The three terms in the title of the workshop are three facets of the same basic question: what information can be gleaned from observed data? In information theory, especially in universal compression, observed data is used to better compress new information; in machine learning, observed data is applied to classify and predict new instances; and in big data, observed data helps with data mining and more general inferences about the domain. This workshop will bring together participants from these three communities to combine different techniques and apply them to problems in diverse applications areas. The techniques of interest include distribution modeling, sublinear sample learning, sparse recovery, and spectral methods in machine learning. Applications may include data compression, data security, natural language processing, advertising, data mining, bioinformatics and genomics, social networks, and finance. One particular theme is the learning of high dimensional structure, on which much recent progress has been made in the various communities.
Further details about this workshop will be posted in due course. Enquiries may be sent to the organizers at this address.