Inference From Machine Learning Models

I've created a machine learning model from a number of predictors using GentleBoost Ensemble Learning. Is there a way to view the relative significance of the predictors, for the purposes of inference, or is the model only useful for prediction?

Thanks.