Learning-based trajectory prediction models have encountered great success, with the promise of leveraging contextual information in addition to motion history. Yet, we find that state-of-the-art forecasting methods tend to overly rely on the agents
dynamics, failing to exploit the semantic cues provided at its input. To alleviate this issue, we introduce CAB, a motion forecasting model equipped with a training procedure designed to promote the use of semantic contextual information. We also introduce two novel metrics -- dispersion and convergence-to-range -- to measure the temporal consistency of successive forecasts, which we found missing in standard metrics. Our method is evaluated on the widely adopted nuScenes Prediction benchmark.
