Skip to:

Past Events

View events by calendar year
Enter one or more keywords to search for
Filter by event type
Thursday, September 25, 2014 -
4:15pm to 5:30pm

Speaker: Dr. Ofer Levi, Department of Industrial Engineering and Management, Ben-Gurion University of the Negev, Beer Sheva, Israel

Laplace inversion of LR-NMR relaxometry data using sparse representation methods
LR-NMR relaxometry is a powerful tool that can be harnessed for characterizing constituents in complex materials. The technology is used for industrial quality control to measure solid-to-liquid and oil-to-water ratios in materials as diverse as oil-bearing rock, food emulsions, and plant seeds.

Conversion of the relaxation signal into a continuous distribution of relaxation components is an ill-posed problem.  We provide a numerical optimization method for analyzing LR-NMR data by including L1 regularization and applying the convex optimization solver PDCO.  Our integrated approach includes validation of analyses by simulations, testing repeatability of experiments, and validation of the model and its statistical assumptions.  The method provides better resolved and more accurate solutions than those suggested by existing tools.

Wednesday, September 24, 2014 -
4:15pm to 5:30pm

Speaker: Stephan Ludwig

Change of financial networks through central clearing and collateralized banking
The introduction of central clearing through the Dodd-Frank act in the US and EMIR in the EU changes the structure of inter banking networks and risk management elementary. Driven by the shock of bank defaults during the financial crises, regulators implemented stabilizing components to absorb single defaults and enable more transparency. This talk presents a few but fundamental changes through central clearing. The focus is to introduce the method of margining, collateral management and default funds in order to ask the question: how does central clearing impacts our financial stability and the cost for it. The goal is to provide input for creating new network models for one of the most-discussed topic in today’s financial markets. 

Tuesday, September 23, 2014 -
12:15pm to 1:15pm

Speaker: TBD

Thursday, September 18, 2014 -
9:00am to 3:45pm

 

Sign-up now for the ICME Summer Refresher!  This review session spans four days and covers basic mathematical material relevant to first-year ICME classes and applied mathematics topics in science and engineering. The ICME Summer Refresher is suitable for incoming ICME graduate students and other graduate students with a technical background. This year, we are offering courses in numerical linear algebra, partial differential equations, statistics and probability, and discrete mathematics and algorithms. Students can enroll in any or all of these courses. 

Details on each session are available here:  http://stanford.edu/~arbenson/refresher.html

Register here:  https://docs.google.com/forms/d/1WrW2jMjre8qbznN-x2LUxxygN2b6TOq9WdxK9-f...

Wednesday, September 17, 2014 -
9:00am to 3:45pm

 

Sign-up now for the ICME Summer Refresher!  This review session spans four days and covers basic mathematical material relevant to first-year ICME classes and applied mathematics topics in science and engineering. The ICME Summer Refresher is suitable for incoming ICME graduate students and other graduate students with a technical background. This year, we are offering courses in numerical linear algebra, partial differential equations, statistics and probability, and discrete mathematics and algorithms. Students can enroll in any or all of these courses. 

Details on each session are available here:  http://stanford.edu/~arbenson/refresher.html

Register here:  https://docs.google.com/forms/d/1WrW2jMjre8qbznN-x2LUxxygN2b6TOq9WdxK9-f...

Tuesday, September 16, 2014 -
9:00am to 3:45pm

 

Sign-up now for the ICME Summer Refresher!  This review session spans four days and covers basic mathematical material relevant to first-year ICME classes and applied mathematics topics in science and engineering. The ICME Summer Refresher is suitable for incoming ICME graduate students and other graduate students with a technical background. This year, we are offering courses in numerical linear algebra, partial differential equations, statistics and probability, and discrete mathematics and algorithms. Students can enroll in any or all of these courses. 

Details on each session are available here:  http://stanford.edu/~arbenson/refresher.html

Register here:  https://docs.google.com/forms/d/1WrW2jMjre8qbznN-x2LUxxygN2b6TOq9WdxK9-f...

Monday, September 15, 2014 -
9:00am to 3:45pm

 

Sign-up now for the ICME Summer Refresher!  This review session spans four days and covers basic mathematical material relevant to first-year ICME classes and applied mathematics topics in science and engineering. The ICME Summer Refresher is suitable for incoming ICME graduate students and other graduate students with a technical background. This year, we are offering courses in numerical linear algebra, partial differential equations, statistics and probability, and discrete mathematics and algorithms. Students can enroll in any or all of these courses. 

Details on each session are available here:  http://stanford.edu/~arbenson/refresher.html

Register here:  https://docs.google.com/forms/d/1WrW2jMjre8qbznN-x2LUxxygN2b6TOq9WdxK9-f...

Tuesday, June 17, 2014 - 8:00am to Friday, June 20, 2014 - 4:00pm

Online registration for the 2014 Workshop on Algorithms for Modern Massive Data Sets (MMDS 2014) is now available at:

http://mmds-data.org/home/registration2014

In addition to the talks, there will be a poster session one evening.
You may apply to present a poster at the link above.

Event: MMDS 2014: Workshop on Algorithms for Modern Massive Data Sets
Dates: June 17-20, 2014
Location: UC Berkeley, Berleley, CA
Website: http://mmds-data.org
Contact: organizers@mmds-data.org

Synopsis: The 2014 Workshop on Algorithms for Modern Massive Data Sets (MMDS 2014) will address algorithmic, mathematical, and statistical challenges in modern statistical data analysis. The goals of MMDS 2014 are to explore novel techniques for modeling and analyzing massive, high-dimensional, and nonlinearly-structured scientific and internet data
sets, and to bring together computer scientists, statisticians, mathematicians, and data analysis practitioners to promote cross-fertilization of ideas.

Organizers: Michael Mahoney (UC Berkeley), Alex Shkolnik (Stanford), Petros Drineas (RPI), Reza Zadeh (Stanford), Fernando Perez (UC Berkeley)

Friday, June 6, 2014 -
1:00pm to 4:00pm

Stanford’s new visualization center, the HIVE, opens to the Stanford Community in June. 

Join us for an open house any time between 1:00- 4:00pm on Friday, June 6 for the HIVE Xpo.  Come see the HIVE in action and learn more about this brand-new collaborative visualization center. 

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 for research visualization projects by Stanford faculty, students, and staff.

The HIVE features a 10-foot-tall, 24-feet-wide video wall with 13440x5400 resolution and 72 million total active pixels.  Thirty-five 46-inch ultra-narrow bezel backlit monitors in a 7 X 5 concaved configuration provide optimal visual stimuli for all research projects. The equipment will allow collaborative visualization in teaching and research across the sciences, social sciences, and engineering.  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.

Monday, June 2, 2014 -
4:15pm to 5:30pm

Speaker: Joel E. Cohen, Laboratory of Populations, Rockefeller University and Columbia University, New York

Abrupt biotic change associated with smooth environmental change: predictions from theories of Taylor's power law of fluctuation scaling
Recent paleoclimatic and paleoecological studies and the 2013 National Academy of Sciences report on Abrupt Impacts of Climate Change: Anticipating Surprises highlight the need to understand better, and develop early warnings of, abrupt changes in the environment and abrupt changes in biota associated with smooth changes in the environment. I will describe recently discovered theoretical examples that predict that smooth changes in the environment can produce abrupt changes (infinite jumps) in a key parameter of Taylor's power law of fluctuation scaling, one of the most widely verified empirical patterns in ecology. A comparable real-world singularity could adversely affect fisheries, forestry, agriculture, conservation, and public health.
 
Sources:
Cohen, J.E., 2014. Taylor's law and abrupt biotic change in a smoothly changing environment. Theoretical Ecology 7:77-86, http://dx.doi.org/10.1007/s12080-013-0199-z.
Cohen, J.E., 2014. Stochastic population dynamics in a Markovian environment implies Taylor's power law of fluctuation scaling. Theoretical Population Biology 93:30-37. DOI:10.1016/j.tpb.2014.01.001
National Research Council (NRC) 2013 Abrupt Impacts of Climate Change: Anticipating Surprises. National Academies Press, Washington, DC. ISBN 978-0-309-28773-9; 250 pp.

 

Wednesday, May 28, 2014 -
4:15pm to 5:00pm

Speaker: Govind Menon

More details to come.

Wednesday, May 21, 2014 -
4:15pm to 5:00pm

Speaker: Gautam Iyer

More details to come.

Monday, May 19, 2014 -
4:15pm to 5:30pm

Speaker: Harley McAdams, Department of Developmental Biology at Stanford University

An essential, conserved embedded logic that directs cell operations
Every cell in every organism has a core logic circuitry that directs the overall functioning of the cell.  Many cell models and simulations have been developed over the last two decades.  They all share problems of daunting complexity and limitations in their experimental underpinnings.  We can surmount these problems in the design of extraordinarily complex engineering systems such as commercial aircraft and communications satellites, but it remains a challenge for the biological cell.  We will discuss the nature of the “essential” logic of a bacterial cell – that is the minimal regulatory logic necessary for viability – and suggest that this essential logic is relatively small in combinational complexity and that it can be almost completely characterized in specific engineering detail with current laboratory technology.  Our goal is to develop an engineering quality simulation of the embedded logic (comprising about a dozen transcription factor proteins) that is necessary to make the ~400 essential genes of the bacterium Caulobacter crescentus into a viable living entity.  The project involves ongoing CHiP-seq experiments to identify the complete set of essential transcriptional pathways and recently completed experiments that that have identified all essential segments of the bacterial DNA, including DNA encoding essential proteins, DNA segments containing essential regulatory binding sites or other regulatory features, all regulatory RNAs (and their cell cycle profiles), plus the transcriptional start sites on the genome to single basepair resolution.  Conventional engineering analysis of the regulatory logic circuitry will elucidate logic design rules underlying evolutionary selection of the regulatory network.  We also seek better understanding of circuit features underlying robust operation of the cell in the face of wide random variations in the timing of operation of critical cellular subsystems.
Thursday, May 15, 2014 -
8:00am to 5:00pm

Registration is now open for the Big Data for Energy and Environment Spring 2014 Conference of Stanford’s Energy and Environment Affiliates ProgramThursday May 15, 2014, 8:00 a.m.-5:00 p.m. at Stanford University, with a welcome reception the prior evening.  Massive data and sophisticated analytical techniques are leading to new opportunities in fields such as electricity generation and distribution, manufacturing, oil and gas, and sustainability.  This conference will present cutting edge research in tools, techniques, and applications.  The agenda will be posted online shortly. Click here to register.

Big Data for Energy and Environment

General data science techniques and sensors

Margot Gerritsen, Data science and analytics

Phil Levis, Embedded sensing systems

Oil and gas exploration and production

Biondo Biondi, Permanent seismic arrays

Jef Caers, Multiple-point geostatistics

Environment and climate change

Noah Diffenbaugh, Climate and earth dynamics

Mary Ruckelshaus, Earth Genome Project

Annie Hazlehurst and Sara Menker, Gro Intelligence

Electric grid, energy efficiency, and social networks

Balaji Prabhakar, Societal networks

Ram Rajagopal, Electric grid analytics

Amit Narayan, AutoGrid: Finding business value from energy analytics

The Energy and Environment Affiliates Program at Stanford University is a membership organization that facilitates interactions between companies and other organizations, and Stanford's faculty and graduate students, across the full range of energy-related and environmental topics. Visit our website at http://eeap.stanford.edu/

Monday, May 12, 2014 -
4:15pm to 5:30pm

Speaker: Henry van den Bedem, SLAC National Accelerator Laboratory

Now You See Me, Now You Don’t: Structurally Characterizing Biomolecular Dynamics from Kinematics
Large biomolecules, such as RNA, proteins and their complexes, fluctuate between conformational substates to mediate their cellular activity. Their functional mechanisms often manifest themselves precisely in transient populations of different conformers. Adequate exploration of conformational space to identify these substates, starting from static crystal structures, remains challenging. We have developed techniques based on the theory of kinematically redundant manipulators to efficiently probe the conformational diversity of biomolecules, ranging from small RNAs to large protein complexes. Our procedure projects degrees of freedom onto a subspace of conformational space defined by distance constraints in the tertiary structure, such as hydrogen bonds. The dimensionality reduction enables efficient exploration of conformational space and reduces the risk of overfitting sparse experimental data. Our procedure can predict coupled substates that coincide with long-range NMR chemical shift perturbations of an allosteric mutation in the enzyme ecDHFR. When coupled with NMR residual dipolar coupling data for the small RNA HIV1 trans-activation response element stem-loop, it revealed a hypothesized transient excited state of the apical binding loop. Our computational procedure thus enables an integrative view of NMR data, which can provide the tools to formulate and test dynamic, motion-based hypotheses of functional mechanisms in protein and RNA.

 

Saturday, May 10, 2014 -
11:00am to 6:30pm

Above and Beyond

Stanford welcomes you to the world “Above and Beyond” with our third TEDxStanford, Saturday, May 10. 
 
Whether you’ve attended our previous events or are new to TEDxStanford, we are delighted to invite you to TEDxStanford2014: Above and Beyond. Join us for a day of inspirational short talks and performances by scholars that go above and beyond the ordinary. 
From a deep dive with extreme creatures of the sea and the new synthetic biology that “teaches” our cells to target cancer, to the creation of the solar car, stripping down Wall Street and a sneak peak at the documentary film made in the world’s “happiest place,” TEDxStanford2014 is all about imagination and exploration without bounds.
 
Go “Above and Beyond” on May 10. The main auditorium is already sold out, but additional overflow seating is available. All details are at https://tedx.stanford.edu/2014. Watch the free livestream.
Friday, May 9, 2014 -
9:00am to 5:00pm

The Stanford Research Computing Center (SRCC) and the Institute for Computational & Mathematical Engineering (ICME) are pleased to sponsor a one-day workshop on Intel® Cluster Studio XE 2013. The workshop will be led by Intel software engineers. They will provide comprehensive and in-depth coverage of Intel(R) MPI and Cluster tools, Parallelism for Intel Processors & Coprocessors, Intel Performance and threading analysis tools, and Intel performance libraries.

The workshop is intended for both beginning and experienced users of Intel architectures and software development tools for both applications and software development to become familiar with the Intel architectures and software development tools, and make the most efficient uses of Intel architectures. Some knowledge of C/C++/Fortran programming languages, MPI programming, and performance profiling and tuning is assumed but not required. 

For the workshop, attendees are expected to bring their own laptops with the necessary software already installed. Detailed instructions will be sent to participants before the workshops.

Space is limited~ register early (before May 7) at <https://www.eventbrite.com/e/intel-cluster-studio-xe-2013-training-workshop-tickets-11263047077>

Wednesday, May 7, 2014 -
10:00am to 3:00pm

MATLAB Technical Seminars at Stanford – May 6 and 7

Join us for complimentary MATLAB seminars at Stanford University.  Faculty, academic researchers and students are all welcome to attend.

Dates:                                  Tuesday and Wednesday, May 6 and 7, 2014

Location:                             Turing Auditorium, Polya Hall

Sessions:                         1. Analyzing and Visualizing Data with MATLAB

                                            2. Parallel Computing with MATLAB to Perform Large-Scale Simulations

                                            3. Data Mining and Machine Learning with MATLAB

                                            4. Image Processing and Computer Vision with MATLAB

Tuesday, May 6, 2014

10:00 a.m. – 12:00 p.m.:  Analyzing and Visualizing Data with MATLAB

Learn how to analyze and visualize data, develop algorithms and applications to automate your workflow, and how to publish and share your results.

1:00 – 3:00 p.m.:  Parallel Computing with MATLAB to Perform Large-Scale Simulations and Data Processing Tasks

Learn how to solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters with very few changes to your code.

To view complete session descriptions and register for Tuesday’s session, visit:  www.mathworks.com/stanford/day1

Wednesday, May 7, 2014

10:00 a.m. – 12:00 p.m.:  Data Mining and Machine Learning with MATLAB

Learn several data mining and machine learning techniques available in MATLAB to quickly explore your data, compare the results and apply the best solutions to your problem.

1:00 – 3:00 p.m.:  Image Processing and Computer Vision

Learn how to perform image analysis, object detection and tracking, feature detection and extraction, feature matching, stereo vision, camera calibration and motion detection tasks.

To view complete session descriptions and register for Wednesday’s session, visit:  www.mathworks.com/stanford/day2

 

Tuesday, May 6, 2014 -
10:00am to 3:00pm

MATLAB Technical Seminars at Stanford – May 6 and 7

Join us for complimentary MATLAB seminars at Stanford University.  Faculty, academic researchers and students are all welcome to attend.

Dates:                                  Tuesday and Wednesday, May 6 and 7, 2014

Location:                             Turing Auditorium, Polya Hall

Sessions:                         1. Analyzing and Visualizing Data with MATLAB

                                            2. Parallel Computing with MATLAB to Perform Large-Scale Simulations

                                            3. Data Mining and Machine Learning with MATLAB

                                            4. Image Processing and Computer Vision with MATLAB

Tuesday, May 6, 2014

10:00 a.m. – 12:00 p.m.:  Analyzing and Visualizing Data with MATLAB

Learn how to analyze and visualize data, develop algorithms and applications to automate your workflow, and how to publish and share your results.

1:00 – 3:00 p.m.:  Parallel Computing with MATLAB to Perform Large-Scale Simulations and Data Processing Tasks

Learn how to solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters with very few changes to your code.

To view complete session descriptions and register for Tuesday’s session, visit:  www.mathworks.com/stanford/day1

Wednesday, May 7, 2014

10:00 a.m. – 12:00 p.m.:  Data Mining and Machine Learning with MATLAB

Learn several data mining and machine learning techniques available in MATLAB to quickly explore your data, compare the results and apply the best solutions to your problem.

1:00 – 3:00 p.m.:  Image Processing and Computer Vision

Learn how to perform image analysis, object detection and tracking, feature detection and extraction, feature matching, stereo vision, camera calibration and motion detection tasks.

To view complete session descriptions and register for Wednesday’s session, visit:  www.mathworks.com/stanford/day2

 

Monday, May 5, 2014 -
4:15pm to 5:30pm

Speaker: Markus Covert, Department of Bioengineering at Stanford University

Towards Whole-Cell Models of Health and Disease
Understanding how complex phenotypes arise from individual molecules and their interactions is a primary challenge in biology that computational approaches are poised to tackle. We report a whole-cellcomputational model of the life cycle of the human pathogen Mycoplasma genitalium that includes all of its molecular components and their interactions. An integrative approach to modeling that combines diverse mathematics enabled the simultaneous inclusion of fundamentally different cellular processes and experimental measurements. Our whole-cell model accounts for all annotated gene functions and was validated against a broad range of data. The model provides insights into many previously unobserved cellular behaviors, including in vivo rates of protein-DNA association and an inverse relationship between the durations of DNA replication initiation and replication. In addition, experimental analysis directed by model predictions identified previously undetected kinetic parameters and biological functions. We conclude that comprehensive whole-cell models can be used to facilitate biological discovery.
Monday, April 28, 2014 -
4:15pm to 5:30pm

Speaker: Michael Mascagni, Professor of Computer Science at Florida State University

Novel Stochastic Methods in Biochemical Electrostatics
Electrostatic forces and the electrostatic properties of molecules in solution are among the most important issues in understanding the structure and function of large biomolecules. The use of implicit-solvent models, such as the Poisson-Boltzmann equation (PBE), have been used with great success as a way of computationally deriving electrostatics properties such molecules. We discuss how to solve an elliptic system of partial differential equations (PDEs) involving the Poisson and the PBEs using path-integral based probabilistic, Feynman-Kac, representations. This leads to a Monte Carlo method for the solution of this system which is specified with a stochastic process, and a score function. We use several techniques to simplify the Monte Carlo method and the stochastic process used in the simulation, such as the walk-on-spheres (WOS) algorithm, and an auxiliary sphere technique to handle internal boundary conditions. We then specify some optimizations using the error (bias) and variance to balance the CPU time. We show that our approach is as accurate as widely used deterministic codes, but has many desirable properties that these methods do not. In addition, the currently optimized codes consume comparable CPU times to the widely used deterministic codes. Thus, we have an very clear example where a Monte Carlo calculation of a low-dimensional PDE is as fast or faster than deterministic techniques at similar accuracy levels.

Thursday, April 24, 2014 -
4:15pm to 5:30pm

Speaker: Randall J. LeVeque, Professor of Applied Mathematics, University of Washington

Using Models to Understand Geophysical Hazards

Geophysical hazards unfold over a wide range of length and time scales. Numerical simulations of such events can uniquely probe these variable scales and allow us to query the connections between multiple processes. However, developing accurate and efficient simulation techniques is challenging. Prof. LeVeque will discuss new approaches for tsunami modeling and probabilistic hazard assessment as a case study, and also briefly describe several related applications to modeling storm surge, debris flows and seismic waves. He will also give a brief overview of the open source Clawpack software he developed for solving general hyperbolic systems of conservation laws and modeling wave propagation, and of GeoClaw, a specialized version for depth-averaged modeling of flows over topography.

Professor Leveque received his PhD in computer science from Stanford in 1982. He is the lead developer of the Clawpack soft ware package and the author of three books on numerical methods for differential equations. His current research interests are focused on algorithm and soft ware development, reproducible research and applications of hyperbolic solvers in geophysics and biophysics. He is a Fellow of the Society for Industrial and Applied Mathematics and of the American Mathematical Society.

This lecture is part of the School of Earth Sciences Distinguished Lecture Series.   

Click here to watch this lecture on YouTube.

Wednesday, April 23, 2014 -
4:15pm to 5:00pm

Speaker: Randall LeVeque

More details to come.

Monday, April 21, 2014 -
4:15pm to 5:30pm

Speaker: Noah Rosenberg, Department of Biology at Stanford University

Mathematical Properties of Statistics Commonly Used in Population Genetics
Many analyses of population-genetic data begin with computations of summary statistics from the frequencies of alleles in populations, relying for biological interpretations on a large body of theory concerning the behavior of these statistics in models of evolutionary processes.  Often, however, these statistics can generate peculiar results that arise not from biological sources, but from mathematical properties of the statistics themselves.  We discuss how counterintuitive features of population-genetic summary statistics can be produced by an often-unexpected or underappreciated phenomenon that the mathematical upper and lower bounds on the statistics can vary with aspects of the allele or frequencies.  To reduce the potential for misinterpretations, we have determined the bounds on a variety of population-genetic statistics in relation to other such statistics, with a focus on measures of genetic diversity and genetic differentiation. Knowledge of these bounds can facilitate sensible biological understanding, and in some instances, useful normalized statistics. We examine how the mathematical bounds on population-genetic statistics provide insight into potentially counterintuitive phenomena observed in such contexts as human population relationships, tests of population-genetic null models, the distribution of rare genetic variants, and the detection of “soft” selective sweeps.
Wednesday, April 16, 2014 -
4:15pm to 5:00pm

Speaker: Georg Menz

A two scale proof of the Eyring-Kramers formula
 
We consider a diffusion on a potential landscape which is given by a smooth Hamiltonian in the regime of small noise. We give a new proof of the Eyring-Kramers formula for the spectral gap of the associated generator of the diffusion. The proof is based on a refinement of the two-scale approach introduced by Grunewald, Otto, Villani, and Westdickenberg and of the mean-difference estimate introduced by Chafai and Malrieu. The Eyring-Kramers formula follows as a simple corollary from two main ingredients : The first one shows that the Gibbs measures restricted to a domain of attraction has a "good" Poincaré constant mimicking the fast convergence of the diffusion to metastable states. The second ingredient is the estimation of the
mean-difference by a new weighted transportation distance. It contains the main contribution of the spectral gap, resulting from exponential long waiting times of jumps between metastable states of the diffusion. This new approach also allows to derive sharp estimates on the log-Sobolev constant.
 
Monday, April 14, 2014 -
4:15pm to 5:30pm

Speaker: Bo Huang, Professor of Pharmaceutical Chemistry at UCSF

Pixels or Coordinates? Computational Methods in Super-resolution Microscopy
In super-resolution microscopy methods based on single-molecule switching (STORM/PALM), each camera frame samples a random, sparse subset of probe molecules in the samples. The final super-resolution image is assembled with the molecule coordinates extracted from thousands of frames. We have developed a sparse-signal recovery technique using compressed sensing to analyze camera images with highly overlapping fluorescent spots. This method allows one camera image to sample an order of magnitude more fluorophores, thus improving the temporal resolution of live imaging. In order to increase the total duration of observation limited by fluorophore photobleaching, we utilized the redundant information from adjacent snapshots in a time sequence. With a deformable registration algorithm, we have effectively reduced the number of camera frames required for super-resolution image reconstruction. Finally, we developed a framework for correlation analysis of coordinate-based super-resolution images. We have shown a variety of applications of this framework, including image alignment, single-molecule tracking and the quantification of colocalization.
 
Friday, April 11, 2014 -
8:30am to 5:30pm

 Please join us on Friday April 11 for the 4th Stanford Conference for Computational Social Science jointly hosted by The Institute for Computational & Mathematical Engineering, directed by Margot Gerritsen; The Center for the Study of Language and Information, directed by Chris Potts; and The Center for Computational Social Science, directed by Dan McFarland. Topics at the spring conference will include Diffusion of Sentiment and Community Building Online, Computation and Markets, Health and Computational Social Science, and Social Movements and Social Media. Lunch will be provided for all confirmed attendees. Reception to follow.

 

Read details and register at https://css-center.stanford.edu/4th-conference-computational-social-science

Wednesday, April 9, 2014 -
3:45pm to 5:30pm

Speaker: Thomas Chen and Josselin Garnier

3:45-4:30pm, 380C
Josselin Garnier, Universite Paris VII 
Time-reversal focusing and correlation-based imaging with moving sensors
 
It is now well known that the Green's function between two receivers can be estimated from the cross correlation of the signals emitted by ambient noise sources and recorded by the receivers. It is also well known that a reflector can be imaged by migration of the cross correlation matrix of the signals emitted by ambient noise sources and
recorded by a passive receiver array. In this talk we extend these results to situations in which the sensors are moving. We show that it is possible to use moving sensors to generate large synthetic apertures. We also exhibit a surprising super-resolution phenomenon when the sensor velocities become of the order of the speed of propagation.
 
4:35-5:20pm, 380C
Thomas Chen, University of Texas, Austin 
Interacting Bose gases, Gross-Pitaevskii hierarchies, and the quantum de Finetti theorem
 
This talk surveys some recent results related to the derivation of the cubic nonlinear Schroedinger equation in R^3, and the related Cauchy problem for Gross-Pitaevskii (GP) hierarchies. Moreover, a new approach to the analysis of GP hierarchies using the quantum de Finetti theorem is presented, based on which we obtain a new proof of
unconditional uniqueness of solutions, as well as a proof of scattering in the defocusing case (joint work with C. Hainzl, N. Pavlovic and R. Seiringer).
 
Monday, April 7, 2014 -
4:15pm to 5:30pm

Speaker: Leor Weinberger, Professor of Biochemistry and Biophysics at UCSF

Viral gene circuits:  simple model systems and a new spectrum of antiviral targets
Viral gene circuits are bounded by many of the same design constraints as electrical circuits (e.g., signal-to-noise ratio) but rapidly mutating viruses, like HIV, evolved unique mechanisms to circumvent and even exploit these constraints.  These adaptations are essential to the evolutionary fitness of viruses and present novel targets for antiviral therapy.  I will describe two viral circuit adaptations, the first in HIV (see Weinberger & Weinberger, Cell, 2013) and the second in the human herpesvirus CMV (Teng et al., Cell, 2012), the leading cause of birth defects and transplant rejections.  I will show how disrupting and exploiting these circuit architectures presents a new therapy spectrum.
Friday, April 4, 2014 -
10:00am to 5:00pm

Announcing a full-day workshop on the high-speed cluster programming framework, Spark. We will discuss Machine Learning and Matrix Computations, along with other projects in the Spark platform such as Shark and MLLib, and new programming tools for big data. The workshop will include introductions to the many Spark features, case studies from current users, best practices for deployment and tuning, future development plans, and hands-on exercises for Amazon EC2. 

Please register now: http://spark-icme.eventbrite.com

Friday April 4th, 10am-5pm

Clark Center Auditorium

Organized by Databricks and Stanford ICME. 

More details on Spark: http://spark.incubator.apache.org

Workshop Schedule:  stanford.edu/~rezab/sparkworkshop/

 

Tuesday, April 1, 2014 -
3:15pm to 4:00pm

Speaker: Albert Fannjiang, University of California, Davis

Phase retrieval with roughly known mask

Fourier phasing is the problem of retrieving Fourier phase information from Fourier intensity data. The standard Fourier phase retrieval (without a mask) is known to have many solutions which cause the standard phasing algorithms to stagnate and produce wrong or
inaccurate solutions. In this talk Fourier phase retrieval is carried out with the introduction of a randomly fabricated mask in measurement and reconstruction. Highly probable uniqueness of solution, up to a global phase, was previously proved with exact knowledge of the
mask. Here the uniqueness result is extended to the case where only rough information about the mask’s phases is assumed. The exponential probability bound for uniqueness is given in terms of the uncertainty-to-diversity ratio of the unknown mask. New phasing
algorithms alternating between the object update and the mask update are systematically tested and demonstrated to have the capability of recovering both the object and the mask (within the object support) simultaneously, consistent with the uniqueness result. Phasing with a
phase-uncertain mask is shown to be robust with respect to the correlation in the mask as well as the Gaussian and Poisson noises.
 
The seminar webpage with more abstracts is

 

Monday, March 31, 2014 -
4:15pm to 5:00pm

Speaker: Dr. Joseph Grcar

Computational Science and Engineering History and Bibliometry

"Numerical analysis", "scientific computing", and "computational science and engineering" all refer to more or less the same subject, yet since the invention of modern computers, these names have been successively introduced for what were described as important new fields.  The history of computing before and after the invention of computers offers some explanation for the repeated reintroduction of the field.  Fewer than a dozen types of major calculations were done before computers.  Some lessons about scientific computing as a
profession can be learned by examining historical calculations in two areas: "multiple correlations" in statistics and "range tables" in ballistics.
 
Presently, research experts for the modern subject are most numerous first in Europe, then second in the United States, and third in China. Although the subject originated here, recent bibliometric data suggest that computational science and engineering may not have been best
integrated into the curricula of American universities.
 
Bio:
Dr. Grcar has written several articles on the history and sociology of science focussing on mathematical sciences and making use of bibliometric data.  Formerly he was a research scientist specializing in reacting fluid flow (combustion) calculations at Sandia National
Lab and at Lawrence Berkeley National Lab.  His expertise in numerical analysis is matrix calculations.  A test matrix and a polynomial preconditioner are named after him.
Monday, March 17, 2014 - 2:15pm

Emmanuel Candès, The Barnum-Simons Chair in Math and Statistics, cordially invites you to the

Math + X Symposium at Stanford University

Monday, March 17, 2014

2:15 p.m. - 5:00 p.m.

Click here to register -- Sign up early as space is limited

The Math+X Program at Stanford University was established in 2012 with funding from the Simons Foundation and additional support from William and Donnalisa Barnum. Its goals are to strengthen the ties between mathematics and statistics and to expand interactions with other disciplines in which the application of mathematics promises to achieve novel results. 

The Math + X Symposium features:

  • Welcome and Introduction to the Math + X Program,
    • Emmanuel Candès, The Barnum-Simons Chair in Math and Statistics
  • “Mathematical Foundations for Quantum Engineering”
    • Hideo Mabuchi, professor of applied physics, and Nikolas Tezak, graduate student in applied physics
  • “Analyzing Large Data Sets from Brain Imaging Studies in Freely Behaving Mice”
    • Mark Schnitzer, associate professor of biology and of applied physics, and Lacey Kitch, graduate student in electrical engineering
  • “Gene Networks Analysis and Density Estimation”
    • Wing Wong, The Stephen R. Pierce Family Goldman Sachs Professor in Science and Human Health, and Michael Kun Yang, graduate student in computational and mathematical engineering
  • “Math + Computation”
    • ​​Lester Mackey, assistant professor of statistics
Wednesday, March 12, 2014 -
4:15pm to 5:15pm

Birth and Future of Multi-Scale Modeling of Macromolecules

Dr. Michael Levitt

The Robert W. and Vivian K. Cahill Professor in Cancer Research in the Department of Structural Biology and Professor, by courtesy, of Computer Science, Dr. Michael Levitt was Awarded the Nobel Prize in Chemistry in 2013 "for the development of multi-scale models for complex chemical systems."

The development of multiscale models for complex chemical systems began in 1967 with publications by Warshel and Levitt recently recognized by the 2013 Nobel Committee for Chemistry.  The simplifications used then at the dawn of the age of computational structural biology were mandated by computers that were almost a billion times less cost-effective than those we use today.  These same multiscale models have become increasingly popular in applications that range from  simulation of atomic protein motion to protein folding and explanation of enzyme catalysis.  

In this talk, Dr. Levitt will describe the origins of computational structural biology and then go on to show some of the most exciting current and future applications.

 

students, faculty, and staff are welcome to attend.  No RSVP is needed. 

Tuesday, March 11, 2014 -
12:30pm to 2:00pm

Jun Mei (Department of Physics, South China University of Technology)
"Dark Acoustic Metamaterials"

A thin-film acoustic metamaterial, comprising an elastic membrane decorated with rigid platelets, is able totally absorb low-frequency sound at selective resonance frequencies ranging from 100-1,000 Hz, with the relevant sound wavelength in air being three orders of magnitude larger than the membrane thickness.  At resonances, the flapping motion of the rigid platelets leads naturally to large elastic curvature energy density at their perimeter regions, and forms in essence an open cavity.

Ying Wu (Division of Computer, Electrical and Mathematical Science and Technology)
"Linear Dispersion Relations in Periodic Structures"

Recently, linear dispersion relations in periodic structures for classical waves have attracted much attention. In this talk, I will introduce our recent work on various types of linear dispersion relations in photonics and phononic crystals.  We developed a k.p
method based on perturbation to study the properties of the system in the region when linear dispersion happens, including the slope, the effective Hamiltonian, and the Berry phase.  For certain cases, we also derived an effective medium theory to reveal the link between the linear dispersion and zero-refrantive-index materials.

Monday, March 10, 2014 -
4:15pm to 5:00pm

Speaker: Ernest Ryu

Stochastic Proximal Iteration: A Non-Asymptotic Improvement Upon Stochastic Gradient Descent

Many large scale machine learning problems give rise to large scale optimization problems. In solving these problems, the extremely simple algorithm, stochastic gradient descent (SGD), is often the method of choice because it performs well, while alternatives from traditional optimization such as batch gradient descent or Newton methods are computationally infeasible.

In this talk, we will first start off by discussing the restraints of these massive scale problems and arguing that the SGD-like (first order) methods are not only computationally feasible but also theoretically desirable. We will then make a simple observation that reveals the sub-optimality of SGD: in many machine learning problems, the learner is provided with random functions, but SGD uses only its  gradient and not the entire function.

To address this issue, we present an algorithm which we call stochastic proximal iteration (SPI). Each iterate of SPI is obtained by applying the proximal mapping with respect to the given random function to the previous iterate. This makes SPI an online algorithm with a computational cost comparable to that of SGD in certain interesting settings. Using machinery from monotone operator theory, we show that SPI has an asymptotic performance equivalent to SGD but does better in the non-asymptotic regime.

Friday, March 7, 2014 -
4:00pm to 5:00pm

This Week's Speaker: Wing H. Wong, Professor of Statistics, of Health Research and Policy and, by courtesy, of Biology at Stanford University

Wing H. Wong

This seminar series is designed for a student audience. The idea of the seminar is for faculty to describe their research, their philosophy or specific projects in broad terms at a level that students can easily follow. The style is informal to encourage a dialog rather than a one-way flow. Computer projectors are rarely used except to display images. Also the lack of faculty in the audience enables the student to be the ones asking the questions. The less preparation the presenter makes the better. Anything that is not completely familiar to the speaker is not something that the students need to know.Since this is informal for the speaker we hope that there will a large number of faculty that will respond to request that they speak. We also extend the list of speakers to include senior students with a particular focus on those who no longer sit in ICME space and do not have an ICME adviser. The class is also encouraged to make suggestions for speakers and can if they wish invite speakers.

Thursday, March 6, 2014 -
4:15pm to 5:15pm

Special Linear Algebra and Optimization seminar

 

Prof Per Christian Hansen

Section for Scientific Computing

DTU Compute

Technical University of Denmark

http://www.imm.dtu.dk/~pch

 

Image Deblurring with Krylov Subspace Methods

 

Image deblurring, i.e., reconstruction of a sharper image from a blurred and noisy one, involves the solution of a large and very ill-conditioned system of linear equations, and regularization is needed in order to compute a stable solution. Krylov subspace methods are often ideally suited for this task: their iterative nature is a natural way to handle such large-scale problems, and the underlying Krylov subspace provides a convenient mechanism to regularize the problem by projecting it onto a low-dimensional "signal subspace" adapted to the particular problem. In this talk we consider the three rylov subspace methods CGLS, MINRES, and GMRES. We describe their regularizing properties, and we discuss some computational aspects such as preconditioning and stopping criteria.

Wednesday, March 5, 2014 -
4:15pm to 5:15pm

Wednesday, March 5, 4:15pm in 380C

Liliana Borcea, University of Michigan

Electromagnetic wave propagation in random waveguides

I will present a study of long range propagation of electromagnetic waves in random waveguides with perfectly conducting boundaries. The waveguide is filled with an isotropic linear dielectric material, with randomly fluctuating electric permittivity. The fluctuations are weak, but they cause significant cumulative scattering over long distances of propagation of the waves. We analyze the wave field by decomposing it in evanescent transverse electric and magnetic modes with random amplitudes that encode the cumulative scattering effects. They satisfy a coupled system of stochastic differential equations driven by the random fluctuations of the electric permittivity. We analyze the solution of this system with the diffusion approximation theorem, under the assumption that the fluctuations decorrelate rapidly in the range direction. The result is a detailed characterization of the transport of energy in the waveguide, the loss of coherence of the modes and the depolarization of the waves due to cumulative scattering. I will explain how we can use such theory for imaging in random waveguides.

Monday, March 3, 2014 -
4:15pm to 5:00pm

Speaker: Xiangrui Meng, Software Engineer at Databricks

 

Scalable Simple Random Sampling Algorithms

 

Analyzing data sets of billions of records has now become a regular task in many companies and institutions. In the statistical analysis of those massive data sets, sampling generally plays a very important role. In this talk, we will present some scalable algorithms for simple random sampling with and without replacement, which use probabilistic thresholds to decide on the fly whether to accept, reject, or wait-list an item independently of others while maintaining high success rates. These algorithms are built from extremely simple ideas and extend naturally to applications such as stratified sampling, bootstrapping, and generating random tuples. The proposed algorithms, when implemented on big data platforms like MapReduce and Spark, can effectively reduce the size of intermediate output and greatly improve load balancing. Empirical evaluation on large-scale data sets clearly demonstrates their superiority.

Friday, February 28, 2014 -
4:00pm to 5:00pm

This Week's Speaker: Jenny Suckale, Assistant Professor of Geophysics at Stanford University

Jenny Suckale

This seminar series is designed for a student audience. The idea of the seminar is for faculty to describe their research, their philosophy or specific projects in broad terms at a level that students can easily follow. The style is informal to encourage a dialog rather than a one-way flow. Computer projectors are rarely used except to display images. Also the lack of faculty in the audience enables the student to be the ones asking the questions. The less preparation the presenter makes the better. Anything that is not completely familiar to the speaker is not something that the students need to know.Since this is informal for the speaker we hope that there will a large number of faculty that will respond to request that they speak. We also extend the list of speakers to include senior students with a particular focus on those who no longer sit in ICME space and do not have an ICME adviser. The class is also encouraged to make suggestions for speakers and can if they wish invite speakers.

Wednesday, February 26, 2014 -
4:15pm to 5:15pm
Wednesday, February 26, 4:15pm in 380C

Bedros Afeyan

 

Nonlinear Optics of Plasmas: Mathematical Opportunities and Challenges

 

High intensity laser-matter interactions give rise to new nonlinear kinetic states of self-organization in plasmas. These nonlinear coherent structures pose challenges and present opportunities for applied mathematical models that combine non-equilibrium statistical mechanics, nonlinear dynamics, plasma physics, statistical optics and nonlinear ultrafast optics. Combining these disciplines to make sense of matter in extreme conditions will require new insights and a close collaboration between microscopic computational approaches, mesoscopic reduced models and macroscopic observable predictions. Attention will be given to the physics background and the numerical approaches popular in the field of high energy density plasmas. The outstanding, compelling applied mathematical questions will be highlighted throughout the presentation.

Monday, February 24, 2014 -
4:15pm to 5:00pm

Speaker: Shuang-hong Yang

Research Scientist at Twitter

Building topic channels on Twitter

Twitter has become a global town hall where people from all over the world come to the platform to communicate or discover all sorts of information in real-time. As the volume of user-generated contents soars, it has become critical to effectively organize contents to facilitate content creation, discovery and consumption. In this talk, I present a collection of data processing, data mining and machine learning algorithms we used in a deployed system that organize a continuous stream of noisy and sparse tweets into hundreds of topic channels with stringently high precision.

Friday, February 21, 2014 -
4:00pm to 5:00pm

This Week's Speaker: Biondo Biondi

Stanford University

This seminar series is designed for a student audience. The idea of the seminar is for faculty to describe their research, their philosophy or specific projects in broad terms at a level that students can easily follow. The style is informal to encourage a dialog rather than a one-way flow. Computer projectors are rarely used except to display images. Also the lack of faculty in the audience enables the student to be the ones asking the questions. The less preparation the presenter makes the better. Anything that is not completely familiar to the speaker is not something that the students need to know.Since this is informal for the speaker we hope that there will a large number of faculty that will respond to request that they speak. We also extend the list of speakers to include senior students with a particular focus on those who no longer sit in ICME space and do not have an ICME adviser. The class is also encouraged to make suggestions for speakers and can if they wish invite speakers.

Wednesday, February 19, 2014 -
4:15pm to 5:15pm
Wednesday, February 19, 4:15pm in 380C

Guillaume Bal, Columbia University

High-contrast high-resolution coupled physics imaging modalities

Abstract: Several recent coupled-physics medical imaging modalities aim to combine a high-contrast, low-resolution, modality with a high-resolution, low-contrast, modality and ideally offer high-contrast, high-resolution, reconstructions. Such modalities often involve the reconstruction of constitutive parameters in partial differential equations (PDE) from knowledge of internal functionals of the parameters and PDE solutions.  This talk presents results of uniqueness, stability and explicit reconstructions for several hybrid inverse problems. In particular, we provide explicit characterizations of what can (and cannot) be reconstructed and offer optimal (elliptic) stability estimates for a large class of coupled-physics imaging modalities including Magnetic Resonance Elastography, Transient Elastography, Photo-Acoustic Tomography and Ultrasound Modulation Tomography. Numerical simulations confirm the high-resolution, high-contrast, potential of these novel modalities.

Monday, February 17, 2014 -
4:15pm to 5:00pm

NO Speaker, Campus is closed for Presidents' Day.  

Friday, February 14, 2014 -
4:00pm to 5:00pm

This Week's Speaker: Patrick Hayden, Professor of Physics at Stanford University

Patrick Hayden

This seminar series is designed for a student audience. The idea of the seminar is for faculty to describe their research, their philosophy or specific projects in broad terms at a level that students can easily follow. The style is informal to encourage a dialog rather than a one-way flow. Computer projectors are rarely used except to display images. Also the lack of faculty in the audience enables the student to be the ones asking the questions. The less preparation the presenter makes the better. Anything that is not completely familiar to the speaker is not something that the students need to know.Since this is informal for the speaker we hope that there will a large number of faculty that will respond to request that they speak. We also extend the list of speakers to include senior students with a particular focus on those who no longer sit in ICME space and do not have an ICME adviser. The class is also encouraged to make suggestions for speakers and can if they wish invite speakers.

 

 
Wednesday, February 12, 2014 -
4:15pm to 5:15pm
Wednesday, February 12, 4:15pm in 380C

Matt Lorig, Princeton University

Pricing Varience Swaps on Time-Changed Markov Process

Often, different financial derivatives are subject to the same sources of risk. In such cases, the prices of these derivatives ought to be related. In this talk, we will explore the relationship between a variance swap, which is a financial derivative whose payoff depends on the entire path of a stock S over a fixed time interval [0,T], and a European contract, which is a financial derivative whose payoff depends only on the value of S at the terminal time T. We will prove that, when a stock is modeled as a time-changed Markov process, the variance swap has the same value as a European contract whose payoff function is the solution of an integro-differential equation (which we will solve). The significance of this result is that the path-dependent variance swap contract can be priced relative to liquidly traded (and efficiently priced) path-independent European options in a semi-nonparametric fashion.

Universal properties related to the KPZ equation is an extremely active research field right now. It is connected to many different areas in mathematics and mathematical physics: probability theory and statistical mechanics, PDEs, random matrices, integrable systems, and many others. While it was a big recent progress, mainly in deriving exact formulas for correlation functions, the problem of universality is still largely open. We shall discuss some recent results in this direction, as well as possible approaches to establishing of the full universality.Universal properties related to the KPZ equation is an extremely active research field right now. It is connected to many different areas in mathematics and mathematical physics: probability theory and statistical mechanics, PDEs, random matrices, integrable systems, and many others. While it was a big recent progress, mainly in deriving exact formulas for correlation functions, the problem of universality is still largely open. We shall discuss some recent results in this direction, as well as possible approaches to establishing of the full universality.Universal properties related to the KPZ equation is an extremely active research field right now. It is connected to many different areas in mathematics and mathematical physics: probability theory and statistical mechanics, PDEs, random matrices, integrable systems, and many others. While it was a big recent progress, mainly in deriving exact formulas for correlation functions, the problem of universality is still largely open. We shall discuss some recent results in this direction, as well as possible approaches to establishing of the full universality.Universal properties related to the KPZ equation is an extremely active research field right now. It is connected to many different areas in mathematics and mathematical physics: probability theory and statistical mechanics, PDEs, random matrices, integrable systems, and many others. While it was a big recent progress, mainly in deriving exact formulas for correlation functions, the problem of universality is still largely open. We shall discuss some recent results in this direction, as well as possible approaches to establishing of the full universality.Universal properties related to the KPZ equation is an extremely active research field right now. It is connected to many different areas in mathematics and mathematical physics: probability theory and statistical mechanics, PDEs, random matrices, integrable systems, and many others. While it was a big recent progress, mainly in deriving exact formulas for correlation functions, the problem of universality is still largely open. We shall discuss some recent results in this direction, as well as possible approaches to establishing of the full universality.

Monday, February 10, 2014 -
4:15pm to 5:00pm

Febryary 10th seminar is cancelled.

Friday, February 7, 2014 -
4:00pm to 5:00pm

This Week's Speaker: Kay Giesecke

Stanford University, MS&E

This seminar series is designed for a student audience. The idea of the seminar is for faculty to describe their research, their philosophy or specific projects in broad terms at a level that students can easily follow. The style is informal to encourage a dialog rather than a one-way flow. Computer projectors are rarely used except to display images. Also the lack of faculty in the audience enables the student to be the ones asking the questions. The less preparation the presenter makes the better. Anything that is not completely familiar to the speaker is not something that the students need to know.Since this is informal for the speaker we hope that there will a large number of faculty that will respond to request that they speak. We also extend the list of speakers to include senior students with a particular focus on those who no longer sit in ICME space and do not have an ICME adviser. The class is also encouraged to make suggestions for speakers and can if they wish invite speakers.

 

 

 

 

Pages