Abstract: Modeling human trajectories in crowded environments is challenging due to the complex nature of pedestrian behavior and interactions. This paper proposes a geometric graph neural network (GNN) architecture that integrates domain knowledge from psychological studies to model pedestrian interactions and predict future trajectories. Unlike prior studies using complete graphs, we define interaction neighborhoods using pedestrians' field of view, motion direction, and distance-based kernel functions to construct graph representations of crowds. Evaluations across multiple datasets demonstrate improved prediction accuracy through reduced average and final displacement error metrics. Our findings underscore the importance of integrating domain knowledge with data-driven approaches for effective modeling of human interactions in crowds.

Guided by a circular reticle that expanded and contracted along with the movements of my diaphragm, I floated gracefully through a Tron-like underwater world shaded by cool blues, cheerful pinks and relaxing purples as low-poly models of marine life swam about. It took only about 10 seconds or so for the game's breathing-based control scheme to become intuitive, at which point I'd become so engrossed in gliding through the environment and surrendering my mental state to the fuzzy murmur of binaural beats soundtracking it, that I'd forgotten about Tribeca and the media frenzy around me. In truth, I only took off the headset because I knew I had to make way for other journalists, not because I was ready to end my play session. When I did, Owen Harris, the video game designer who originated the concept for Deep VR, was delighted: According to Harris and his Dutch co-creator Niki Smit, I'd officially stayed in the experience longer than any other viewer who'd demoed it. "We want to help teach people these breathing techniques so that they can then manage these conditions outside of the game," says Harris of Deep VR's intended stress- and anxiety-reducing goal.

This paper argues for the position that experimental human studies are relevant to most facets of AI research and that closer ties between AI and experimental psychology will enhance the development of booth the principles of artificial intelligence and their implementation in computers. Raising psychological assumptions from the level of ad hoc intuitions to the level of systematic empirical observation, in the long run, will improve the quality of AI research and help to integrate it with related studies in other disciplines.

This paper argues for the position that experimental human studies are relevant to most facets of AI research and that closer ties between AI and experimental psychology will enhance the development of booth the principles of artificial intelligence and their implementation in computers. Raising psychological assumptions from the level of ad hoc intuitions to the level of systematic empirical observation, in the long run, will improve the quality of AI research and help to integrate it with related studies in other disciplines.