Alternative Graph Neural Networks: Synergizing GEV Models and Deep Learning for Travel Mode Choice Modeling
arXiv:2509.07123v2 Announce Type: replace-cross Abstract: Generalized extreme value models capture dependence among choice alternatives in discrete choice modeling, but require this dependence to be predefined, symmetric, and shared uniformly across individuals. Recent efforts to synergize discrete choice models with deep neural networks have improved predictive performance but still cannot explicitly represent alternative dependence within neural architectures. To address these gaps, we introduce the alternative graph -- a graph in which nodes represent choice alternatives and edges encode th
This paper leverages recent advancements in deep learning to address long-standing limitations in discrete choice modeling, particularly in explicitly representing complex dependencies between choices.
Improving the accuracy and sophistication of predictive models for human decision-making, like travel mode choice, has broad implications for urban planning, transportation infrastructure, and the development of more adaptive AI agents.
Traditional discrete choice models are enhanced by integrating graph neural networks, allowing for more nuanced and dynamic representation of alternative dependencies previously hardcoded or oversimplified.
- · AI researchers in transport and urban planning
- · Smart city initiatives
- · Transportation modelling software developers
- · Developers of less sophisticated discrete choice models
- · Entities relying on static, predefined dependency assumptions
More accurate predictions of consumer and commuter behavior will improve resource allocation and infrastructure planning.
The methodology could be extended to other areas of discrete choice, such as product selection or policy adoption, enhancing AI agent capabilities.
As AI models better understand human choice, they could enable more personalized and dynamic interventions in various sectors, raising new ethical considerations around algorithmic influence.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG