New Brown Corpus

New Brown Corpus是由布朗大学创建的一个数据集，旨在模拟儿童早期语言学习环境中的自然语言数据。该数据集包含18000个单词，通过虚拟现实厨房环境收集，提供了丰富的视觉和空间信息。数据集的创建过程涉及在VR环境中进行自然语言的收集，并设计了多种任务以确保词汇的广泛覆盖。该数据集主要用于研究基于环境的语言模型，特别是动词学习，旨在解决从自然语言数据中学习语义表示的挑战。

The New Brown Corpus is a dataset developed by Brown University, which aims to simulate natural language data in early childhood language learning environments. This dataset contains 18,000 words, collected in a virtual reality (VR) kitchen environment, and provides rich visual and spatial information. The process of creating this dataset involves collecting natural language within the VR environment and designing multiple tasks to ensure broad lexical coverage. It is primarily used for research on environment-grounded language models, particularly verb learning, and aims to address the challenges of learning semantic representations from natural language data.