DIARC-embodied-nlu-styled-4k

DIARC-LLM-Parser-Embodied-NLU-Styled-4K This dataset contains about ~4k utterances together with their semantic parses as interpretable by the DIARC cognitive robotic architecture. The parses are meant to capture the speech-theoretic aspects of NL and parse the intent, referents, and descriptors in the utterance. This dataset is one in a set of datasets. For this particular one, we programmatically built 127 utterances and semantics that are groundable in a robotic architecture… See the full description on the dataset page: https://huggingface.co/datasets/vsarathy/DIARC-embodied-nlu-styled-4k.