1 paper
Michael Fore, Akshay Jain, Justin Downes +2
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 par…