The quantitative model is the core of quantitative investment and the key carrier to realize the investment concept. With the continuous development of technology, AI models are becoming more and more widely used in quantitative investment. With its powerful learning, cognition, and reasoning ability, an AI model can develop trading models for different markets and asset types more quickly, and has faster self-iteration ability for rapidly changing markets. But the learning process of AI models often ignores the causal relationship and interpretability between data and targets, and this "black box" operation makes it difficult to understand how they work. By introducing AI-interpretable technology, interpretable constraints are introduced at the data level, architectural level, and decision-making level.