Statistics Across Campuses

Instability, Computational Efficiency and Statistical Accuracy *** Update 3 June 2022: This seminar is cancelled because the speaker has an unexpected family issue. We are trying to re-schedule the seminar with him.  Speaker: Nhat Ho (University of Texas at Austin) Time: 14-15h Friday 3 June 2022 Link: https://uni-sydney.zoom.us/j/85712683684 Abstract: Many statistical estimators are defined as the fixed point of a data-dependent operator, with estimators based on minimizing a cost function being an important special case. The limiting performance of such estimators depends on the properties of the population-level operator in the idealized limit of infinitely many samples. We develop a general framework that yields bounds on statistical accuracy based on the interplay between the deterministic convergence rate of the algorithm at the population level, and its degree of (in)stability when applied to an empirical object based on n samples. Using this framework, we analyze both stab

Quasi-score matching estimation for spatial autoregressive models with random weights matrix and regressors Speaker: Tao Zou (ANU) Time: 14-15pm Friday 27 May 2022 Location: Zoom at https://uni-sydney.zoom.us/j/85850162973 Abstract: Due to the rapid development of social networking sites, the spatial autoregressive (SAR) model has played an important role in social network studies. However, the commonly used quasi-maximum likelihood estimation (QMLE) for the SAR model is not computationally scalable as the network size is large. In addition, when establishing the asymptotic distribution of the parameter estimators of the SAR model, both weights matrix and regressors are assumed to be non-stochastic in classical spatial econometrics, which is perhaps not realistic in real applications. Motivated by the machine learning literature, quasi-score matching estimation for the SAR model is proposed. This new estimation approach is still likelihood-based, but significantly reduces t

T-Cal: An optimal test for the calibration of predictive models Date :  Thursday, 26 May 2022 Time:   10:00  - 11:00 am Speaker:    Assistant Professor Edgar Dobinban  (University of Pennsylvania)  Abstract: The prediction accuracy of machine learning methods is steadily increasing, but the calibration of their uncertainty predictions poses a significant challenge. Numerous works focus on obtaining well-calibrated predictive models, but less is known about reliably assessing model calibration. This limits our ability to know when algorithms for improving calibration have a real effect, and when their improvements are merely artifacts due to random noise in finite datasets. In this work, we consider detecting mis-calibration of predictive models using a finite validation dataset as a hypothesis testing problem. The null hypothesis is that the predictive model is calibrated, while the alternative hypothesis is that the deviation from calibration is sufficiently large. We find that detect

Computing Entropies with Nested Sampling Date: 03 June 2022, Friday Time: 4pm AEDT Speaker: Dr Brendon Brewer  (University of Auckland) Abstract: The Nested Sampling algorithm, invented in the mid-2000s by John Skilling, represented a major advance in Bayesian computation. Whereas Markov Chain Monte Carlo (MCMC) methods are usually effective for sampling posterior distributions, Nested Sampling also calculates the marginal likelihood integral used for model comparison, which is a computationally demanding task. However, there are other kinds of integrals that we might want to compute. Specifically, the entropy, relative entropy, and mutual information, which quantify uncertainty and relevance, are all integrals whose form is inconvenient in most practical applications. I will present my technique, based on Nested Sampling, for estimating these quantities for probability distributions that are only accessible via MCMC sampling. This includes posterior distributions, marginal distributio

Bayesian Multiple Instance Learning with Input from Attention Mechanism for Text Analysis Speaker: Professor Jing Cao (Southern Methodist University) Zoom link: https://uni-sydney.zoom.us/j/83582933974  As a branch of machine learning, multiple instance learning (MIL) learns from a collection of labeled bags, each containing a set of instances. Each instance is described by a feature vector. Since its emergence, MIL has been applied to solve various problems including content-based image retrieval, object tracking/detection, and computer-aided diagnosis. In this study, we apply MIL to text sentiment analysis. The current neural-network-based approaches in text analysis enjoy high classification accuracies but usually lack interpretability. The proposed Bayesian MIL model treats each text document as a bag, where the words are the instances. The model has a two-layered structure. The first layer identifies whether a word is essential or not (i.e., primary instance), and the second laye

Improved estimation of partially-specified models (Joint work with: Nicola Lunardon, University of Milano-Bicocca, Milan, Italy) Date :  Thursday, 19 May 2022 Time:   4 pm   - 5 pm Speaker:    Professor Ioannis Kosmidis  ( University of Warwick) Abstract: Many popular methods for the reduction of estimation bias rely on an approximation of the bias function under the assumption that the model is correct and fully specified. Other bias reduction methods, like the bootstrap, the jackknife and indirect inference require fewer assumptions to operate but are typically computer-intensive, requiring repeated optimization.   We present a novel framework for reducing estimation bias that: i) can deliver estimators with smaller bias than reference estimators even for partially-specified models, as long as estimation is through unbiased estimating functions; ii) always results in closed-form bias-reducing penalties to the objective function if estimation is through the maximization of one, like m

    Weakly non-Gaussian formulas of cosmological random fields. Date: Tuesday, 10 May 2022   Time: 6:00pm (AEST)   Speaker:  Prof. Takahiko Matsubara (High Energy Accelerator Research Organization (KEK), Japan) Contact the organizer: Andriy Olenko a.olenko@latrobe.edu.au Abstract: In cosmology, various kinds of random fields play important roles, including 3D distributions of galaxies and other astronomical objects, 2D distributions of cosmic microwave background radiations and weak lensing fields, etc. The features of non-Gaussianity in these fields contain a lot of cosmological information. In this talk, I will present a method to analytically describe the effects of weak non-Gaussianity field statistics, such as the peak abundance, peak correlations, Minkowsky functionals, etc. See zoom meeting details below. This seminar is a part of an international online seminars series on Statistics and Data Science applications in Astronomy : https://sites.google.com/view/iau-iaa-seminar/home