Knowledge Grafts in the Age of Responsible Ea

If you know something about the physics of your problem, why should you start from zero and reinforcement learning and brute force your way through the problem? Ritsa, same thing with knowledge graps. If knowledge, knowledge grafts will infuse your with a lot of information. In some applications, you may not as o have enough data to label anyway, but you have enough knowledge to get going. Especially fora out of distribution rit cases, where and black swan events and things like that,. There's no data right around meter a pandemic rit, so all the models suddenly wrong. So what is the what is the e? Knowledge graft community thinking in terms of this