Abstract Learning class-conditional data distributions is crucial for Generative Adversarial Networks (GAN) in semisupervised learning. To improve both instance synthesis and classifification in this setting, we propose an enhanced TripleGAN (EnhancedTGAN) model in this work. We follow the adversarial training scheme of the original TripleGAN, but completely re-design the training targets of the generator and classififier. Specififically, we adopt featuresemantics matching to enhance the generator in learning class-conditional distributions from both the aspects of statistics in the latent space and semantics consistency with respect to the generator and classififier. Since a limited amount of labeled data is not suffificient to determine satisfactory decision boundaries, we include two classififiers, and incorporate collaborative learning into our model to provide better guidance for generator training. The synthesized high-fifidelity data can in turn be used for improving classififi- er training. In the experiments, the superior performance of our approach on multiple benchmark datasets demonstrates the effectiveness of the mutual reinforcement between the generator and classififiers in facilitating semi-supervised instance synthesis and classifification