Microsoft Research Blog

English

  1. Memory Management in the Tera MTA Computer System 

    June 1, 1990 | Richard Korry, Cathy McCann, and Burton Smith

    This paper describes memory scheduling for the Tera MTA (Multi Threaded Architecture) computer system. The Tera MTA is intended to support a mixture of large and small tasks running in parallel, and ensure that they all make progress commensurate with their importance. We describe the…

  2. Lapped Transforms for Efficient Transform/Subband Coding 

    June 1, 1990 | Henrique S. Malvar

    Two new lapped transforms for subband/transform coding of signals are introduced: a new version of the lapped orthogonal transform (LOT), which can be efficiently computed for any transform length; and the modulated lapped transform (MLT), which is based on a modulated quadrature mirror filter (QMF)…

  3. Conceptual Set Covering: Improving Fit-and-Split Algorithms 

    June 1, 1990 | Carl Kadie

    Many learning systems implicitly use the fit-and-split learning method to create a comprehensive hypothesis from a set of partial hypotheses. At the core of the fit-and-split method is the assignment of examples to partial hypotheses. To date, however, this core has been neglected. This paper…

  4. Fast surface interpolation using hierarchical basis functions 

    May 31, 1990 | Rick Szeliski

    An alternative to multigrid relaxation that is much easier to implement and more generally applicable is presented. Conjugate gradient descent is used in conjunction with a hierarchical (multiresolution) set of basis functions. The resultant algorithm uses a pyramid to smooth the residual vector before the…

  5. Live Text 

    May 2, 1990 | Balachauder Krishnamurthy and Christopher Fraser

    Describes a software package that allows the user to edit the output of several common software tools and to cause the changes to be written back to the input files. For example, it is possible to edit the output of a spelling checker and have…

  6. Local Spatial Frequency Analysis for Computer Vision 

    May 1, 1990 | John Krumm and Steven A. Shafer

    A sense of vision is a prerequisite for a robot to function in an unstructured environment. However, real-world scenes contain many interacting phenomena that lead to complex images which are difficult to interpret automatically. Typical computer vision research proceeds by analyzing various effects in isolation…

  7. An overview of the SPHINX speech recognition system 

    April 30, 1990 | Kai-Fu Lee, Hsiao-Wuen Hon, and R. Reddy

    A description is given of SPHINX, a system that demonstrates the feasibility of accurate, large-vocabulary, speaker-independent, continuous speech recognition. SPHINX is based on discrete hidden Markov models (HMMs) with LPC- (linear-predictive-coding) derived parameters. To provide speaker independence, knowledge was added to these HMMs in several…

  8. VideoDraw: A Video Interface for Collaborative Drawing 

    April 15, 1990 | John Tang

    This paper describes VideoDraw, a shared drawing tool, and the process by which it is being designed and developed. VideoDraw is a prototype, video-based, tool that provides a shared “virtual sketchbook” among two or more collaborators. It not only allows the collaborators to see each…