About The Forum
Calendar 2017 Poster Sessions : Semi-Supervised Deep Regression through Posterior Regularization
Abstract:
Large amounts of data are typically required to train deep learning models. For many important problems, however, acquiring additional data can be expensive or even impossible. We present semi-supervised deep kernel learning (SSDKL), a semi-supervised regression model based on minimizing predictive variance in the posterior regularization framework. SSDKL combines the hierarchical representation learning of neural networks with the probabilistic modeling capabilities of Gaussian processes. By leveraging unlabeled data, we show improved performance on a diverse set of real-world regression tasks.
Bio:
Neal is a third-year PhD student advised by Prof. Stefano Ermon. He is interested in how machine learning can help to solve global sustainability challenges such as the elimination of poverty. Besides computational sustainability, his research interests include semi-supervised and unsupervised learning.
Large amounts of data are typically required to train deep learning models. For many important problems, however, acquiring additional data can be expensive or even impossible. We present semi-supervised deep kernel learning (SSDKL), a semi-supervised regression model based on minimizing predictive variance in the posterior regularization framework. SSDKL combines the hierarchical representation learning of neural networks with the probabilistic modeling capabilities of Gaussian processes. By leveraging unlabeled data, we show improved performance on a diverse set of real-world regression tasks.
Bio:
Neal is a third-year PhD student advised by Prof. Stefano Ermon. He is interested in how machine learning can help to solve global sustainability challenges such as the elimination of poverty. Besides computational sustainability, his research interests include semi-supervised and unsupervised learning.