TL;DR本研究探讨了轨迹（即物体位置和旋转随时间变化的道路）能否自然地编码动词语义，研究结果表明，轨迹本身与某些动词（如 fall）相关联，而通过自我监督预训练来进行进一步的抽象可以更好地捕捉动词含义上的细微差别（如 roll 与 slide）。
Abstract
distributional models learn representations of words from text, but are
criticized for their lack of grounding, or the linking of text to the
non-linguistic world. grounded language models have had success in lea