Model Inference and Averaging Prof. Liqing Zhang Dept. Computer Science & Engineering, Shanghai Jiaotong University

Model Inference and Averaging Contents The Bootstrap and Maximum Likelihood Methods Bayesian Methods Relationship Between the Bootstrap and Bayesian Inference The EM Algorithm MCMC for Sampling from the Posterior Bagging Model Averaging and Stacking Stochastic Search: Bumping 2017/4/22 Model Inference and Averaging

Bootstrap by Basis Expansions Consider a linear expansion The least square error solution The Covariance of \beta 2017/4/22 Model Inference and Averaging

Model Inference and Averaging 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Parametric Model Assume a parameterized probability density (parametric model) for observations 2017/4/22 Model Inference and Averaging

Maximum Likelihood Inference Suppose we are trying to measure the true value of some quantity (xT). We make repeated measurements of this quantity {x1, x2, … xn}. The standard way to estimate xT from our measurements is to calculate the mean value: and set . DOES THIS PROCEDURE MAKE SENSE? The maximum likelihood method (MLM) answers this question and provides a general method for estimating parameters of interest from data. 2017/4/22 Model Inference and Averaging

The Maximum Likelihood Method Statement of the Maximum Likelihood Method Assume we have made N measurements of x {x1, x2, …, xn}. Assume we know the probability distribution function that describes x: f(x, a). Assume we want to determine the parameter a. MLM: pick a to maximize the probability of getting the measurements (the xi's) we did! 2017/4/22 Model Inference and Averaging

The MLM Implementation The probability of measuring If the measurements are independent, the probability of getting the measurements we did is: We can drop the dxn term as it is only a proportionality constant. 2017/4/22 Model Inference and Averaging

Log Maximum Likelihood Method We want to pick the a that maximizes L: Often easier to maximize lnL. L and lnL are both maximum at the same location. we maximize rather than L itself because converts the product into a summation. 2017/4/22 Model Inference and Averaging

Log Maximum Likelihood Method The new maximization condition is: could be an array of parameters (e.g. slope and intercept) or just a single variable. equations to determine a range from simple linear equations to coupled non-linear equations. 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Gaussian Let f(x, ) be given by a Gaussian distribution function. Let =μ be the mean of the Gaussian. We want to use our data+MLM to find the mean. We want the best estimate of a from our set of n measurements {x1, x2, …, xn}. Let’s assume that s is the same for each measurement. 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Gaussian Gaussian PDF The likelihood function for this problem is: 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Gaussian We want to find the a that maximizes the log likelihood function: 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Gaussian If are different for each data point then is just the weighted average: Weighted Average 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Poisson Let f(x, ) be given by a Poisson distribution. Let = μ be the mean of the Poisson. We want the best estimate of a from our set of n measurements {x1, x2, … xn}. Poisson PDF: 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Poisson The likelihood function for this problem is: 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example: Poisson Find a that maximizes the log likelihood function: Average 2017/4/22 Model Inference and Averaging

General properties of MLM For large data samples (large n) the likelihood function, L, approaches a Gaussian distribution. Maximum likelihood estimates are usually consistent. For large n the estimates converge to the true value of the parameters we wish to determine. Maximum likelihood estimates are usually unbiased. For all sample sizes the parameter of interest is calculated correctly. 2017/4/22 Model Inference and Averaging

General properties of MLM Maximum likelihood estimate is efficient: the estimate has the smallest variance. Maximum likelihood estimate is sufficient: it uses all the information in the observations (the xi’s). The solution from MLM is unique. Bad news: we must know the correct probability distribution for the problem at hand! 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Maximum Likelihood We maximize the likelihood function Log-likelihood function 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Score Function Assess the precision of using the likelihood function Assume that L takes its maximum in the interior parameter space. Then 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Likelihood Function We maximize the likelihood function We omit normalization since only adds a constant factor Think of L as a function of with Z fixed Log-likelihood function 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Fisher Information Negative sum of second derivatives is the information matrix is called the observed information, should be greater 0. Fisher information ( expected information ) is 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Sampling Theory Basic result of sampling theory The sampling distribution of the max-likelihood estimator approaches the following normal distribution, as when we sample independently from This suggests to approximate the distribution with 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Error Bound The corresponding error estimates are obtained from The confidence points have the form 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Simplified form of the Fisher information Suppose, in addition, that the operations of integration and differentiation can be swapped for the second derivative of f(x;θ) as well, i.e., In this case, it can be shown that the Fisher information equals The Cramér–Rao bound can then be written as 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Single-parameter proof If the expectation of T is denoted by ψ(θ), then, for all θ, Let X be a random variable with probability density function f(x;θ). Here T = t(X) is a statistic, which is used as an estimator for ψ(θ). If V is the score, i.e. then the expectation of V, written E(V), is zero. If we consider the covariance cov(V,T) of V and T, we have cov(V,T) = E(VT), because E(V) = 0. Expanding this expression we have 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Using the chain rule and the definition of expectation gives, after cancelling f(x;θ), because the integration and differentiation operations commute (second condition). The Cauchy-Schwarz inequality shows that Therefore 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example Consider a linear expansion The least square error solution The Covariance of \beta 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example The confidence region 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Bayesian Methods Given a sampling model Pr(Z | ) and a prior Pr( ) for the parameters, estimate the posterior probability By drawing samples or estimating its mean or mode Differences to mere counting ( frequentist approach ) Prior: allow for uncertainties present before seeing the data Posterior: allow for uncertainties present after seeing the data 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Bayesian Methods The posterior distribution affords also a predictive distribution of seeing future values In contrast, the max-likelihood approach would predict future data on the basis of not accounting for the uncertainty in the parameters 2017/4/22 Model Inference and Averaging

Model Inference and Averaging An Example Consider a linear expansion The least square error solution 2017/4/22 Model Inference and Averaging

Model Inference and Averaging The posterior distribution for \beta is also Gaussian, with mean and covariance The corresponding posterior values for \mu(x), 2017/4/22 Model Inference and Averaging

Model Inference and Averaging 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Bootstrap vs Bayesian The bootstrap mean is an approximate posterior average Simple example: Single observation z drawn from a normal distribution Assume a normal prior for : Resulting posterior distribution 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Bootstrap vs Bayesian Three ingredients make this work The choice of a noninformative prior for The dependence of on Z only through the max-likelihood estimate Thus The symmetry of 2017/4/22 Model Inference and Averaging

Model Inference and Averaging Bootstrap vs Bayesian The bootstrap distribution represents an (approximate) nonparametric, noninformative posterior distribution for our parameter. But this bootstrap distribution is obtained painlessly without having to formally specify a prior and without having to sample from the posterior distribution. Hence we might think of the bootstrap distribution as a \poor man's" Bayes posterior. By perturbing the data, the bootstrap approximates the Bayesian effect of perturbing the parameters, and is typically much simpler to carry out. 2017/4/22 Model Inference and Averaging

Model Inference and Averaging The EM Algorithm The EM algorithm for two-component Gaussian mixtures Take initial guesses for the parameters Expectation Step: Compute the responsibilities 2017/4/22 Model Inference and Averaging

Model Inference and Averaging The EM Algorithm Maximization Step: Compute the weighted means and variances Iterate 2 and 3 until convergence 2017/4/22 Model Inference and Averaging

The EM Algorithm in General Baum-Welch algorithm Applicable to problems for which maximizing the log-likelihood is difficult but simplified by enlarging the sample set with unobserved ( latent ) data ( data augmentation ). 2017/4/22 Model Inference and Averaging

The EM Algorithm in General 2017/4/22 Model Inference and Averaging

Model Inference and Averaging 2017/4/22 Model Inference and Averaging

The EM Algorithm in General Start with initial params Expectation Step: at the j-th step compute as a function of the dummy argument Maximization Step: Determine the new params by maximizing Iterate 2 and 3 until convergence 2017/4/22 Model Inference and Averaging

Model Averaging and Stacking Given predictions Under square-error loss, seek weights Such that Here the input x is fixed and the N observations in Z are distributed according to P 2017/4/22 Model Inference and Averaging

Model Averaging and Stacking The solution is the population linear regression of Y on namely Now the full regression has smaller error, namely Population linear regression is not available, thus replace it by the linear regression over the training set 2017/4/22 Model Inference and Averaging