<< September 2019 >>

Tuesday, September 3, 2019

Seminar 217, Risk Management: Does the Leverage Effect Affect the Distribution of Return

Seminar: Risk Seminar | September 3 | 11 a.m.-12:30 p.m. | 1011 Evans Hall

 Speaker: Dangxing Chen, UC Berkeley

 Consortium for Data Analytics in Risk

The leverage effect refers to the generally negative correlation between the return of an asset and the changes in its volatility. There is broad agreement in the literature that the effect should be present, and it has been consistently found in empirical work. However, a few papers have pointed out a puzzle: the return distribution of many assets do not appear to be affected by the leverage...   More >

Wednesday, September 4, 2019

Functional inequalities of the Infinite swapping algorithm: theory and applications

Seminar: Probability Seminar | September 4 | 3:10-4:10 p.m. | 1011 Evans Hall

 Wenpin Tang, Berkeley IEOR

 Department of Statistics

Sampling Gibbs measures at low temperature is a very important task but computationally very challenging. Numeric evidence suggest that the infinite-swapping algorithm (isa) is a promising method. The isa can be seen as an improvement of replica methods which are very popular. We rigorously analyze the ergodic properties of the isa in the low temperature regime deducing Eyring-Kramers formulas...   More >

Statistics over algorithms as a model of human learning: Neyman Seminar

Seminar | September 4 | 4-5 p.m. | 1011 Evans Hall

 Steve Piantadosi, UC Berkeley

 Department of Statistics

Human learning differs qualitatively from state of the art machine learning both in scale and power. People are able to discover much richer latent structures in data than are typically captured in statistical models. In particular, people seem able to discover algorithmically sophisticated representations, often real computational processes like computer programs. This ability can be seen in...   More >

Tuesday, September 10, 2019

Seminar 217, Risk Management: Towards theoretical understanding of large batch training in stochastic gradient descent

Seminar: Risk Seminar | September 10 | 11 a.m.-12:30 p.m. | 1011 Evans Hall

 Speaker: Xiaowu Dai, UC Berkeley

 Consortium for Data Analytics in Risk

ABSTRACT: Stochastic gradient descent (SGD) is almost ubiquitously used in training non-convex optimization tasks. Recently, a hypothesis by Keskar et al. (2017) that large batch SGD tends to converge to sharp minima has received increasing attention. We justify this hypothesis by providing new properties of SGD in both finite-time and asymptotic regimes, using tools from Partial Differential...   More >

Wednesday, September 11, 2019

Gambler's ruin in three dimensions.

Seminar: Probability Seminar | September 11 | 3:10-4 p.m. | 1011 Evans Hall

 Persi Diaconis, Stanford University

 Department of Statistics

Picture three gamblers with initial capital x(A), x(B), x(C) summing to N. Each time a pair of gamblers is chosen uniformly and they flip a fair coin. Consider the first time one of them hits zero.
How are the fortunes of the other two distributed and how does this depend on how they start?
Approximations (upper and lower bounds with reasonable constants) are derived for parallel problems on...   More >

What do we know about how to make good predictions?: Neyman Seminar

Seminar: Neyman Seminar | September 11 | 4-5 p.m. | 1011 Evans Hall

 Danny Hernandez, OpenAI

 Department of Statistics

Everyone makes bets. Scientists bet years of their lives on research agendas, CEO’s bet billions of dollars on new products, and world leaders bet our welfare through their policies. Their decisions often hinge on implicit judgement based predictions about relatively one-off events rather than on data. We’ll review the most promising existing techniques for improving one’s predictions. I’ll...   More >

Tuesday, September 17, 2019

Seminar 217, Risk Management: Characteristics of Mutual Fund Portfolios: Where Are the Value Funds?

Seminar: Risk Seminar | September 17 | 11 a.m.-12:30 p.m. | 1011 Evans Hall

 Speaker: Martin Lettau, UC Berkeley

 Consortium for Data Analytics in Risk

ABSTRACT: This paper provides a comprehensive analysis of portfolios of active mutual funds, ETFs and hedge funds through the lens of risk (anomaly) factors. We show that that these funds do not systematically tilt their portfolios towards profitable factors, such as high book-to-market (BM) ratios, high momentum, small size, high profitability and low investment growth. Strikingly, there are...   More >

Wednesday, September 18, 2019

General selection models: Bernstein duality and minimal ancestral structures

Seminar: Probability Seminar | September 18 | 3:10-4 p.m. | 1011 Evans Hall

 Sebastian Hummel, Bielefeld University

 Department of Statistics

We construct a sequence of Moran models that converges for large populations under suitable conditions to the $\Lambda$-Wright-Fisher process with a drift that is vanishing at the boundaries. The genealogical structure inherent in the graphical representation of the finite population models leads in the large population limit to a generalisation of the ancestral selection graph of Krone and...   More >

Regression analysis of longitudinal data with omitted asynchronous longitudinal covariate: Neyman Seminar

Seminar: Neyman Seminar | September 18 | 4-5 p.m. | 1011 Evans Hall

 Hongyuan Cao, Florida State University

 Department of Statistics

Extended follow-up with longitudinal data is common in many medical investigations. In regression analyses, a longitudinal covariate may be omitted, often because it is not measured synchronously with the longitudinal response. Naive approach that simply ignores the omitted longitudinal covariate can lead to biased estimators. In this article, we establish conditions under which estimation is...   More >

Tuesday, September 24, 2019

Seminar 217, Risk Management: Self-excited Black-Scholes models for option pricing

Seminar: Risk Seminar | September 24 | 11 a.m.-12:30 p.m. | 1011 Evans Hall

 Speaker: Alec Kercheval, Florida State University

 Consortium for Data Analytics in Risk

Beginners first learn to price stock options with a simple binomial tree model for random price changes. It is well known that this classical one-dimensional random walk converges weakly to Brownian motion in the proper space-time scaling limit. Actual stock prices changes occur not at regular times but at random times according to the order flow in an electronic limit order book...   More >

Wednesday, September 25, 2019

Localization of Gaussian disordered systems at low temperature

Seminar: Probability Seminar | September 25 | 3:10-4 p.m. | 1011 Evans Hall

 Erik Bates, U.C. Berkeley

 Department of Statistics

The fundamental premise of statistical mechanics is that a physical system's state is random according to some probability measure, which is determined by the various forces of interaction between the system's constituent particles. In the ``disordered" setting, these interactions are also random (meant to capture the effect of a random medium), meaning the probability measure is itself a random...   More >

Overlapping Clustering Models, and One (class) SVM to Bind Them All: Neyman Seminar

Seminar: Neyman Seminar | September 25 | 4-5 p.m. | 1011 Evans Hall

 Purnamrita Sarkar, UT Austin

 Department of Statistics

People belong to multiple communities, words belong to multiple topics, and books cover multiple genres; overlapping clusters are commonplace. Many existing overlapping clustering methods model each person (or word, or book) as a non-negative weighted combination of “exemplars” who belong solely to one community, with some small noise. Geometrically, each person is a point on a cone whose corners...   More >