Driven by a wide variety of technologies, major societal shifts in demographics and values, and evolving policy instruments and planning practices, unprecedented changes are underway in transportation. It has never been more vital to understand and predict the behavioral impacts of these changes. However, our ability to do so is severely hampered by the absence from our models of a major class of variables that has been repeatedly demonstrated to be vital to nearly every decision individuals make — specifically, attitudes (including opinions, feelings, preferences, perceptions, and personality). Several factors have historically prevented the incorporation of attitudes into large-scale travel-demand forecasting models, including the challenges associated with measuring them in traditional travel behavior surveys, and a current inability to forecast them in the way that socioeconomic variables are forecast. The TOMNET research team is engaged in creating and testing a variety of innovative and practical approaches to overcoming these barriers. These approaches have in common that they use attitudinal data collected from one sample to inform models built on a different sample. The Center will conduct extensive, coordinated, and systematic exploration of various machine learning and statistical data fusion approaches, involving applications to a diverse array of important topics (such as equity, vehicle ownership, the adoption of autonomous vehicles and ride-hailing apps, safety, resilience, active transportation, and land use impacts on travel) in multiple geographic regions (e.g., Phoenix, Seattle, San Francisco, Los Angeles, Tampa, and Atlanta). Through its work, the Center will identify the most promising approaches for integrating attitudinal variables and latent constructs in regional travel demand forecasting models, and quantifying the effects of these traditionally unobserved traits on behavioral choices and transport outcomes.