汉语韵律短语切分及语调降阶和下倾研究

Firstly, prosodic boundaries of 1991 common sentences were labeled based on speech perception experiment, relation between prosodic structure and syntactic structure was examined after immediate constituent analysis, an example of prosodic phrasing from text sentences was provided using CART. Then, using designed sentences, phenomena of downstep and declination in pitch downtrend of Chinese declarative sentences were examined, commonness and speciality of Chinese intonation were discussed. The main results of the study are: 1 The distribution patterns of prosodic phrase boundaries for different syntactic structures are different, and there is great freedom in prosodic chunking. The relation between syntactic structure and prosodic structure can only be discussed in statistical sense. 2 Besides of syntactic relation, the second most important factor which influences prosodic phrase boundaries is length. The distances to the front boundary and the back boundary are more important than the lengths of the left syntactic contituent and the right one. In our corpus, the length distributions of prosodic phrases are 5±3 syllables. 3 Automatic downstep can lower intonation linearly, but is affected by stress easily. Non-automatic downstep lowers the higher part of pitch contours and has no effect on the lower one of the intonation. 4 The downtrend reason of low point is declination. The extent of declination relates to not only tones of low points, but also their positions in prosodic words, the baselines decline much faster when low point are in the initial position of a prosodic word. In long sentences, the baselines of prosodic phrases are the basic declination units, and the whole declination pattern of a sentence is related to syntactic relations between two neighboring prosodic phrases.

Minimizing Statistical Bias with Queries

I describe an exploration criterion that attempts to minimize the error of a learner by minimizing its estimated squared bias. I describe experiments with locally-weighted regression on two simple kinematics problems, and observe that this "bias-only" approach outperforms the more common "variance-only" exploration approach, even in the presence of noise.

Surface Reflectance Estimation and Natural Illumination Statistics

Humans recognize optical reflectance properties of surfaces such as metal, plastic, or paper from a single image without knowledge of illumination. We develop a machine vision system to perform similar recognition tasks automatically. Reflectance estimation under unknown, arbitrary illumination proves highly underconstrained due to the variety of potential illumination distributions and surface reflectance properties. We have found that the spatial structure of real-world illumination possesses some of the statistical regularities observed in the natural image statistics literature. A human or computer vision system may be able to exploit this prior information to determine the most likely surface reflectance given an observed image. We develop an algorithm for reflectance classification under unknown real-world illumination, which learns relationships between surface reflectance and certain features (statistics) computed from a single observed image. We also develop an automatic feature selection method.

An Accelerated Chow and Liu Algorithm: Fitting Tree Distributions to High Dimensional Sparse Data

Chow and Liu introduced an algorithm for fitting a multivariate distribution with a tree (i.e. a density model that assumes that there are only pairwise dependencies between variables) and that the graph of these dependencies is a spanning tree. The original algorithm is quadratic in the dimesion of the domain, and linear in the number of data points that define the target distribution $P$. This paper shows that for sparse, discrete data, fitting a tree distribution can be done in time and memory that is jointly subquadratic in the number of variables and the size of the data set. The new algorithm, called the acCL algorithm, takes advantage of the sparsity of the data to accelerate the computation of pairwise marginals and the sorting of the resulting mutual informations, achieving speed ups of up to 2-3 orders of magnitude in the experiments.

Inferring 3D Structure with a Statistical Image-Based Shape Model

We present an image-based approach to infer 3D structure parameters using a probabilistic "shape+structure'' model. The 3D shape of a class of objects may be represented by sets of contours from silhouette views simultaneously observed from multiple calibrated cameras. Bayesian reconstructions of new shapes can then be estimated using a prior density constructed with a mixture model and probabilistic principal components analysis. We augment the shape model to incorporate structural features of interest; novel examples with missing structure parameters may then be reconstructed to obtain estimates of these parameters. Model matching and parameter inference are done entirely in the image domain and require no explicit 3D construction. Our shape model enables accurate estimation of structure despite segmentation errors or missing views in the input silhouettes, and works even with only a single input view. Using a dataset of thousands of pedestrian images generated from a synthetic model, we can perform accurate inference of the 3D locations of 19 joints on the body based on observed silhouette contours from real images.

Permutation Tests for Classification

We introduce and explore an approach to estimating statistical significance of classification accuracy, which is particularly useful in scientific applications of machine learning where high dimensionality of the data and the small number of training examples render most standard convergence bounds too loose to yield a meaningful guarantee of the generalization ability of the classifier. Instead, we estimate statistical significance of the observed classification accuracy, or the likelihood of observing such accuracy by chance due to spurious correlations of the high-dimensional data patterns with the class labels in the given training set. We adopt permutation testing, a non-parametric technique previously developed in classical statistics for hypothesis testing in the generative setting (i.e., comparing two probability distributions). We demonstrate the method on real examples from neuroimaging studies and DNA microarray analysis and suggest a theoretical analysis of the procedure that relates the asymptotic behavior of the test to the existing convergence bounds.

A Unified Statistical and Information Theoretic Framework for Multi-modal Image Registration

We formulate and interpret several multi-modal registration methods in the context of a unified statistical and information theoretic framework. A unified interpretation clarifies the implicit assumptions of each method yielding a better understanding of their relative strengths and weaknesses. Additionally, we discuss a generative statistical model from which we derive a novel analysis tool, the "auto-information function", as a means of assessing and exploiting the common spatial dependencies inherent in multi-modal imagery. We analytically derive useful properties of the "auto-information" as well as verify them empirically on multi-modal imagery. Among the useful aspects of the "auto-information function" is that it can be computed from imaging modalities independently and it allows one to decompose the search space of registration problems.

Statistical Evaluation of the Predictive Toxicology Challenge 2000-2001

Toivonen, H., Srinivasan, A., King, R. D., Kramer, S. and Helma, C. (2003) Statistical Evaluation of the Predictive Toxicology Challenge 2000-2001. Bioinformatics 19: 1183-1193

Vegetation and Topographic Control of Wind-blown Snow Distributions in Distributed and Aggregated Simulations for an Arctic Tundra Basin

Essery, RLH & JW, Pomeroy, (2004). Vegetation and topographic control of wind-blown snow distributions in distributed and aggregated simulations. Journal of Hydrometeorology, 5, 735-744.