Dynamic Knowledge Graph Construction for Real-Time Natural Language Understanding in Autonomous Systems

This work presents Real- Time Semantic Mapping (RTSM), a new idea for improving the language comprehension in autonomous systems using the knowledge graph. RTSM solely relies on ConceptNet as its dataset and incorporate Spacy to parse and synchronize commends and signals from natural language appropriately. However, through the continuous updating of the knowledge graph, RTSM allows an autonomous system to provide accurate responding to real-time contextual changes with relatively low computational cost. Constant learning and semantic enhancement in this model make RTSM light and easily scalable, putting it at a higher ground as a natural language processing technique for use in autonomous systems. Due to a simple framework that makes it easy to incorporate all the knowledge, the RTSM is ideal in environments that are challenged for time and resources, but need to provide real time interaction capabilities.