The MINERvA experiment is designed to precisely measure the cross sections of different neutrino interactions on various nuclear targets via the neutrino nucleus scattering. In the first part of the presentation I will show the results for the direct measurement of nuclear dependence of the charged current quasielastic(CCQE)-like neutrino interactions on various nuclear targets. In the second part of my talk, I will report on the ongoing and recent ventures in the application of machine learning algorithms in the analysis to boost the sensitivity. Specifically, I will show the vertex reconstruction via ML approach and the application of domain adversarial neural network to remove the possible biasses coming from the physics model used for the training. Then I will show initial results for particle identification from the Semantic Segmentation approach.