108643
Seminar 217, Risk Management: Institutional Investor Behavior and Market Dynamics
2017-04-25
11:00:00
13:00:00
11 a.m.
1 p.m.
Evans Hall
On Campus
639
32
Seminar
Primary
1793
Other Related Seminars
Secondary
Speaker: John Arabadjis
Speaker - Featured
State Street
Center for Risk Management Research
Happening As Scheduled
108885
Random walk driven by two-dimensional discrete Gaussian free field
2017-04-26
15:10:00
16:00:00
3:10
4 p.m.
Evans Hall
On Campus
1011
32
Seminar
Primary
776
Probability Seminar
Secondary
I will discuss the random walk driven by two-dimensional pinned discrete Gaussian Free Field (pDGFF). Explicitly, I will consider the Markov chain on the square lattice that jumps across an edge with probability proportional to the exponential of the gradient of pDGFF across that edge. The chain thus tends to move in the direction of increasing values of the pDGFF and this results in trapping. I...
Marek Biskup
U.C.L.A. Mathematics
Statistics, Department of
Happening As Scheduled
108270
Center for Computational Biology Seminar
Dr. Anne Carpenter, Broad Institute of Harvard and MIT
2017-04-26
16:00:00
17:00:00
4
5 p.m.
Li Ka Shing Center
On Campus
125
32
Seminar
Primary
1793
Other Related Seminars
Secondary
Title: Complex traits and simple systems
Light refreshments will be provided at reception from 3:30-4:00m, 125 LKS foyer.
Computational Biology, Center for
Happening As Scheduled
108464
Cross-validation with Confidence
2017-04-26
16:00:00
17:00:00
4
5 p.m.
Evans Hall
On Campus
1011
32
Seminar
Primary
773
Neyman Seminar
Secondary
Cross-validation is one of the most popular model selection methods
in statistics and machine learning. Despite its wide applicability,
traditional cross-validation methods tend to overfit, unless the ratio
between the training and testing sample sizes is very small.
We argue that such an overfitting tendency of cross-validation
is due to the ignorance of the uncertainty in the testing...
Jing Lei
Department of Statistics, CMU
Statistics, Department of
Happening As Scheduled
108704
Subtle but not malicious? The (high) computational cost of non-smoothness in learning from big data
2017-05-03
16:00:00
17:00:00
4
5 p.m.
Evans Hall
On Campus
1011
32
Seminar
Primary
773
Neyman Seminar
Secondary
What can we learn from big data? First, more data allows us to more precisely estimate probabilities of uncertain outcomes. Second, data provides better coverage to approximate functions more precisely. I will argue that the second is key to understanding the recent success of large scale machine learning. A useful way of thinking about this issue is that it is necessary to use many more...
Mikhail Belkin
Department of Computer Science and Engineering, Ohio State University
Statistics, Department of
Happening As Scheduled