
StorEBook: How Goldilocks and the Three Bears is Driving Text-to-Speech Technology
When it comes to critics of machine-generated speech, three-year-olds might be the toughest around. If they are going to sit and listen to their favorite stories, a witch better sound evil and a wolf better huff and puff.
Current text-to-speech (TTS) systems are not up to the challenge of expressing the often strong emotions in children’s stories. AT&T researchers are looking to add expression for dramatic effect with the StorEBook project, which uses word prominence to not only please the young set but also to improve TTS for everyone. Read more
Technical Documents
Using n-Grams for Syndromic Surveillance in a Turkish Emergency Department without English Translation: A Feasibility Study
Sylvia Halasz, Philip Brown, Cem Oktay, Arif Alper Cevik, Isa Kilicaslan, Colin Goodall, Dennis Cochrane, Tom Fowler, Guy Jacobson, Simon Tse, John Allegra
Accurate and Efficient Private Release of Datacubes and Contingency Tables
Graham Cormode, Cecilia Procopiuc, Divesh Srivastava, Grigory Yaroslavtsev
Yank: Enabling Green Data Centers to Pull the Plug
Kadangode Ramakrishnan, Univ. of Massachusetts Rahul Singh, Univ. of Massachusetts David Irwin, Univ. of Massachusetts Prashant Shenoy
Previous Editions