Using machine learning, we explore the utility of various deep neural networks when applied to high harmonic generation scenarios. First, we train the neural networks to predict the time-dependent dipoles and spectra of high harmonic emission from reduced-dimensionality models of di- and triatomic systems based on sets of randomly generated parameters (laser pulse intensity, internuclear distance, and molecular orientation). These networks, once trained, are useful tools to rapidly simulate the high harmonic spectra of our systems. Similarly, we have trained the neural networks to solve the inverse problem—to determine the molecular parameters based on high harmonic spectra or dipole acceleration data. The latter types of networks could then be used as spectroscopic tools to invert high harmonic spectra in order to recover the underlying physical parameters of a system. Next, we demonstrate that transfer learning can be applied to our networks to expand the range of applicability of the networks with only a small number of new test cases added to our training sets. Finally, we demonstrate neural networks that can be used to classify molecules by type: di- or triatomic, symmetric or asymmetric. With outlooks toward training with experimental data, these neural network topologies offer a novel set of spectroscopic tools that could be incorporated into high harmonic generation experiments.