Non-Parametric Spatiotemporal Trajectory Prediction via State-Conditioned Transition Sampling

A new training-free method for multi-modal trajectory prediction achieves accuracy comparable to a 57M-parameter transformer while requiring no GPU and zero learned parameters, according to a paper posted on arXiv (2608.14349v1). The method builds a transition table of historical state-to-next-position pairs and retrieves neighbors using a product kernel over spatial proximity, bearing, speed, and temporal context. On the TrAISformer benchmark (Danish Maritime AIS), it remains stable down to 10% of training data where TrAISformer degrades catastrophically, enabling deployment in new geographic regions with an order of magnitude less historical data.

arXiv:2608.14349v1 Announce Type: new Abstract: We present a training-free method for multi-modal trajectory prediction that achieves comparable accuracy to a 57M-parameter transformer while requiring no GPU and zero learned parameters. The method builds a transition table of historical state-to-next-position pairs and retrieves neighbors using a product kernel over spatial proximity, bearing, speed, and temporal context. Two inference modes operate over this shared representation: diversity-penalized sampling produces trajectories covering distinct plausible routes, while beam search finds the highest-likelihood path. On the TrAISformer benchmark Danish Maritime AIS , our method achieves competitive accuracy at full data availability and dramatically outperforms the transformer in data-scarce regimes---remaining stable down to 10% of training data where TrAISformer degrades catastrophically. This enables deployment in new geographic regions from an order of magnitude less historical data, and with no GPU training.