Skip to:

Past Events

View events by calendar year
Enter one or more keywords to search for
Filter by event type
Wednesday, March 8, 2017 -
4:30pm to 5:30pm

Tamara Kolda, Sandia National Labs

Title: Tensor Decompositions: A Mathematical Tool for Data Analysis

Abstract: Tensors are multiway arrays, and tensor decompositions are powerful tools for data analysis and compression. In this talk, we demonstrate the wide-ranging utility of both the canonical polyadic (CP) and Tucker tensor decompositions with examples in neuroscience, chemical detection, and combustion science. The CP model is extremely useful for interpretation, as we show with an example in neuroscience. However, it can be difficult to fit to real data for a variety of reasons. We present a novel randomized method for fitting the CP decomposition to dense data that is more scalable and robust than the standard techniques. The Tucker model is useful for compression and can guarantee the accuracy of the approximation. We show that it can be used to compress massive data sets by orders of magnitude; this is done by determining the latent low-dimensional multilinear manifolds. This talk features joint work with Woody Austin (University of Texas), Casey Battaglino (Georgia Tech), Grey Ballard (Wake Forrest), Alicia Klinvex (Sandia),  Hemanth Kolla (Sandia), and Alex Williams (Stanford University).

Bio: Tamara (Tammy) Kolda is a member of the Data Science and Cyber Analytics Department at Sandia National Laboratories in Livermore, CA. She earned her Ph.D. in applied mathematics from the University of Maryland at College Park. She was awarded the Householder Postdoctoral Fellowship in Scientific Computing at Oak Ridge National Lab in 1997. She joined Sandia in 1999 and was named a Distinguished Member of Technical Staff in 2010. She received a 2003 Presidential Early Career Award for Scientists and Engineers (PECASE), was named a Distinguished Scientist of the Association for Computing Machinery (ACM) in 2011 and a Fellow of the Society for Industrial and Applied Mathematics (SIAM) in 2015. She was the winner of an R&D100 award and three best paper prizes at international conferences. Her work has been cited more than 8,500 times, and her 2009 article on Tensor Decompositions and Applications is currently the #1 most-downloaded article over all SIAM journals. She has led numerous projects in computational science and data analysis on topics in multilinear algebra and tensor decompositions, graph models and algorithms, data mining, optimization, nonlinear solvers, parallel computing and the design of scientific software. She has mentored more than nineteen summer interns and six postdocs. She has given keynotes talks at a variety of meetings including the International Symposium on Mathematical Programming (ISMP), the SIAM Conference on Computational Science & Engineering (CS&E), the SIAM Annual Meeting, and the IEEE International Conference on Data Mining (ICDM). She is currently a two-time elected member of the SIAM Board of Trustees, in her second term as Section Editor for the Software and High Performance Computing Section of SISC, and in her second term as Associate Editor for SIMAX.

This seminar replaces CME 500 for this week.

Time: 4:30-5:30 p.m.
Location: Hewlett 201 *Note change in room

Tuesday, March 7, 2017 -
12:30pm to 1:15pm

Presenters: Dr. Andrzej Banaszuk, Dr. Kishore Reddy, Dr. Tuhin Sahai, and Dr. Tong Sun

Abstract: We will present a broad overview of UTRC’s Systems Department research with particular focus on the areas of robotics, intelligent systems, machine learning, and advanced uncertainty management methods. The research, conducted by a diverse team of researchers in robotics, dynamical systems, control, applied mathematics, computer vision, machine learning, and human factors (in partnership with several leading universities including CMU, MIT, UPenn, and UC Berkeley) includes:

  • The cutting edge predictive and prescriptive analytics to automate decision processes and optimize human-machine efficiency for aerospace engineering, smart building and digital manufacturing. We will also present how we apply the state-of-art deep learning techniques to tackle the challenges in relevant industrial domains. 
  • The scalable algorithms and framework for uncertainty quantification and graph analytics. We leverage a combination of distributed graph clustering and adaptive waveform relaxation. Then we explore extensions to scalable inference for rare events in a Bayesian setting. We will give a short overview of our work on efficient algorithms for structure learning of Bayesian networks that is inspired from the efficient computations for NP-hard problems.

We will conclude with research problems of interest to UTRC and discuss existing and future career and internship opportunities in the broad area of robotics and machine learning.  

Speakers Bios: 

Dr. Andrzej Banaszuk is a Director of Systems Department at the United Technologies Research Center. Before his current appointment he was a Program Leader of Autonomous and Intelligent Systems and Sikorsky Program Offices at UTRC. Since joining UTRC in 1997, he has conducted research in analysis, design, and control of dynamical systems applied to jet engines, rotorcraft, electric power networks, and buildings. Since 2000 he has led collaborative multi-university research teams in the area of flow control, control of combustion instabilities, robust design of large uncertain dynamic networks, and autonomy. He is an author of 44 journal papers, 71 conference papers, and 9 patents. For his work on active and passive control of flow instabilities in jet engines he received IEEE Controls Systems Technology Award in 2007. He became an IEEE Fellow in 2011. He was elected to the Connecticut Academy of Science and Engineering in 2015. He holds Ph.D. in EE from Warsaw University of Technology and Ph.D. in Mathematics from Georgia Institute of Technology. 

Dr. Kishore Reddy is a Staff Research Scientist at the United Technologies Research Center (UTRC) working in the area of computer vision, human machine interaction (HMI) and machine learning. He is currently leading the Data Analytics initiative at UTRC primarily focusing on Deep Learning applications in aerospace and building systems to perform outliers and anomalies detection, multi-modal sensor fusion and data compression. He is also working with Department of Homeland Security (DHS) to develop algorithms for continuous authentication of mobile devices. Kishore earned his Ph.D. in 2012 from University of Central Florida, where he developed advanced video and image analysis algorithms, primarily segmentation and classification approaches, for multiple contracts funded by DARPA, IARPA and NIH. 

Dr. Tuhin Sahai is a Principal Research Scientist at the United Technologies Research Center (UTRC), broadly interested in the design, analysis and uncertainty quantification of complex systems. At UTRC, Tuhin has served as a principal investigator on multiple DARPA programs. In 2013, he was invited by the National Academy of Engineering to attend the Frontiers of Engineering Symposium. Furthermore, he was awarded the Grainger Foundation Frontiers of Engineering (FOE) grant by the National Academy of Engineering in 2014. Tuhin earned his Ph.D. in January 2008 from Cornell University, where he was a McMullen Fellow and won the H.D. Block teaching award. He received his Masters and Bachelors in Aerospace Engineering from the Indian Institute of Technology, Bombay in 2002.

Dr. Tong Sun is Group Leader for Decision Support & Machine Intelligence at the United Technologies Research Center. Her research portfolio includes using large-scale text and graph analytics to understand user behavior and social influence, applying deep neural network, active learning and online stream analytics to uncover actionable insights from heterogeneous data in distributed computing environments. Tong held 22 US Patents and co-authored over 35 peer-reviewed publications. She earned her Ph.D. in Distributed and Parallel Computing from University of Rhode Island in 1996. She is also an accomplished research thought-leader and technology innovator with 10+ years proven track of leadership in incubating new concepts through state-of-art machine learning methods/tools, developing advanced rapid prototypes that lead to impactful outcomes.

Monday, March 6, 2017 -
9:00am to 6:00pm

Each year, Stanford ICME hosts an Xtrapolate Roundtable on a topic relevant to computational mathematics. This year, ICME Xtrapolate focused on machine and deep learning. The goal of the roundtable was to explore ways that industry, academia and government can collaborate to address challenges,
and further the field. We covered two main topics during the day:
  • Research and Beyond: What machine/deep learning research is being undertaken that could have major impact in the short term or longer term? Where are the early wins in industry, and what are key success factors?
  • Data Ethics: Machine learning algorithms and potential for bias and discrimination -- how can we overcome these challenges?  
The slides and links shared by the speakers can be found in the agenda below. Please check back soon for a summary of the discussion.


Professor Ramesh Johari, Management Science & Engineering, Stanford 
Professor Sharad Goel, Management Science & Engineering, Stanford 
Professor James Zou, School of Medicine, Stanford 
Professor Margot Gerritson, ICME Director, Stanford
Timnit Gebru, AI Lab, Stanford 
Julie Bernauer, NVIDIA
Jeremy Pack, Google
Rukmini Ayer, Microsoft
Swaroop Kalasapur, Schlumberger
Stan Baginskis, Cisco
Reza Zadeh, Matroid/Stanford University
and more speakers from industry and academia.


8:30-9:00 Registration and Breakfast
9:00-9:15 Welcome Comments by Margot Gerritsen, ICME Director, Stanford University
9:15-10:15 Faculty Talks on Machine/Deep Learning Technologies
9:15-9:35: Reza Zadeh, Matroid/Stanford University (Link to slides can be found here)
9:35-9:55: Timnit Gebru, AI Lab, Stanford (Relevant papers can be found here)
9:55-10:15:Fireside Chat with Ramesh Johari, Management Science & Engineering, Stanford University
Interviewed by Esteban Arcaute, @WalmartLabs
10:15-10:30 Break
10:30-11:15 Industry Panel on Use Cases
  Julie Bernauer, NVIDIA
  Rukmini Iyer, Microsoft
  Swaroop Kalasapur, Schlumberger
  Stan Baginskis, Cisco
  Moderated by Karen Matthys, ICME
11:15-12:15 Roundtable Discussions
12:15-1:00 Lunch and Readouts
1:00-1:30 Break
1:30-2:30 Faculty and Industry Talks on Algorithms and Bias/Discrimination
       1:30-1:50: Sharad Goel, Management Science & Engineering, Stanford University (Link to slides can be found here and paper can be found here)
       1:50-2:10: Jeremy Pack, Google (Referenced articles can be found here)
       2:10-2:30: James Zou School of Medicine, Stanford University (Paper can be found here)
2:30-3:30 Roundtable Discussions
3:30-4:15 Readouts
4:15-4:45 Close
4:45-6:00 Reception


Friday, February 3, 2017 -
8:00am to 5:00pm

ICME proudly announces the Women in Data Science (WiDS) Conference will take place at Stanford University on Friday, February 3, 2017. 

To get a feeling for the event, view the video highlights from our inaugural event, including conference sessions and speaker interviews.

This one-day technical conference provides an opportunity to hear about the latest data science related research in a number of domains, learn now leading-edge companies are leveraging data science for success, and connect with potential mentors, collaborators, and others in the field. For more information, visit our WiDS conference website.

Join us! Live, via live stream, or become a WiDS Ambassador.



Thursday, January 19, 2017 -
8:30am to 5:30pm

Thursday, January 19, 2017

This event is now sold out, but click here if you would like to be added to the waitlist.
This event is free of charge.
The inaugural Artificial Intelligence in Fintech Forum at Stanford School of Engineering is sponsored by the Stanford Institute for Computational & Mathematical Engineering (ICME), Stanford Management Science and Engineering, Stanford Center for Financial and Risk Analytics, and White & Case Law Firm.
This event is organized by Kapil Jain, Director of the Math and Computational Finance Program in the Institute for Computational and Mathematical Engineering and Kay Giesecke, Associate Professor of Management Science and Engineering and Director of the Stanford Center for Financial and Risk Analytics.
The goals of this technical forum are:
  • To share latest technology advancements and research,
  • Discuss use cases, technical trends/challenges, and regulatory considerations,
  • Foster connections between industry and academia, and
  • Establish a model for ongoing technical dialog and partnerships

Tentative Agenda:

8:30 Registration and Breakfast
8:45 Welcome by Kapil Jain and Kay Giesecke
9:00 Apaar Sadhwani, Google Brain
9:30 Ashish Goel, Stanford MS&E and Stripe (keynote seminar)
10:15 Coffee
10:30 Tech Panel: Jeff Bohn from State Street, Andres Villaquiran from Alkanza, Adam Coates from Baidu, and John Ashley from NVIDIA
11:15 Steve Jurvetson, DFJ (keynote interview by Kapil Jain and Kay Giesecke)
12:00 Posters and Lunch
1:00 Benjamin Saul and James Greig, White & Case Law Firm
1:30 Amir Khosrowshahi, Intel/Nervana (keynote seminar)
2:15 Gustavo Schwenkler, Boston University and SQIntel
2:45 Coffee
3:15 Timnit Gebru, Stanford AI Lab
3:45 Keith Rabois, Khosla Ventures and Opendoor (keynote interview by Joe Grundfest)
4:30 Business Panel: Angela Strange from Andreessen Horowitz Venture Capital, Loek Janssen from Nova Credit, Chris Brahm from Bain & Company,
       and Kevin Petrasic from White & Case Law Firm
5:15 Reception
6:00 Finale
Click here to view the pictures and view bios of all the speakers.
For more info, please email


Saturday, December 3, 2016 -
9:00am to 4:15pm


The Bay Area Scientific Computing Day (BASCD) is an annual one-day meeting focused on fostering interactions and collaborations between researchers in the fields of scientific computing and computational science and engineering from the San Francisco Bay Area and nearby regions. BASCD provides junior researchers a venue to present their work to the local community and an opportunity for the Bay Area scientific and computational science and engineering communities at large to interchange views on today’s multidisciplinary computational challenges and state-of-the-art developments. Click here to visit the event website for schedule and other information.
Date: Saturday, December 3, 2016
Time: 9 a.m. - 4:15 p.m.
Location: Mackenzie Room of Huang Engineering Center at Stanford University (Address: 475 Via Ortega, Stanford, CA 94305)
While BASCD is a free event, registration is required. Please click here to register. 
Last day to register is on Monday, November 21, 2016.
Thursday, November 10, 2016 - 9:00am to Friday, November 11, 2016 - 1:00pm

UPDATE: Links to the presentations for the faculty tech talks are listed below.


This one-of-a-kind event brings ICME students and faculty together with our partners from industry and national labs for two days of networking and recruiting, discussions on current trends in our fields, workshops, and mixers.  Our main event will take place on Thursday and Friday, November 10-11, 2016. 

For details on special pre-Xtend events, please see the 'more information' section below.  

More Information

Information for ICME students:

Information for ICME partners:

Information on joining the ICME partnership programs:



Xtend Day 1, November 10, 2016

9:00-11:00 a.m.

Networking Breakfast

Huang basement, Student Commons

11:00 a.m.- 5:00 p.m.


See your schedule for specific appointments and locations

5:00 p.m.-7:30 p.m.


ICME lobby, Huang basement, Suite 060

Xtend Day 2, November 11, 2016

8:30- 9:00 a.m.


Mackenzie Room (300), Huang Engineering Center

9:00- 11:00 a.m.

Faculty Technology Talks:

John Carlsson, ICME alumnus and visiting scholar at Stanford, and Professor at USC, on Applying Computational Geometry to Modern Transportation Problems.

Click here to view John Carlsson's presentation. Here are the video clips for the presentation: Clip 1, Clip 2, Clip 3, Clip 4, and Clip 5.

Itai Ashlagi, professor in the department of Management Science & Engineering at Stanford, on Design and Analysis of Marketplaces.

Mackenzie Room (300), Huang Engineering Center

11:00- 11:15 a.m.



11:15- 12:30 p.m.

HIVE Overview & Demonstrations:

Alison Marsden, ICME affiliated faculty, will talk about patient-specific cardiovascular blood flow simulations.

G. Salim Mohammed, Head and Curator of the David Rumsey Map Center, will talk about the new David Rumsey Map Center. It also has high resolution screens such as the HIVE and will show some of the maps/visualizations in the HIVE.

HIVE Visualization Facility, Huang Basement, Room 050

12:30 p.m. Lunch  ICME lobby, Huang basement, Suite 060


Friday, November 4, 2016 -
1:00pm to 3:00pm

Data Visualization Center Open House

Calling all Stanford students, faculty, researchers, and data & technology aficionados. 

You are invited to the HIVE Xperience, an open house opportunity to see Stanford’s immersive visualization environment in action.                                                                                                                                                                                                                                                                                                                                

Friday, November 4, 2016

Drop by anytime between 1:00- 3:00pm

Huang Engineering Center, room 050


The HIVE features a 10-foot-tall by 24-feet-wide display with 13440x5400 resolution and 72 million total active pixels.  Come see how the HIVE helps researchers to visualize various aspects of data analysis and simulation, and to zoom-in to discover detail and nuance at previously unheard-of-levels. 


This event features collaborative visualization and research projects already taking place in the HIVE

1:00pm - Stanford's PBL Lab will show their research on how distributed settings affect individuals’ cognitive performance and learning
1:30pm - Stace Maples, of the Stanford Geospatial Center, will provide a brief demonstration of Google Earth Engine, which combines a multi-petabyte catalog of satellite imagery and geospatial datasets with planetary-scale analysis capabilities
2:00pm - A researcher with Professor Alison Marsden's Lab will show results from patient-specific cardiovascular blood flow simulations
2:30pm - Johnny Kuo of Global Business Services will present on Stanford's international projects, illustrating the University's activities in over 150 countries

No R.S.V.P. is needed; light refreshments will be served


About the HIVE

The HANA Immersive Visualization Environment (HIVE), was built in partnership between the Institute for Computational and Mathematical Engineering (ICME), the Army High Performance Computing Research Center (AHPCRC), and with the generous support of SAP.  Located in Huang 050, the HIVE may be reserved by Stanford faculty, students, and staff, free of charge, for research visualization projects and courses.

The HIVE features a 10-foot-tall by 24-feet-wide display with 13440x5400 resolution and 72 million total active pixels.  Researchers may use multiple displays simultaneously to investigate various aspects of data collection, simulation, and visualization and to zoom in to see detail at previously unheard-of-levels.

Learn more:



Wednesday, October 12, 2016 - 9:00am


GPUniversity of Deep Learning and Beyond

Formerly known as CUDA on Campus, GPUniversity of Deep Learning and Beyond is brought to you by NVIDIA and ICME (GPU Center of Excellence). Come explore the future of AI computing—from robotics and medicine to finance and self-driving cars. Discover how GPUs are powering this revolution with hands-on demos, deep learning workshops, and tech talks with experts who are creating new realities with AI.

Talks include Kay Giesecke on Deep Learning for Mortgage Risk, Silvio Savarese on Deep Learning for Vision and Robotics, and Daniel Rubin on Deep Learning for Medical Imaging.

To register for this event, please visit:


10:00AM – 11:00AM Keynote, Jen-Hsun Huang (Seating is limited) 
11:00AM – 12:30PM Speakers
-          11:00 am: Kay Giesecke: Deep Learning for Housing Markets
-          11:30am: Silvio Savarese: Deep Learning for Vision and Robotics
-          12:00pm: Daniel Rubin: Deep Learning for Medical Imaging 
12:30PM Hands-On Demos and Lunch
1:30PM Deep Learning Workshop: Approaches to Object Detection using DIGITS
3:00PM Deep Learning Workshop: Deep Learning Network Deployment
4:30PM Speaker, Bryan Catanzaro (VP,  Applied Deep Learning Research, NVIDIA) 
5:00PM Reception
Monday, September 19, 2016 - 9:00am to Thursday, September 22, 2016 - 3:45am

The ICME refresher course is a free, four-day long review intended to provide incoming students with an opportunity to review material relevant to their upcoming coursework. The material covered will help participants prepare for the first-year ICME core classes, as well as other classes in applied mathematics, science, and engineering. The courses are suitable for incoming ICME graduate students and other graduate students with a technical background.

Click here for the ICME Refresher Course website for more information including registration and schedule.

Monday, August 15, 2016 - 9:00am to Saturday, August 20, 2016 - 4:45pm

Each year, ICME offers a variety of summer workshops open to students, external partners, and the wider community. This year's series of day-long workshops is happening from August 15-20, 2016, as detailed below.  All workshops are from 9:00 a.m.-4:45 p.m., four 75-minute sessions, separated by time for breaks. Please note that these are not Stanford for-credit courses.

Registration is required, and will be issued on a first-come first-served basis to students and ICME External Partners. Other community members are welcome to attend, space permitting. Room locations will be given to those who register. 


2016 Summer Workshop Series

Monday, August 15

  • Introduction to Scientific Python   Register here  Instructor: Nick Henderson
  • Introduction to Programming in R    Register here   Instructors: Andreas Santucci and Lan Huong Nguyen
  • Advanced MATLAB for Scientific Computing   Register here  Instructor: Danielle Maddix

Tuesday, August 16

  • Introduction to Matrix Computations    Register here  Instructor: Margot Gerritsen
  • Connections Between Computational Neuroscience and Deep Learning    Register here  Instructor: Dave Deriso

Wednesday, August 17

  • Introduction to Mathematical Optimization   Register here  Instructor: AJ Friend

Thursday, August 18

  • Introduction to Statistical Data Analysis   Register here  Instructor: James Lambers

Friday, August 19

  • Introduction to Machine Learning   Register here  Instructors: Alex Ioannidis and Gabriel Maher

Saturday, August 20

  • Data Visualization    Register here  Instructor: Dave Deriso
Thursday, May 26, 2016 -
4:30pm to 5:30pm

Speaker: Svetlana Bryzgalova, Stanford

Title: Spurious Factors in Linear Asset Pricing Models

Abstract: When a risk factor has small covariance with asset returns, risk premia in the linear asset pricing models are no longer identified. Weak factors, similar to weak instruments, make the usual estimation techniques unreliable. When included in the model, they generate spuriously high significance levels of their own risk premia estimates, overall measures of fit and may crowd out the impact of the true sources of risk. I develop a new approach to the estimation of cross-sectional asset pricing models that: a) provides simultaneous model diagnostics and parameter estimates; b) automatically removes the effect of spurious factors; c) restores consistency and asymptotic normality of the parameter estimates, as well as the accuracy of standard measures of fit; d) performs well in both small and large samples. I provide new insights on the pricing ability of various factors proposed in the literature. In particular, I identify a set of robust factors (e.g. Fama-French ones, but not only), and those that suffer from severe identification problems that render the standard assessment of their pricing performance unreliable (e.g. consumption growth, human capital proxies and others).
Bio: Svetlana Bryzgalova is an Assistant Professor of Finance at the Stanford Graduate School of Business. She joined the GSB in September 2015 after receiving her PhD and MRes degrees in Economics from London School of Economics. Prior to attending LSE, Professor Bryzgalova graduated summa cum laude with a MSc in Financial Economics and BA in Economics (Mathematics) from the Higher School of Economics (Russia).
Thursday, May 26, 2016 -
4:30pm to 5:45pm

Speaker: James Lambers, Dept of Mathematics, University of Southern Mississippi

Title: A Crash Course on Matrices, Moments and Quadrature
Abstract: The aim of this talk is to give an overview of the beautiful mathematical relationships between matrices, moments, orthogonal polynomials, quadrature rules, and Krylov subspace methods.  The underlying goal is to obtain efficient numerical methods for estimating quantities of the form u^T f(A) v, where u and v are given vectors, A is a symmetric nonsingular
matrix, and f is a smooth function.
An obvious application is the computation of some elements of the matrix f(A) when all of f(A) is not required.  Computation of quadratic forms can yield error estimates in methods for solving systems of linear equations. Bilinear or quadratic forms also arise naturally for the computation of parameters in some numerical methods for solving least squares or total least squares problems, and also in Tikhonov regularization for solving ill-posed problems.  Furthermore, computation of bilinear forms is also useful in spectral methods for the numerical solution of PDE.
The main idea is to write I[f] as a Riemann-Stieltjes integral and then to apply Gaussian quadrature rules to approximate the integral.  The nodes and weights of these quadrature rules are given by the eigenvalues and eigenvectors of tridiagonal matrices whose nonzero coefficients describe the three-term recurrences satisfied by the orthogonal polynomials associated with the measure of the Riemann-Stieltjes integral.  Beautifully, these orthogonal polynomials can be generated by the Lanczos algorithm.
Results pertaining to orthogonal polynomials and quadrature rules may not be so well known in the matrix computation community, and conversely, researchers working with orthogonal polynomials and quadrature rules may not be very familiar with their applications to matrix computations.  We will see that it can be very fruitful to mix techniques coming from different areas. The resulting algorithms can also be of interest to scientists and engineers who are solving problems in which computation of bilinear forms arises naturally.
Bio: Jim Lambers graduated in 2003 from Gene Golub's SCCM program (the forerunner of ICME).  His current research is on spectral methods for time-dependent variable-coefficient PDE; simulation of gas injection processes for enhanced oil recovery (in collaboration with Margot Gerritsen), and denoising of images via nonlinear diffusion (in collaboration with Patrick Guidotti).
Tuesday, May 24, 2016 -
12:30pm to 1:30pm

Speaker: Dustin Schroeder, Stanford

Title: Radio Glaciology: A Window into the Physical Processes of Ice Sheets

Abstract: Radio echo sounding is a uniquely powerful geophysical technique for studying the interior of ice sheets, glaciers, and icy planetary bodies. It can provide broad coverage and deep penetration as well as interpretable ice thickness, basal topography, and englacial radio stratigraphy. However, despite the long tradition of glaciological interpretation of radar images, quantitative analyses of radar sounding data are rare and face several technical challenges. These include attenuation uncertainty from unknown ice temperature and chemistry, clutter and losses from surface and volume scattering, and a lack of problem-specific radar theory. However, there is rich, often underexploited, information in modern radar sounding data, which is being collected over terrestrial and planetary ice at an unprecedented rate. The development and application of hypothesis-driven analysis approaches for these data can place observational constraints on the morphologic, hydrologic, geologic, mechanical, thermal, and oceanographic configurations of ice sheets and glaciers. These boundary conditions – and the physical processes which they express and control – are filling a fundamental gap our ability to understand and predict the evolution, stability, and sea level contributions of marine ice sheets.

Bio: Dustin Schroeder is an Assistant Professor of Geophysics in the School of Earth, Energy, and Environmental Sciences at Stanford University. He works on the fundamental problem of observing, understanding, and predicting the configuration and evolution of ice sheet boundary conditions using ice penetrating radar sounding data. Before coming to Stanford he worked as a radar systems engineer with the Jet Propulsion Laboratory at the California Institute of Technology. He is also a science team member and co-investigator on the REASON (Radar for Europa Assessment and Sounding: Ocean to Near-Surface) radar sounder on NASA’s Europa Clipper mission and is an active collaborator on the RIME (Radar for Icy Moon Exploration) radar sounder for EASA’s JUICE mission to Ganymede. He received his PhD in geophysics from the University of Texas at Austin where he served as the lead radar engineer and operator during three Antarctic field seasons with the ICECAP and Operation Ice-Bridge projects.

Monday, May 23, 2016 -
4:30pm to 5:30pm

More details to follow.

Friday, May 20, 2016 -
1:00pm to 8:30pm

Video playlist of the Xpo Faculty Vision Talks are here

Some featured research:

Eileen Martin uses fast algorithms to detect and track permafrost

Gabriel Maher's 3D models of the cardiovascular system provide insight for deciding patient care options

Luke de Oliveira and Alfredo Lainez Rodrigo use machine learning to better categorize and understand text

May 20, 2016 ICME Xpo gives an up-close and inside look at current research and future plans for ICME faculty and students. ICME is engaged with over 50 faculty from 18 departments throughout Stanford. This is a unique opportunity to see how computational mathematics, data science, scientific computing, and related fields are applied across a wide range of domain areas.

ICME Xpo features an afternoon poster session followed by a series of faculty vision talks and promises plenty of opportunities to connect with ICME faculty and students, alumni, and partners from industry and laboratories.  ICME Xpo will take place in the Huang Engineering Center, Mackenzie Room (Room 300) followed by ICME Xtravaganza--the ICME end-of-year celebration held in the ICME suite in Huang Engineering Center, Suite B060.


1:00-2:40 p.m.: Poster Session. We expect 40+ posters on research in computational mathematics and data science applied to a wide range of fields.

Click here for a complete list of poster titles.

2:00- 2:40 p.m.: HANA Immersive Visualization Environment (HIVE) Demos. Visit the HIVE (Huang B050) to see demonstrations of research in action.

2:45-5:00 p.m.: Faculty Vision Talks. Topics will include: Autonomous Vehicles, Scalable Machine Learning, Modeling Extreme Events in Natural World, Experiments with Network Effects, Estimating Asset Pricing Factors, Recent Progress and Exploration of Linear Programming Algorithms, along with additional topics to be announced soon.

5:00-6:30 p.m.: Reception.

6:30- 8:30 p.m.: ICME Xtravaganza.  The ICME end-of-year celebration will begin immediately after Xpo, complete with music, food, end-of-year accolades, and conversations with students, faculty, and colleagues. ICME Xtravaganza will be held in the ICME suite in Huang Engineering Center, Suite B060.

Click here for the tentative agenda for Xpo. Additionally, here is a complete list of all the titles, abstracts, and bios for the faculty speakers and poster presenters.

Who should attend:

  • ICME Faculty, Students, Staff and Stanford colleagues
  • ICME's Partners in Industry and National Laboratories
  • ICME, SCCM, and Financial Math Alumni
  • Others who are interested in learning about ICME research
Registration closed on by May 12, 2016.


Thursday, May 19, 2016 -
4:30pm to 5:30pm

Speaker: Mikhail Chernov, UCLA

Title: Macroeconomic-driven Prepayment Risk and the Valuation of Mortgage-Backed Securities

Abstract: We introduce a reduced-form modeling framework for mortgage-backed securities in which we solve for the implied prepayment function from the cross section of market prices. From the implied prepayment function, we find that prepayment rates are driven not only by interest rates, but also by two macroeconomic factors: turnover and rate response. Intuitively, turnover represents prepayments for exogenous reasons like employment-related moves, household income shocks, and foreclosures, while rate response reflects frictions faced by borrowers in refinancing into a lower rate. We find that the implied turnover and rate response measures are in fact significantly related to macroeconomic measures such as consumption growth, the unemployment rate, housing values, credit availability, and market uncertainty. Implied prepayments are substantially higher than actual prepayments, providing direct evidence of significant prepayment risk premia in mortgage-backed security prices. We analyze the properties of the prepayment risk premium and find that it is almost entirely due to compensation for turnover risk. We also find evidence that mortgage-backed security prices were significantly affected by Fannie Mae credit risk and the Federal Reserve’s Quantitative Easing Programs.

Bio: Professor of Finance Mikhail Chernov focuses on macro-based asset pricing, derivatives, fixed income and financial econometrics. “My research could be characterized as measurement of various risks that financial markets are facing, and understanding how these risks could translate into expected returns,” he says. “My particular focus is on the importance of market crashes, private and sovereign defaults, and unexpected changes in policy — events that may occur infrequently, but whose impact can be devastating on financial markets and on the economy overall.” Chernov’s academic publications have earned him numerous awards.

Chernov is a research fellow at the Center for Economic and Policy Research and associate editor of Journal of Business and Economic Statistics, Journal of Econometrics, Journal of Finance, Journal of Financial Econometrics, and Journal of Financial and Quantitative Analysis. His professional interests lead him to international cross-disciplinary collaborations in statistics and macroeconomics. “I view my work as a never-ending pursuit of learning and discovering new things,” he says. “I try to pick co-authors from whom I can learn the most.”
Thursday, May 19, 2016 -
4:30pm to 5:45pm

Speaker: Nick Trefethen, Oxford

Bio: Nick Trefethen is Professor of Numerical Analysis and Head of the Numerical Analysis Group in the Oxford University Mathematical Institute. He was educated at Harvard University (AB 1977, summa cum laude) and Stanford University (MS 1980 and PhD 1982) and has held professorial positions at New York University (Courant Institute), MIT, Cornell University, and Oxford University. He is a Fellow of the Royal Society (London) and a member of the National Academy of Engineering (USA), and served as President of SIAM (Society for Industrial and Applied Mathematics) during 2011-2012.

As an author he is known for his books Numerical Linear Algebra (SIAM, 1997, with David Bau, III), Spectral Methods in MATLAB (SIAM, 2000), Schwarz-Christoffel Mapping (Cambridge U. Press, 2002, with Tobin Driscoll) and Spectra and Pseudospectra: The Behavior of Nonnormal Matrices and Operators (Princeton, 2005, with Mark Embree); he also won the Catherine Richards Prize for the best article in Mathematics Today in 2000. He has served as editor for many of the leading research journals in numerical analysis, and spent four years as Section Editor in charge of the major review articles in SIAM Review.
As a teacher he has won awards for graduate-level instruction at MIT, Cornell, and Oxford. An outgrowth of one of his graduate courses at Oxford was the 2002 "SIAM 100-Dollar, 100-Digit Challenge", which became the subject of a 2004 SIAM book by Bornemann, Laurie, Wagon and Waldvogel and its 2006 Springer successor in German.
Trefethen is a Highly Cited Researcher according to His approximately 100 publications in research journals span a number of areas within numerical analysis and applied mathematics, including non-normal eigenvalue problems and applications, spectral methods for differential equations, numerical linear algebra, fluid mechanics, computational complex analysis, and approximation theory. He is well known for his work on the pseudospectra, for the study of non-normal matrices and operators, and as the inventor of Chebfun.
Trefethen was the first winner of the Fox Prize in Numerical Analysis and was a Presidential Young Investigator. During 1998/99 he was the honorary Rouse Ball Lecturer in Applied Mathematics at the University of Cambridge. He received fellowships from the NSF (both graduate and post-doctoral), Hertz Foundation, and IBM. He is a frequent invited speaker at international conferences, including the quadrennial International Congress of Mathematicians (Berlin, 1998) and the International Congress on Applied Mathematics (Hamburg, 1995).
Trefethen was born August 30, 1955, grew up in Lexington, Massachusetts, and attended Shady Hill School and Phillips Exeter Academy.

Video of seminar can be found here.


Tuesday, May 17, 2016 -
12:30pm to 1:30pm

Speaker: Kapil Jain, Lecturer at ICME

This talk will be a brief overview of opportunities for quants, or folks with ICME skill set in finance, followed by a roundtable discussion format with interactive Q&A. If there are any specific related topics you wish to hear about, but wish to ask anonymously, feel free to email in advance.
Bio: Kapil Jain directs the MCF program within ICME; and previously worked as a quant at Citigroup, Perry Capital and DE Shaw, and interned at Microsoft Research and Trilogy software. He also ran a financial technology startup from his Rains dorm room as a Stanford graduate student.
Monday, May 16, 2016 -
4:30pm to 5:30pm

Speaker: Adam Backer, Graduate student at ICME

Title: Enhanced DNA imaging using super-resolution microscopy and simultaneous single-molecule orientation measurements

Abstract: In the past decade, a revolution has occurred in biological imaging. Using a conventional fluorescence microscope, it is now possible to image structures an order of magnitude smaller than the wavelength of light--thus achieving super-resolution. In this first part of my talk, I will briefly describe the basic experimental principles and computational methods that underlie single-molecule super-resolution microscopy. Next, I will present a novel image-processing technique for measuring the orientations and rotational dynamics of single fluorescent dye molecules over the course of a typical experiment. To demonstrate our approach, we use fluorescent dyes to obtain super-resolution images of stretched viral DNA. By leveraging our image data to obtain thousands of dye molecule orientation measurements, we develop a means of probing the nano-scale structure of individual DNA strands, while also characterizing dye-DNA interactions. 

Friday, May 13, 2016 -
4:00pm to 5:00pm

Join ICME for a TGIF seminar on Big Math in Small Companies!

  • Hear from industry experts about start-ups involved in computational math and data science
  • Get the scoop on how Start-X can help you get your venture off the ground
  • Practice pitching your own ideas and get expert feedback

Bucky Moore, Investor at Costanoa Venture Capital

Bucky is an early stage venture investor at Costanoa Venture Capital, where he works closely with various companies including Bugcrowd and Directly. Prior to joining Costanoa, Bucky was an investor at Battery Ventures, where he focused on early stage investments in enterprise software and infrastructure. During his time at Battery, he invested in various enterprise technology companies including Vera and Lightcyber.

Earlier in his career, Bucky was a member of the corporate development team at Cisco, where he led M&A and venture investment activity within the data center and cloud infrastructure segments. During his time at Cisco, he led investments in companies including Platfora, Embrane, 6Wind, Nantero, Whiptail, and Insieme Networks.

Bucky graduated from the University of Southern California with a B.S in Business Administration, with a concentration in Finance. He serves as a mentor for the Alchemist Accelerator program.

Sid Singh, Vice President of Operations at StartX

Siddhartha (Sid) Singh is the Vice President of Operations at StartX. Prior to that, he was the Vice President & National Head Retail at Citibank, where he created and implemented marketing, revenue enhancement, and customer engagement strategies. Sid received his M.B.A. at the University of Delhi in 2002. He graduated with his M.S. in Management in 2014 at Stanford University Graduate School of Business.

Thursday, May 12, 2016 -
4:30pm to 5:30pm

Speaker: Robert Anderson, UC Berkeley

Title: PCA with Model Misspecification

Abstract: The theoretical justifications for Principal Component Analysis (PCA) typically assume that the data is IID over the estimation window. In practice, this assumption is routinely violated in financial data. We examine the extent to which PCA-like procedures can be justified in the presence of two specific kinds of misspecification present in financial data: time-varying volatility, and the presence of regimes in factor loadings. Joint work with Stephen W. Bianchi (Berkeley).
Bio: Robert Anderson is a Professor of Economics and Mathematics at the University of California Berkeley, and Director of the Coleman Fung Risk Management Research Center. He has been on the Berkeley faculty since 1983. At Berkeley, he has taught calculus, real analysis, mathematics for incoming Economics Ph.D. students, microeconomic theory, and continuous-time finance, among other subjects. His research is in mathematical economics and nonstandard analysis, an alternative formulation of analysis and probability theory derived from mathematical logic. Currently, his work focuses on the economic foundations of continuous-time finance models. He obtained his Bachelor's Degree from the University of Toronto in 1974 and his Ph.D from Yale in 1977, both in Mathematics. After a one-year postdoctoral appointment, he became an Assistant Professor of Economics and Mathematics at Princeton. He was promoted to Associate Professor with tenure in 1982, and came to Berkeley a year later. He was an Alfred P. Sloan Research Fellow from 1982 to 1986, and has been a Fellow of the Econometric Society since 1987. He served two terms as Chair of the Berkeley Economics Department. His involvement with the Academic Senate began in 1992, when he began an effort to provide equal University benefits to the families of lesbian and gay University employees, resulting in the provision of health benefits in 1997 and pension benefits in 2002, with ongoing updates in response to subsequent changes in California family and tax law. He has served as Chair of the University Committee on Faculty Welfare, as parliamentarian of the Berkeley Division of the Academic Senate, and eight years as chair of the Senate's Task Force on Investments and Retirement, serving concurrently as an Academic Senate representative on the University of California Retirement System Advisory Board. Over the last eight years, he has led the Senate's increasingly urgent appeals to take action to ensure the stability of UCRP. He was awarded the Berkeley Faculty Service Award in 2009.



Tuesday, May 10, 2016 -
12:30pm to 1:30pm

Speaker: Christian Linder, Assistant Professor of Civil and Environmental Engineering, Stanford

Title: Stretchability by Design - Advanced Computational Methods to Understand Mechanical Phenomena in Microarchitectured Soft Materials
Abstract: The weakness of conjugated polymers, the basic material for organic semiconductors, is that they are not stretchable. While generally flexible, their stretchability is restricted up to a few percent and at large deformations, cracks can deteriorate electronic device performance. This weakness limits their use in industrial applications that require large stretchability of 100% and new applications demanding complete flexibility. This presentation will show advanced computational simulations to predict the development of a highly stretchable semiconducting material by modeling the blending process of a conjugating semiconducting polymer with an elastomer, thereby triggering nanoconfined morphologies due to phase separation. It will be shown in detail, how advanced computational scale bridging techniques can provide insight on the complex mechanisms involved in that process. In particular, a Cahn Hilliard framework to model morphology evolution in polymer blends, homogenization techniques to understand elastic and inelastic effects in polymers, and a computational framework to account for crystallization based enthalpic effects will be covered in this presentation.
Bio: Professor Christian Linder is the principal investigator of the Micromechanics of Materials Lab at Stanford University. The lab advances modeling aspects, numerical algorithms, and visualization tools to improve the performance and reliability of simulations to (i) understand physical mechanisms in materials, (ii) create innovative sustainable building materials and structures, and (iii) enable upscaled devices and engineered systems of the environment. In-house (iv) computational method development in the area of Computational Mechanics and Computational Materials Science constitutes the foundation of our research. 


Professor Linder received his Ph.D. in Civil and Environmental Engineering from UC Berkeley, an MA in Mathematics from UC Berkeley, an M.Sc. in Computational Mechanics from the University of Stuttgart, and a Dipl.-Ing. degree in Civil Engineering from TU Graz. Before joining Stanford in 2013 he was a Junior-Professor of Micromechanics of Materials at the Applied Mechanics Institute of Stuttgart University where he also obtained his Habilitation in Mechanics. Notable honors include a Fulbright scholarship, the 2013 Richard-von-Mises Prize, and the 2016 NSF CAREER Award.

Monday, May 9, 2016 -
4:30pm to 5:30pm

Speakers: Karianne Bergen and Eileen Martin, ICME

Friday, May 6, 2016 -
4:00pm to 5:00pm

Speaker: Margot Gerritsen, ICME Director

During this talk, you will learn about different strategies on how to improve your presentation skills:
  • How to overcome presentation anxiety - tricks from 30 years experience
  • How to deal with questions
  • How to start and end powerfully
... and much more!
Thursday, May 5, 2016 -
4:30pm to 5:30pm

Speaker: Eric Aldrich, UC Santa Cruz

Title: The Flash Crash: A New Deconstruction

Abstract: On May 6, 2010, in the span of a mere four and half minutes, the Dow Jones Industrial Average lost approximately 1,000 points. In the following fifteen minutes it recovered essentially all of its losses. This “Flash Crash” occurred in the absence of fundamental news that could explain the observed price pattern and is generally viewed as the result of endogenous factors related to the complexity of modern equity market trading. We present the first analysis of the entire order book at millisecond granularity, and not just of executed transactions, in an effort to explore the causes of the Flash Crash. We also examine information flows as reflected in a variety of data feeds provided to market participants during the Flash Crash. While assertions relating to causation of the Flash Crash must be accompanied by significant disclaimers, we suggest that it is highly unlikely that, as alleged by the United States Government, Navinder Sarao’s spoofing orders, even if illegal, could have caused the Flash Crash, or that the crash was a foreseeable consequence of his spoofing activity. Instead, we find that the explanation offered by the joint CFTC-SEC Staff Report, which relies on prevailing market conditions combined with the introduction of a large equity sell order implemented in a particularly dislocating manner, is consistent with the data. We offer a simulation model that formalizes the process by which large sell orders of the sort observed in the CFTC-SEC Staff Report, combined with prevailing market conditions, could generate a Flash Crash in the absence of fundamental information. Our research also documents the emergence of heretofore unobserved anomalies in market data feeds that correlate very closely with the initiation of and recovery from the Flash Crash. Our analysis of these data feed anomalies is ongoing as we attempt to discern whether they were a symptom of the rapid trading that accompanied the Flash Crash or whether they were causal in the sense that they rationally contributed to traders’ decisions to withdraw liquidity and then restore it after the anomalies were resolved.

Bio: Eric Aldrich is an Assistant Professor of Economics at the University of California, Santa Cruz. He holds a B.S. in Economics from Duke University (2002), an M.S. in Statistics from the University of Washington (2005) and a Ph.D. in Economics from Duke University (2011). His research interests include macroeconomic asset pricing, computational economics, financial econometrics and finance.

Thursday, May 5, 2016 -
4:30pm to 5:45pm

Speaker: Ernest Ryu, Stanford

Title: Risk-Constrained Kelly Gambling
Abstract: We consider the classic Kelly gambling problem with general distribution of outcomes, and an additional risk constraint that limits the probability of a drawdown of wealth to a given undesirable level.  We develop a bound on the drawdown probability; using this bound instead of the original risk constraint yields a convex optimization problem that guarantees the drawdown risk constraint holds.  Numerical experiments show that our bound on drawdown probability is reasonably close to the actual drawdown risk, as computed by Monte Carlo simulation.  Our method is parametrized by a single parameter that has a natural interpretation as a risk-aversion parameter, allowing us to systematically trade off asymptotic growth rate and drawdown risk.  Simulations show that this method yields bets that outperform fractional-Kelly bets for the same drawdown risk level or growth rate.
Joint work with Enzo Busseti and Stephen Boyd.
Tuesday, May 3, 2016 -
12:30pm to 1:30pm

Speaker: Lexing Ying, Professor of Mathematics, Stanford

Title: Fast algorithms for physical sciences
Abstract: In this talk, I will give an introduction to fast algorithms for computational problems in physical sciences. I will start with a brief discussion of the fast multipole method and then highlight some key ideas in other important examples.
Bio: Lexing Ying received his PhD from New York University in 2004 and was a postdoctoral scholar at Caltech from 2004 to 2006. From 2006 to 2012, he was a professor at The University of Texas at Austin. Since late 2012, he is a professor at Stanford University. His main research area is computational mathematics and scientific computing. He has received the Sloan Foundation Research Fellowship in 2007, the NSF Career award from National Science Foundation in 2009, the Feng Kang prize in Scientific Computing from Chinese Academy of Sciences in 2011, and the James H. Wilkinson Prize in Numerical Analysis and Scientific Computing from Society for Industrial and Applied Mathematics (SIAM) in 2013.
Friday, April 29, 2016 -
4:00pm to 5:00pm

Speaking Skills for Scientists
Every speaking opportunity is a chance to engage and inform your audience about your research and findings. Creating and delivering a compelling talk is a learned skill, combining audience engagement, clear visual design of information, and public speaking techniques. Come to this Friday's TGIF to learn how to make the most of your audience's attention.

Instructor: After completing her Ph.D in Immunology at Stanford, Trisha Stan became a fellow in the Program for Writing and Rhetoric. She focuses on helping scientists and future scientists find clear and compelling ways to share their ideas. She is also a co-founder and producer of the science news podcast Goggles Optional.


Thursday, April 28, 2016 -
4:30pm to 5:45pm

Speaker: Ruoyu Sun, Stanford

Bio: Ruoyu Sun is a Post Doctoral Scholar in the Department of Managment Science and Engineering at Stanford University, working on large-scale optimization (a.k.a. big data optimization) and low-rank structure learning (with application in, e.g. recommendation systems). Ruoyu's research interests lie in optimization, signal processing, machine learning, information theory, wireless communications and their intersections. He completed his Ph.D. in Electrical Engineering at the University of Minnesota in Spring 2015. He received the B.Sc. degree in mathematics from Peking University, Beijing, China in 2009.

Tuesday, April 26, 2016 -
12:30pm to 1:30pm

Speaker: Steven Scott, Senior Economic Analyst at Google

Title: Scaling Bayesian learning through consensus Monte Carlo

Abstract: A useful definition of “big data” is data that is too big to comfortably process on a single machine, either because of processor, memory, or disk bottlenecks. Graphics processing units can alleviate the processor bottleneck, but memory or disk bottlenecks can only be eliminated by splitting data across multiple machines. Communication between large numbers of machines is expensive (regardless of the amount of data being communicated), so there is a need for algorithms that perform distributed approximate Bayesian analyses with minimal communication. Consensus Monte Carlo operates by running a separate Monte Carlo algorithm on each machine, and then averaging individual Monte Carlo draws across machines. Depending on the model, the resulting draws can be nearly indistinguishable from the draws that would have been obtained by running a single machine algorithm for a very long time. Examples of consensus Monte Carlo are shown for simple models where single-machine solutions are available, for large single-layer hierarchical models, and for Bayesian additive regression trees (BART).

Bio: Steven Scott is a Senior Economic Analyst at Google, where he has worked since 2008.  He received his PhD from the Harvard statistics department in 1998. He spent 9 years on the faculty of the Marshall School of Business at the University of Southern California.  Between USC and Google he also had a brief tenure at Capital One, where he was a Director of Statistical Analysis. Dr. Scott is a Bayesian statistician specializing in Monte Carlo computation.  In his academic life he has written papers on Bayesian methods for hidden Markov models, multinomial logistic regression, item response models, support vector machines. These methods have been applied to network intrusion detection, web traffic modeling, educational testing, health state monitoring, and brand choice, among others.  Since joining Google he has focused on models for time series with many contemporaneous predictors, on scalable Monte Carlo computation, and on Bayesian methods for the multi-armed bandit problem. 

Monday, April 25, 2016 -
4:30pm to 5:30pm

More details to follow.

Friday, April 22, 2016 -
4:00pm to 5:00pm

Presenter: Matt Vassar, Technical Communication ProgramStanford

Description: If a person walks up to your poster and you have only 30 seconds to grab their attention, what would you say? In this fast-paced and interactive workshops, Matt Vassar will give you the tips to interest anybody in your research in a very short amount of time. You will learn to make your research engaging to any audience and to convey a lot of meaning in a very short statement.

Bio: Matt Vassar has worked his entire professional life helping researchers to make their work fascinating to audiences of all sorts. He teaches engineers, mathematicians, and scientists to make their research interesting and informative. He currently works in Stanford's Technical Communication Program, and teaches ENGR 103: Public Speaking.


Thursday, April 21, 2016 -
4:30pm to 5:30pm

Speaker: Charles-Albert Lehalle, Capital Fund Management

Title: Optimal Trading

Abstract: We will go from the role of the financial system described as a large network of intermediaries to a fine description of high frequency market makers. The role of regulation in the recent transformations of participants practices will be exposed too. The viewpoint taken is the one of a practitioner or a researcher who has to put in place models. Existing models will be reviewed, and new challenges and the stakes of possible improvements will be discussed. Important stylized facts and important mechanisms that models should reproduce will be exposed.

Thursday, April 21, 2016 -
4:30pm to 5:45pm

Speaker: Anil Damle, Stanford

Title: Sparse representations and fast algorithms for Kohn-Sham orbitals
Abstract: Kohn-Sham density functional theory is the most widely used electronic structure theory for molecules and systems in condensed phase. The Kohn-Sham orbitals (a.k.a. Kohn-Sham wavefunctions) are eigenfunctions of the Kohn-Sham Hamiltonian and are generally delocalized, i.e. each orbital has significant magnitude across the entire computational domain. Given a set of Kohn-Sham orbitals from an insulating system, it is often desirable to build a set of localized basis functions for the associated subspace. In this talk we present a simple, robust, and parallelizable algorithm to construct a set of (optionally orthogonal) localized basis functions known as the selected columns of the density matrix (SCDM). In addition, we will discuss recently developed variants of the SCDM algorithm that drastically reduce the computational cost while maintaining the quality of the basis.
Bio: Anile Damle is a PhD candidate in the Institute for Computational and Mathematical Engineering (ICME) at Stanford University. His general interests include numerical linear algebra, non-linear approximations, matrix analysis, and fast algorithms for structured matrices. Anil's current research projects focus on localization of Kohn-Sham orbitals, updating of certain tree based matrix factorizations, and non-negative matrix factorizations. Visit Anil's personal webpage at:


Tuesday, April 19, 2016 -
12:30pm to 1:30pm

Speaker: Celso Ferreira, George Mason University

Title: Hurricane storm surge and waves attenuation by nature-based coastal defenses in the Chesapeake Bay

Abstract: Hurricane Sandy demonstrated the vulnerability of the East Coast to extreme events causing wide spread damage and highlighting the need for resilient coastal defenses. The potential of natural wetlands to attenuate storm surge and dampen wave energy has been investigated in many laboratory and numerical studies, however fewer field experiments exist to validate or quantify these processes in a natural environment. Current research indicates that the capacity of wetlands to attenuate storm surge and waves is highly dependent not only on spatial scales but also seasonally dependent vegetation biomechanics, micro-topography, groundwater levels and storm characteristics. We are currently performing a 2-year field campaign to investigate storm surge and wave attenuation in 4 protected areas in the Chesapeake Bay. Our study sites range from lower bay areas with semi-direct contact to the ocean (Eastern Shore of Virginia Wildlife Refuge), mid-bay areas (Dameron Marsh Natural Area Reserve) and upper bay areas in the tidal Potomac. Pressure transducers and Acoustic Doppler Current Profilers (ADCPs) have been permanently deployed to capture time series of water depths and vertical velocity gradients. In collaboration with partners from the United States Geological Survey (USGS), investigations of vegetation characteristics (height, diameter, and stem density) are conducted to allow for improved understanding of the factors contributing to flow-resistance. Additionally, high resolution topo-bathymetric surveys are been conducted to map the complex geomorphology of these areas. The field data set is used to calculate rates of attenuation across marsh transects, and supports a regional storm surge model calibration for an accurate parameterization of coastal wetland vegetation on a regional scale. For this study, storm surge was modeled along the Virginia portion of the Delmarva Peninsula to improve our understanding of the sensitivity of simulated storm surge in these environments to: 1) bottom friction due to wetlands parametrization; 2) bathymetric uncertainty; 3) uniformity of shallow channels; and 4) barrier island connectivity. The coupled hydrodynamic-wave model (ADCIRC+SWAN) was used for this study along with a numerical mesh adapted from the FEMA region III grid. In the adapted mesh, resolution was increased in areas of interest from 30-100m in the FEMA mesh to 10-30 m for this study. Here we also take advantage of a recently developed domain decomposition approach to improve model efficiency and evaluate multiple scenarios while minimizing computational expenses. Hurricane hindcasts and synthetic storms were used to evaluate model sensitivity to friction, topo-bathymetry and configuration. Initial results indicate water level dependency highly correlated to inlet conditions at barrier islands and topographic features of the marsh system within the Back Bay. Results from this study will improve the ability of decision makers to evaluate the value of marsh systems for coastal defense considerations and to improve understanding of the hydrodynamic interaction of barrier island-back bay systems and storm-surge.

Bio: Dr. Celso Ferreira is currently a visiting scholar at Stanford University. He is an Assistant Professor of Water Resources Engineering in the Civil, Infrastructure and Environmental Engineering Department of George Mason University. He is also an Associate Researcher at the USGS National Research Program in Reston, VA. He has a PhD from Texas A&M University in Civil Engineering with a focus on Water Resources Engineering. He has a ME in Hydrology from the CEDEX Institute in Spain, a MS in Environmental Engineering and a BS in Civil Engineering from the Federal University of Santa Catarina in Brazil. His current research interests are associated to water related extreme weather hazards and its impacts to civil engineering infrastructure. His research is currently funded by the National Science Foundation (NSF), the Department of Interior (DOI), the National Fish and Wildlife Foundation (NFWF) and several private organizations. His research group is currently investigating the potential of nature to reduce hurricane flooding impacts to critical infrastructure. His research also strive to incorporate climate variability and human induced environmental changes to engineering practice by considering climate change, sea-level rise, urbanization and environmental degradation impacts on engineering design. His work ranges across spatial scales and the projects range from international to regional and local applications.

Monday, April 18, 2016 -
4:30pm to 5:30pm

Speaker: Yingzhou Li, Stanford

Title: Distributed-memory Hierarchical Interpolative Factorization

Abstract: This work proposes a distributed-memory algorithm for the hierarchical interpolative factorization, which is a novel framework for efficient solution of partial differential equations and integral equations. This algorithm can be applied to most elliptic problems as well as low-to-medium frequency hyperbolic problems. Numerical experiments demonstrate the efficiency of the parallel algorithm.


Thursday, April 14, 2016 -
4:30pm to 5:30pm

Speaker: Shota Ishii, State Street Global Exchange, GX Labs

Title: Data Driven Analysis of Multi-Asset Class Portfolios

Abstract: Longer-horizon, multi-asset class portfolios characterize those of large pensions and sovereign funds, which drive some of the largest flows in the market.  Meanwhile, much of financial theory from the 1950s to the 1990s was driven by examining only the behavior of equities, particularly in the United States.  The rise of high-performance computing using a broader array of market and non-market data is allowing us to create richer models for understanding the sources of performance in these portfolios.  We explore some of the tools, methodologies, and new areas of potential exploration for these complex multi-asset class portfolios through the lens of contemporary computing and data science.
Bio: Shota joined State Street Global Exchange in 2014 to help build out its portfolio analytics and advisory capabilities as well as to work on product strategy, design, and complex data visualization. He also engages in strategic initiatives and business development as part of GXLabs, a new innovation center formed in San Francisco to exploit the recent advances in massive data management and high-performance computing as applied to finance. Prior to State Street, he worked at DCI, a systematic long/short investment manager based in San Francisco specializing in corporate credit. His other experiences include running a social knowledge management start-up venture in Paris and consulting extensively on allocation and risk as regional manager for Asia at Moody’s KMV, a quantitative credit risk advisory firm, in Asia and in the United States for regulators, large banks, asset managers, and insurance companies. He has also consulted extensively for the Financial Strategies group at Shinsei Bank in Tokyo. He has a BS from Cornell University in applied physics and an MBA from INSEAD.  
Thursday, April 14, 2016 -
4:30pm to 5:45pm

Speaker: Alan Karp, HP Labs, Ret.

Title: Patenting Division or How I Ran Afoul of the US Supreme Court

Abstract: Trying to patent division sounds like the action of a crazy person. Nevertheless, I succeeded, not once, but three times. Two of those patents are actually legitimate. In this talk, I'll explain what we did, why some of that is worth patenting and why one of those patents is explicitly ruled out by a US Supreme Court decision. Along the way, I'll show you more about computer arithmetic than you care to know, explain the big part serendipity played in the work, and give a case study on the value of quality refereeing. As an added bonus, I'll even throw in a bit of mathematics.

Bio: Alan Karp received his PhD in Astronomy from the University of Maryland, spent two years at IBM Research modeling exploding stars, and one year as an assistant professor of physics at Dartmouth College before joining IBM's Palo Alto Scientific Center to work on vector and parallel computing.He moved to Hewlett-Packard Laboratories, where he became one of the architects of the chips in Intel's Itanium line.  He then joined HP's newly formed E-speak Operation to productize the technology he helped develop at HP Labs, later returning to HP Labs to work on usable security and subsequently joining HP's Enterprise Services organization as an architect for enterprise-scale distributed systems.

Dr Karp has served on the editorial boards of the Journal of Quantitative Spectroscopy and Radiative Transfer, the Journal of Transport Theory and Statistical Physics, and the editorial advisory board of the journal Scientific Programming.  Dr Karp chaired the committee judging the entries for the Gordon Bell Prize for parallel processing for its first 10 years. He was awarded two IBM Outstanding innovation awards and holds over 70 patents.
Tuesday, April 12, 2016 -
12:30pm to 1:30pm

Speaker: Juan Jose Alonso, Professor of Aeronautics and Astronautics, Stanford

Bio: Professor Alonso's work is focused on research and development of new high-fidelity, multidisciplinary methods and techniques for the analysis and design of complex aerospace systems. He is interested in the development of these methods and their use in realistic test cases. Past and current research includes transonic, supersonic, and hypersonic vehicles, rotorcraft, turbomachinery, and launch vehicles. Specific current interests include advanced methods for design, multi-fidelity optimization, environmentally-friendly aircraft, uncertainty quantification and robust design, and system-level challenges for NextGen.

Monday, April 11, 2016 -
4:30pm to 5:30pm

More details to follow.

Friday, April 8, 2016 -
4:00pm to 5:00pm

Presenter: Kelly Harrison, Technical Communication Program, Stanford

At this workshop, you'll learn the standard practices for designing a conference poster. You'll discover how to tell your story, what to put on your poster, and how to balance text, images, fonts, and colors. Come prepared for an interactive seminar! This workshop will be conducted by Kelly A. Harrison, an instructor in the Technical Communication Program in the School of Engineering at Stanford University. She teaches technical communication (ENGR 202W) and independent writing (ENGR 202S). Before academia, Kelly worked in the computer software industry, studied aviation, and trained as a pilot and aircraft mechanic.

Thursday, April 7, 2016 -
4:30pm to 5:30pm

Speaker: Pierre Spatz, Murex

Title: Effects of GPU, AAD and XVA on the Future Computing Architecture of Banks

Abstract: The 2008 crisis has tremendously changed the way we approach financial computing in the banks. While the complexity and diversity of traded products have been reduced, volumes and regulatory computations needs have exploded while budgets became tight and we do not see any relief in the future. Several solutions including GPU– powerful parallel coprocessor – , AAD– an algorithm – or both of them have been implemented to cope with today workload. All these methods imply at least a partial rewrite of the code. We will come back on our experience and see how well each solution fit different test cases with current or future hardware and extrapolate how the future calculation servers of banks will look like.

Bio: Pierre Spatz heads the quantitative analysis team of Murex, a world leader in trading and risk management software. He holds a master’s degree in computer engineering and applied mathematics from ENSIMAG in Grenoble, France.

Thursday, April 7, 2016 -
4:30pm to 5:45pm

Speaker: Christine Klymko, Lawrence Livermore National Laboratory

Title:Detection of highly-cyclic communities in directed networks

Abstract: Many large, real-world complex network have rich community structure that a network scientist seeks to understand. These communities may overlap or have intricate internal structure. Extracting communities with particular topological structure, even when they overlap with other communities, is a powerful capability that would provide novel avenues of focusing in on structure of interest. In this work we consider extracting highly-cyclic regions of directed graphs (digraphs). We demonstrate that embeddings derived from complex-valued eigenvectors associated with stochastic propagator eigenvalues near roots of unity are well-suited for this purpose. We prove several fundamental theoretic results demonstrating the connection between these eigenpairs and the presence of highly-cyclic structure and we demonstrate the use of these vectors on a few real-world examples.
Bio: Christine Klymko is currently a postdoctoral researcher at the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory. She received her PhD in Computational Mathematics from the Department of Mathematics and Computer Science at Emory University in 2014. During her doctoral studies, she spent summers at Oak Ridge National Laboratory and Sandia National Laboratories. Her research interests include numerical linear algebra, graph algorithms, data mining, matrix analysis, machine learning, and scientific computing.
Tuesday, April 5, 2016 -
12:30pm to 1:30pm

Speaker: Yinyu Ye, Stanford

Title: Multi-Block ADMM and its Convergence

Abstract: We show that the direct extension of alternating direction method of multipliers (ADMM) with three blocks is not necessarily convergent even for solving a square system of linear equations, although its convergence proof was established 40 years ago with one or two blocks. However, we prove that, in each iteration if one randomly and independently permutes the updating order of variable blocks followed by the regular multiplier update, then ADMM will converge in expectation when solving any system of linear equations with any number of blocks. This is probably the first theoretical evidence for applying random permutation in computational optimization, where empirical results have shown the effectiveness of random permutation in either ADMM or block coordinate descent method. We also discuss its extension to solve general convex optimization problems.

Bio: Yinyu Ye is currently the Kwoh-Ting Li Professor in the School of Engineering at the Department of Management Science and Engineering and Institute of Computational and Mathematical Engineering and the Director of the MS&E Industrial Affiliates Program, Stanford University. He received the B.S. degree in System Engineering from the Huazhong University of Science and Technology, China, and the M.S. and Ph.D. degrees in Engineering-Economic Systems and Operations Research from Stanford University. Ye's research interests lie in the areas of optimization, complexity theory, algorithm design and analysis, and applications of mathematical programming, operations research and system engineering. He is also interested in developing optimization software for various real-world applications. Current research topics include Liner Programming Algorithms, Markov Decision Processes, Computational Game/Market Equilibrium, Metric Distance Geometry, Dynamic Resource Allocation, and Stochastic and Robust Decision Making, etc. He is an INFORMS (The Institute for Operations Research and The Management Science) Fellow, and has received several research awards including the winner of the 2014 SIAG/Optimization Prize awarded every three years to the author(s) of the most outstanding paper, the inaugural 2012 ISMP Tseng Lectureship Prize for outstanding contribution to continuous optimization, the 2009 John von Neumann Theory Prize for fundamental sustained contributions to theory in Operations Research and the Management Sciences, the inaugural 2006 Farkas prize on Optimization, and the 2009 IBM Faculty Award. He has supervised numerous doctoral students at Stanford who received received the 2015 and 2013 Second Prize of INFORMS Nicholson Student Paper Competition, the 2013 INFORMS Computing Society Prize, the 2008 Nicholson Prize, and the 2006 and 2010 INFORMS Optimization Prizes for Young Researchers. Ye teaches courses on Optimization, Network and Integer Programming, Semidefinite Programming, etc. He has written extensively on Interior-Point Methods, Approximation Algorithms, Conic Optimization, and their applications; and served as a consultant or technical board member to a variety of industries, including MOSEK.

Thursday, March 31, 2016 -
4:30pm to 5:45pm

Speaker: Joseph Grcar

Title: Episodic History of Calculations for Exterior Ballistics

Abstract: Exterior ballistics studies the trajectories of bullets and artillery shells, or in Leonard Euler's phrase, the path of an object thrown into the air.  Military technologists predicted trajectories using Isaac Newton's model of a point-mass with aerodynamic drag from Euler in 1753 to ENIAC in 1946.  For this purpose they invented various computing methods: ballistic tables, numerical integration, differential analyzers, and finally electronic computers.  Indeed, the United States supported inventing computers precisely to construct artillery firing tables in the second World War.  Beyond the many fascinating historical events, this long period offers a test for understanding the history of engineering technology.  I propose a concept of episodic change in the history of engineering analogous to the paradigm shifts in the history of science.

Bio: Joseph Grcar conducts research on the history of computing.  His 2010 article in Historia Mathematica identified Newton as the inventor of "Gaussian" elimination in Europe.  Grcar was a computational scientist at national laboratories in Livermore and in Berkeley for many years, where he was one of the original developers of the Chemkin software for chemical kinetics.  He specialized in calculations for reacting fluid flows, combustion, and linear algebra in which he originated the partial reorthogonalization technique for Lanczos methods.  A matrix and a polynomial are named after him.  Grcar received a doctorate in mathematics from the University of Illinois in 1980.

Tuesday, March 8, 2016 -
12:30pm to 1:30pm

Speaker: Markus Pelger

Title: Large Dimensional Factor Modeling Based on High-Frequency Observations. 
Abstract: This paper develops a statistical theory to estimate an unknown factor structure based on financial high-frequency data. I derive a new estimator for the number of factors and derive consistent and asymptotically mixed-normal estimators of the loadings and factors under the assumption of a large number of cross-sectional and high-frequency observations. The estimation approach can separate factors for normal “continuous” and rare jump risk. The estimators for the loadings and factors are based on the principal component analysis of the quadratic covariation matrix. The estimator for the number of factors uses a perturbed eigenvalue ratio statistic. The results are obtained under general conditions, that allow for a very rich class of stochastic processes and for serial and cross-sectional correlation in the idiosyncratic components. 
Bio: Markus Pelger is an Assistant Professor at the Management Science & Engineering Department at Stanford University. His research interests are in statistics, financial econometrics, asset pricing and risk management. Markus received his Ph.D. in Economics from the University of California, Berkeley. He has a Diplom in Mathematics and a Diplom in Economics from the University of Bonn in Germany.
Thursday, March 3, 2016 -
4:30pm to 5:45pm

Speaker: Austin Benson

Title: The Spacey Random Walk: a Stochastic Process for Higher-order Data
Abstract: Recent work on eigenvalues of hypermatrices and tensors has generated an algebraic analogue of the stationary distribution vector for a Markov chain. We show that this tensor eigenvector corresponds to the stationary distribution of a new stochastic process called a spacey random walk; it is a hybrid of a higher-order Markov chain and a vertex-reinforced random walk. Our insight provides a solid probabilistic foundation for these tensor eigenvectors, their interpretation, and their application to data problems with higher-order structure.
Tuesday, March 1, 2016 -
12:30pm to 1:30pm

Speaker: Johan Ugander

About the Speaker: Johan Ugander is an Assistant Professor of Management Science and Engineering. His research develops algorithmic and statistical frameworks for analyzing social networks, social systems, and other large-scale social data. His work commonly falls at the intersections of graph theory, probability theory, statistics, optimization, and algorithm design. Johan obtained his Ph.D. in Applied Mathematics from Cornell University in June 2014, advised by Jon Kleinberg. From 2010-14, he held an affiliation with the Facebook Data Science team. For the 2014-15 academic year, he was a post-doctoral researcher at Microsoft Research, hosted by Eric Horvitz. Johan joined Stanford in September 2015, where he is a member of MS&E's Social Algorithms Lab (SOAL) and affiliated with the Center for Computational Social Science.

Monday, February 29, 2016 -
4:30pm to 5:30pm

Speaker: Austin Benson