The National Science Foundation (NSF) Division of Information and Intelligent Systems awarded a $484,509 Project Grant to The Ohio State University to develop a computational model for extracting world knowledge from pre-trained language models and applying it to improve the accuracy of grammar induction from text. The project, titled "COMPCOG: RI: SMALL: HUMAN-LIKE SEMANTIC GRAMMAR INDUCTION THROUGH KNOWLEDGE DISTILLATION FROM PRE-TRAINED LANGUAGE MODELS", aims to create a semantic grammar induction model that can better capture the logical structure of human language by integrating knowledge from large neural language models. The grant period runs from Sep 1, 2023 to Aug 31, 2026 under the NSF's Computer and Information Science and Engineering (CFDA 47.070) program. The proposed work will continue efforts to develop resources for evaluating structural models of grammar learning.
Generated 5/14/24, 8:37 AM