Design of a Knowledge Visualization Framework: From Speech to Knowledge Graph

Organizations are increasingly focused on preserving their knowledge base, even when skilled employees depart. Similarly, educational institutions aim to make classrooms dynamic spaces for seamless knowledge transfer between teachers and students. This study proposes a design for a system that can tackle knowledge-sharing challenges by offering innovative solutions. In our system, as a teacher or expert delivers a lecture, the audio is captured via a microphone and saved in WAV or MP3 format. This audio is then processed using script-recognition APIs to extract the core insights of the lecture. By converting tacit knowledge into an explicit format, we minimize the risk of losing valuable information. The outcome of this work is the construction of a knowledge graph in which concepts are the nodes and the relationships between concepts are the edges. Each node carries a definition of the concept, which can be toggled to display or hide, allowing learners to gain a comprehensive understanding with visual support. In addition to audio files, a transcript of any other audio file in PDF or any document in PDF can also be used for converting tacit knowledge into a visualized knowledge form.