Developing a Linguistically-wise and Computationally Feasible Digital System of Semantic Representation for NLP Applications

Project: Research

View graph of relations

Description

This study aims to promote interdisciplinary collaboration between linguistics and computation science, by constructing an innovative, large-scaled, linguistically-wise, deep-annotated and web-accessible verbal resource for potential NLP tasks, such as digital language learning, readability assessment and metaphor detection. The proposed two-year project will be achieved with a plan of three implementational stages, including Stage 1: to improve deep learning model for the automatic annotator by using LSTIM-CRF; Stage 2: to develop online and offline interfaces and formulate a human-machine interactive annotation workflow; Stage 3: to expand the data coverage of the existing Mandarin VerbNet (MVN) to include verbs in HSK graded vocabulary for better utilization. The proposed interdisciplinary research is an initial attempt to team up theoretical and computational linguists to launch a significant joint venture in developing a linguistically wise and applicationally useful database for the advancement of digital society 

Detail(s)

Project number7020010
Grant typeSIRG
StatusActive
Effective start/end date30/06/21 → …