Tony Jebara - - Machine Learning

Academic Home Research learning vision behaviour wearables Papers Courses Resumé Code Recent Personal Hobbies Trips Links
			Machine Learning We are primarily interested in methods to automatically build models and make predictions from real-world data. The types of datasets we work with include images and computer vision problems, text, time series, social networks and web data. The applications include computational models of humans from interaction, behavior and appearance. In the general area of machine learning we are interested in both probabilistic models such as Bayesian networksand discriminative models such as support vector machines. We are actively involved in uncovering new learning paradigms, algorithms and models that can directly find patterns in real data and can then be used for inference, prediction, resynthesis, etc. In addition, we are interested in graph-based algorithms and invariance. Please see the papers link for details. Also, follow the link to my group's home page. Columbia Machine Learning Laboratory.