, Belgrave D, Fan A, Paquet U, Tomczak JM, Zhang C, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 37 (Curran Associates, Red Hook, NY), 124547–124574.Google Scholar
• Wei J, Bosma M, Zhao VY, Guu K, Yu AW, Lester B, Du N, Dai AM, Le QV (2021) Finetuned language models are zero-shot learners. Preprint, submitted September 3, https://arxiv.org/abs/2109.01652.Google Scholar
• Wong M, Farooq B (2021) ResLogit: A residual neural network logit model for data-driven choice modelling. Transportation Res. Part C Emerging Tech. 126:103050.Crossref, Google Scholar
• Yang Y, Zhang W, Lin H, Liu Y, Qu X (2024) Applying masked language model for transport mode choice behavior prediction. Transportation Res. Part A Policy Practice 184:104074.Crossref, Google Scholar
• Yao S, Yu D, Zhao J, Shafran I, Griffiths T, Cao Y, Narasimhan K (2023) Tree of thoughts: Deliberate problem solving with large language models. Oh A, Naumann T, Globerson A, Saenko K, Hardt M, Levine S, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 36 (Curran Associates, Red Hook, NY), 11809–11822.Google Scholar
• Yuksekgonul M, Bianchi F, Boen J, Liu S, Lu P, Huang Z, Guestrin C, Zou J (2025) Optimizing generative AI by backpropagating language model feedback. Nature 639(8055):609–616.Crossref, Google Scholar
• Zhai Y, Tong S, Li X, Cai M, Qu Q, Lee YJ, Ma Y (2023) Investigating the catastrophic forgetting in multimodal large language models. Preprint, submitted September 19, https://arxiv.org/abs/2309.10313.Google Scholar
• Zhang Z, Sun Y, Wang Z, Nie Y, Ma X, Li R, Sun P, Ban X (2026) Large language models for mobility analysis in transportation systems: A survey on forecasting tasks. Transportation Res. Rec. 2680(2):756–774.Crossref, Google Scholar
• Zhao Y, Zhao Y, Du H, Yang HF (2025) Personalized decision modeling: Utility optimization or textualized-symbolic reasoning. Belgrave D, Zhang C, Lin H, Pascanu R, Koniusz P, Ghassemi M, Chen N, eds. Annual Conf. Neural Inform. Processing Systems, Advances in Neural Information Processing Systems, vol. 38 (Curran Associates, Red Hook, NY), 89317–89364.Google Scholar
• Ziems C, Held W, Shaikh O, Chen J, Zhang Z, Yang D (2024) Can large language models transform computational social science? Comput. Linguistics 50(1):237–291.Crossref, Google Scholar