This paper proposes GoAI, a tool to bridge the gap between information acquisition and innovation in artificial intelligence (AI) learning. GoAI constructs an educational knowledge graph from AI research papers and leverages it to plan personalized learning paths and support creative idea generation. To address the limitations of existing large-scale language model (LLM)-based approaches, which lack semantic information from prior knowledge and citation relationships, GoAI constructs a knowledge graph where papers and prior knowledge (concepts, technologies, tools, etc.) serve as nodes and semantic information from citation relationships serve as edges. Using beam search-based path exploration, GoAI traces recent trends in the field from specific papers and plans learning paths. The integrated Idea Studio provides feedback on problem clarification, design comparisons, novelty, clarity, feasibility, and alignment with learning objectives.
