Statistics
http://events.berkeley.edu/index.php/calendar/sn/stat.html
Upcoming EventsSeminar 217, Risk Management: Nonstandard Analysis and its Application to Markov Processes, Sep 18
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=118740&date=2018-09-18
Nonstandard analysis, a powerful machinery derived from mathematical logic, has had many applications in probability theory as well as stochastic processes. Nonstandard analysis allows construction of a single object - a hyperfinite probability space - which satisfies all the first order logical properties of a finite probability space, but which can be simultaneously viewed as a measure-theoretical probability space via the Loeb construction. As a consequence, the hyperfinite/measure duality has proven to be particularly in porting discrete results into their continuous settings.<br />
<br />
In this talk, for every general-state-space continuous-time Markov process satisfying appropriate conditions, we construct a hyperfinite Markov process to represent it. Hyperfinite Markov processes have all the first order logical properties of a finite Markov process. We establish ergodicity of a large class of general-state-space continuous-time Markov processes via studying their hyperfinite counterpart. We also establish the asymptotical equivalence between mixing times, hitting times and average mixing times for discrete-time general-state-space Markov processes satisfying moderate condition. Finally, we show that our result is applicable to a large class of Gibbs samplers and a large class of Metropolis-Hasting algorithms.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=118740&date=2018-09-18TAP free energy, spin glasses, and variational inference., Sep 19
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=119297&date=2018-09-19
We consider the Sherrington-Kirkpatrick model of spin glasses with ferromagnetically biased couplings. For a specific choice of the couplings mean, the resulting Gibbs measure is equivalent to the Bayesian posterior for a high-dimensional estimation problem known as ‘Z2 synchronization’. Statistical physics suggests to compute the expectation with respect to this Gibbs measure (the posterior mean in the synchronization problem), by minimizing the so-called Thouless-Anderson-Palmer (TAP) free energy, instead of the mean field (MF) free energy. We prove that this identification is correct, provided the ferromagnetic bias is larger than a constant (i.e. the noise level is small enough in synchronization). Namely, we prove that the scaled l2 distance between any low energy local minimizers of the TAP free energy and the mean of the Gibbs measure vanishes in the large size limit. Our proof technique is based on upper bounding the expected number of critical points of the TAP free energy using the Kac-Rice formula.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=119297&date=2018-09-19Correcting Bias in Eigenvectors of Financial Covariance Matrices, Sep 19
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=119956&date=2018-09-19
There is a source of bias in the sample eigenvectors of financial covariance matrices, when unchecked, distorts weights of minimum variance portfolios and leads to risk forecasts that are severely biased downward. Recent work with Lisa Goldberg and Alex Shkolnik develops an eigenvector bias correction. Our approach is distinct from the regularization and eigenvalue shrinkage methods found in the literature. We provide theoretical guarantees on the improvement our correction provides as well as estimation methods for computing the optimal correction from data.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=119956&date=2018-09-19Center for Computational Biology Seminar, Sep 19
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=119117&date=2018-09-19
Title: Two-phase differential expression analysis for single cell RNA-seq<br />
<br />
Abstract: Single-cell RNA-sequencing (scRNA-seq) has brought the study of the transcriptome to higher resolution and makes it possible for scientists to provide answers with more clarity to the question of ‘differential expression’. Specifically, it allows us to observe binary (On/Off) as well as continuous (the amount of expression) regulations. We present a method, SC2P, that identifies the phase of expression a gene is in, by taking into account of both cell- and gene-specific contexts, in a model-based and data-driven fashion. We then identify two forms of transcription regulation: phase transition, and magnitude tuning. We demonstrate that compared with existing methods, SC2P provides substantial improvement in sensitivity without sacrificing the control of false discovery, as well as better robustness. The ability to separately detect different forms of differential expression provides better interpretation of the nature of expression regulation. <br />
<br />
Bio:<br />
Dr. Zhijin (Jean) Wu is Associate Professor of Biostatistics at Brown University. She received her PhD in Biostatistics from Johns Hopkins in 2005 and has been a faculty at Brown since then. Her research interest is in statistical methods for analyzing gene expression and methylation.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=119117&date=2018-09-19Seminar 217, Risk Management: A Deep Learning Investigation of One-Month Momentum, Sep 25
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=118741&date=2018-09-25
The one-month return reversal in equity prices was first documented by Jedadeesh (1990), who found that there was a highly significant negative serial correlation in the monthly return series of stocks. This is in contrast to the positive serial correlation of the annual stock returns. Explanations for this effect differ, but the general consensus has been that the trailing one-month return includes a component of overreaction by investors. Since 1990, the one-month return reversal effect has decayed substantially, which has led others to refine it. Asness, Frazzini, Gormsen, and Pedersen (2017) refine this idea by adjusting MAX5 (the average of the five highest daily returns over the trailing month) for trailing volatility. They define a measure SMAX (scaled MAX5), which is the MAX5 divided by the trailing month daily return volatility. SMAX is designed to capture lottery demand in excess of volatility. They show that SMAX has an even stronger one-month return reversal than trailing month return.<br />
<br />
In this talk, I first replicate the results of Jedadeesh and Asness as benchmark models. I confirm that SMAX outperforms simple return reversal over the test period 1993-2017. However, the effectiveness of SMAX declines substantially over the test period. Using an enhanced combination of return statistics, I improve upon SMAX. I further improve upon SMAX by applying Neural Networks to trailing daily active returns. Note that all of these signals decay substantially in effectiveness over the common test period 1998-2017.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=118741&date=2018-09-25Machine Learning Panel, Sep 25
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=120147&date=2018-09-25
The Berkeley Master of Financial Engineering Program invites you to join us on September 25 at UC Berkeley's Haas School of Business for a Machine Learning Panel in the Spieker Forum of Chou Hall. Industry veterans will discuss applications of machine learning within their firms / industries and the state of the field.<br />
<br />
The Haas School of Business is located on the UC Berkeley campus at 2220 Piedmont Avenue, Berkeley, CA 94720.<br />
<br />
Schedule:<br />
3:30 PM Registration<br />
4:00 PM Program Begins: Panel Discussion and Q&A<br />
5:15 PM Reception and Networking (Refreshments will be served)<br />
<br />
Panelists: <br />
- Laurent El Ghaoui - Professor, EECS and IEOR; Berkeley Artificial Intelligence Research; Co-Founder, SumUp Analytics<br />
- Xin Heng - Senior Director, Data, Punchh<br />
- Bulent Kiziltan - Head of Deep Learning, Aetna<br />
- Stephen Malinak - Chief Data and Analytics Officer, TruValue Labs<br />
- Mike Ryerson - Senior Researcher, The Voleon Group<br />
- Frank Xia - Data Scientist, Opendoorhttp://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=120147&date=2018-09-25Stability of geodesics in the Brownian map, Sep 26
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=120101&date=2018-09-26
The Brownian map is a random non-differentiable surface, homeomorphic to the sphere, which was first identified as a scaling limit of random planar maps (Le Gall 2011 and Miermont 2011). More recently its connections with quantum gravity were established (Miller and Sheffield 2016). In this talk we show that the cut locus of the Brownian map is continuous almost everywhere, and discuss other features of its rich geodesic structure. <br />
<br />
Joint work with Omer Angel and Gregory Miermont.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=120101&date=2018-09-26Unraveling Controversy on Vexed Environmental Risks, Sep 26
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=120191&date=2018-09-26
Scientific assessment of many contemporary risks is plagued by controversy, persistent uncertainty, and polarized societal contexts. Decision makers often become mired in contested evidence, beset by uncertainties and contradictions. This leads to inaction on early warnings, paralysis-by-analysis, and erodes trust in science and its institutions. But why do controversies persist?<br />
<br />
A new conceptualisation of controversy seeks to unraveling its underlying sources and mechanisms. A new analytical framework maps the interpretive space in scientific assessment stemming from: (1) the multitude of ways in which risk issues can be translated into technical problems (translational diversity); (2) the multitude of tenable styles of scientific reasoning in interpreting evidence (argumentative flexibility) and (3) the existence of deep uncertainty (manufactured and actual) in the science.<br />
<br />
This new framework enables to unravel the interplay of scientific complexity, institutionalized practices of risk appraisal and societal discourses: Societal conflicts and interests co-shape the ways in which evidence is produced, communicated and used and how uncertainty is dealt with, in often hidden ways. Regulatory institutional settings co-define whose evidence counts and what style of scientific reasoning dominates. By integrating perspectives from 4 fields into an interdisciplinary analytical model, the interplay of scientific assessment with its polarised contexts can by analysed more systematically. Examples in the talk will draw on the controversy on neonicotinoids and pesticides.http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=120191&date=2018-09-26SPH Brown Bag Research Presentation, Sep 27
http://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=118829&date=2018-09-27
Malaria, dengue, Zika and other mosquito-borne diseases continue to pose a major global health burden through much of the world, despite the widespread distribution of insecticide-based tools and antimalarial drugs. Consequently, there is interest in novel strategies to control these diseases, including the release of mosquitoes transfected with Wolbachia and engineered with CRISPR-based gene drive systems. The safety and efficacy of these strategies and considerations regarding field trial design are critically dependent upon a detailed understanding of the distribution of mosquitoes and their movement between habitat patches. In this talk, I will discuss the work of my research group in using mathematical models to characterize the implementation of gene drive strategies to control mosquito populations, and new research directions.<br />
<br />
John Marshall is an Assistant Professor in Residence of Biostatistics and Epidemiology at the SPH. He has worked on several aspects of the project to engineer mosquitoes incapable of transmitting human diseases – social, cultural and regulatory issues at the UCLA Center for Society & Genetics, ecological field work at the Malaria Research and Training Center in Mali, molecular biology and population genetics at Caltech, and infectious disease modeling and epidemiological field work at Imperial College London. At UC Berkeley, his research group (www.MarshallLab.com) focuses on the use of mathematical models to inform novel genetics-based strategies for mosquito control, and to support efforts to control and eliminate mosquito-borne diseaseshttp://events.berkeley.edu/index.php/calendar/sn/stat.html?event_ID=118829&date=2018-09-27