You can find patterns without necessarily overfitting to them. It's more a question of correlation vs causation. Looking for patterns reported in research paper

You can find patterns without necessarily overfitting to them. It's more a question of correlation vs causation. Looking for patterns reported in research papers would be more promising. But cannot automate that.