Human beings are social creatures. We routinely reason about other agents, and a crucial component of this social reasoning is inferring people's goals as we learn about their actions. In many settings, we can perform intuitive but reliable goal inference from language descriptions of agents, actions, and the background environments. In this paper, we study this process of language driving and influencing social reasoning in a probabilistic goal inference domain. We propose a neuro-symbolic model that carries out goal inference from linguistic inputs of agent scenarios. The "neuro" part is a large language model (LLM) that translates language descriptions to code representations, and the "symbolic" part is a Bayesian inverse planning engine. To test our model, we design and run a human experiment on a linguistic goal inference task. Our model closely matches human response patterns and better predicts human judgements than using an LLM alone.

翻译：人类是社会性生物。我们日常推理他人的行为，而社会推理的关键组成部分是在了解他人行动时推断其目标。在许多情境中，我们能够根据关于行为者、行动及背景环境的语言描述，进行直觉性但可靠的目标推断。本文研究了语言驱动并影响社会推理的这一过程，聚焦于概率性目标推断领域。我们提出了一种神经符号模型，能从行为场景的语言输入中完成目标推断：其"神经"部分使用大语言模型（LLM）将语言描述转化为代码表征，"符号"部分则采用贝叶斯逆规划引擎。为测试该模型，我们设计并开展了一项针对语言目标推断任务的人类实验。结果表明，我们的模型与人类反应模式高度吻合，且在预测人类判断方面优于单独使用LLM的方法。