Solving Ill-Posed Problems in Scientific Computing

By By J. William Bell

May 5, 2006

It's only fair that a puzzle have only one solution, right? That's the great thing about a crossword or one of those Sudoku number games that the lifestyle pages can't stop talking about right now. There's only one way to fill in the boxes and get a complete answer.

Well, nobody said science was always fair. In what are known as ill-posed problems, there is no unique solution. A slight change in the data fed into the system of functions that rule a given ill-posed problem can produce a large, unpredictable change in the results.

“In the late 1800s, a scientist named Hadamard proposed that ill-posed problems didn't exist [or those that did weren't scientifically significant]. He was totally wrong. They're everywhere,” says Rebecca Hartman-Baker, who recently completed her PhD in computer science at the University of Illinois at Urbana-Champaign.

They're found, for example, in medical imaging, financial modeling, environmental modeling, and astronomy. Though Hartman-Baker's PhD thesis focused on an ill-posed problem found in the field of geoprospecting, the approach applies to any of those fields.

“Any ill-posed problem for which you have an educated guess of where to start could use this,” says Hartman-Baker, now a post-doc at Oak Ridge National Laboratory. “As a scientific computing type, the thing that brings me the most joy is to contribute to real-world scientific problems through my little thing. I want to get people who didn't know they needed to be involved in computer science to be in it.”

Making hundreds of runs on NCSA's Platinum and Tungsten clusters, she developed a selection method for choosing the parameters that go into the problem and an optimization method for finding the ideal result among a sea of solutions.

Geoprospecting typically involves placing transmitters and receivers deep in the earth. The transmitter projects electromagnetic energy, or in some cases sound, that is picked up by the receiver miles away. The electromagnetic energy is altered in transit, based on the conductivity of the rock, water, oil, or other materials in the ground. From the data collected by the receiver, researchers can deduce what lies between it and the transmitter.

The challenge is — and this is where the ill-posed problem comes in — that different sizes, shapes, and orientations of underground deposits can produce the same data profile.

“Basically, you're trying to find the size and shape of an ellipse [that represents the deposit that a geoprospector might be targeting, such as oil]. Where the center is at. The rotation or orientation. How you do that is really indirect — kinda backhanded,” Hartman-Baker says.

The traditional method of solving this sort of ill-posed problem, known as Tikhonov regularization, gives a blurry picture of this ellipse. It stabilizes the problem around a single solution by adding things like smoothness constraints to the functions. But in situations like this, researchers tend to prefer distinct boundaries. To get these boundaries, Hartman-Baker proposed another class of stabilizing, known as selection methods. With selection methods, the solution is limited to some reasonable set of possible solutions, and the parameters fed into the problem are reduced to a manageable number (about 10 in the case of Hartman-Baker's work). These decisions limit the computational expense of solving the problem and provide a distinct ellipse.

“We're bringing selection methods back into the world. People have forgotten about them,” she says. In her numerous runs on NCSA's systems, she tested the viability of the approximate quasisolution method.

The selection method reduces the number of parameters and shrinks the size of the solution space, but the optimal solution still has to be found. To do so, researchers make an educated guess of the input that will produce the optimal solution, run a simulation using that input, and compare the output to real-world data. They repeat the process until they find the input that most closely resembles the real-world output. In the example used by Hartman-Baker, that reveals the shape of the ellipse that produces the output.

To find the output that most closely resembled the real world, Hartman-Baker's example required finding the global minimum for a system of mathematical functions. After again trying several different methods, she settled on the diffusion equation method (DEM), which finds local minima and then traces back through a series of those local minima to the global minimum. Hartman-Baker compares this process to finding the lowest point in Kentucky. It's easy to find the lowest point in a region; you just head downhill until you are forced to start uphill again. But there are a lot of regional low points spread about, so ensuring that you've found the lowest of the low is no small task. DEM blurs the function “so that at first you just see overall trends, but not all the details. Then as it progresses, you unblur the function more and more until the full features [of the global minimum that you homed in on] are visible,” she explains.

As part of her thesis work and again using time at NCSA, Hartman-Baker discretized this involved method so that it can be parallelized and can use multiple processors at the same time. Currently, the evaluation of a function can be run on eight processors and takes about a minute. This has to be done several thousand times to solve a single optimization problem in geoprospecting.

This research was supported by the National Computational Science Alliance and, more recently, the UIUC computer science department's Fulton Watson Copp Endowed Chair held by Professor Michael Heath.

—–

Source: Access Online, NCSA

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

Hyperion: AI-driven HPC Industry Continues to Push Growth Projections

November 21, 2019

Three major forces – AI, cloud and exascale – are combining to raise the HPC industry to heights exceeding expectations. According to market study results released this week by Hyperion Research at SC19 in Denver, Read more…

By Doug Black

At SC19: Bespoke Supercomputing for Climate and Weather

November 20, 2019

Weather and climate applications are some of the most important uses of HPC – a good model can save lives, as well as billions of dollars. But many weather and climate models struggle to run efficiently in their HPC en Read more…

By Oliver Peckham

Microsoft, Nvidia Launch Cloud HPC Service

November 20, 2019

Nvidia and Microsoft have joined forces to offer a cloud HPC capability based on the GPU vendor’s V100 Tensor Core chips linked via an InfiniBand network scaling up to 800 graphics processors. The partners announced Read more…

By George Leopold

Hazra Retiring from Intel Data Center Group, Successor Not Known

November 20, 2019

Rajeeb Hazra, corporate VP of Intel’s Data Center Group and GM for the Enterprise and Government Group, is retiring after more than 24 years at the company. At this writing, his successor is unknown. An earlier story on... Read more…

By Doug Black

Jensen Huang’s SC19 – Fast Cars, a Strong Arm, and Aiming for the Cloud(s)

November 20, 2019

We’ve come to expect Nvidia CEO Jensen Huang’s annual SC keynote to contain stunning graphics and lively bravado (with plenty of examples) in support of GPU-accelerated computing. In recent years, AI has joined the s Read more…

By John Russell

AWS Solution Channel

Making High Performance Computing Affordable and Accessible for Small and Medium Businesses with HPC on AWS

High performance computing (HPC) brings a powerful set of tools to a broad range of industries, helping to drive innovation and boost revenue in finance, genomics, oil and gas extraction, and other fields. Read more…

IBM Accelerated Insights

Data Management – The Key to a Successful AI Project

 

Five characteristics of an awesome AI data infrastructure

[Attend the IBM LSF & HPC User Group Meeting at SC19 in Denver on November 19!]

AI is powered by data

While neural networks seem to get all the glory, data is the unsung hero of AI projects – data lies at the heart of everything from model training to tuning to selection to validation. Read more…

SC19 Student Cluster Competition: Know Your Teams

November 19, 2019

I’m typing this live from Denver, the location of the 2019 Student Cluster Competition… and, oh yeah, the annual SC conference too. The attendance this year should be north of 13,000 people, with the majority attende Read more…

By Dan Olds

Hyperion: AI-driven HPC Industry Continues to Push Growth Projections

November 21, 2019

Three major forces – AI, cloud and exascale – are combining to raise the HPC industry to heights exceeding expectations. According to market study results r Read more…

By Doug Black

At SC19: Bespoke Supercomputing for Climate and Weather

November 20, 2019

Weather and climate applications are some of the most important uses of HPC – a good model can save lives, as well as billions of dollars. But many weather an Read more…

By Oliver Peckham

Hazra Retiring from Intel Data Center Group, Successor Not Known

November 20, 2019

Rajeeb Hazra, corporate VP of Intel’s Data Center Group and GM for the Enterprise and Government Group, is retiring after more than 24 years at the company. At this writing, his successor is unknown. An earlier story on... Read more…

By Doug Black

Jensen Huang’s SC19 – Fast Cars, a Strong Arm, and Aiming for the Cloud(s)

November 20, 2019

We’ve come to expect Nvidia CEO Jensen Huang’s annual SC keynote to contain stunning graphics and lively bravado (with plenty of examples) in support of GPU Read more…

By John Russell

Top500: US Maintains Performance Lead; Arm Tops Green500

November 18, 2019

The 54th Top500, revealed today at SC19, is a familiar list: the U.S. Summit (ORNL) and Sierra (LLNL) machines, offering 148.6 and 94.6 petaflops respectively, Read more…

By Tiffany Trader

ScaleMatrix and Nvidia Launch ‘Deploy Anywhere’ DGX HPC and AI in a Controlled Enclosure

November 18, 2019

HPC and AI in a phone booth: ScaleMatrix and Nvidia announced today at the SC19 conference in Denver a joint offering that puts up to 13 petaflops of Nvidia DGX Read more…

By Doug Black

Intel Debuts New GPU – Ponte Vecchio – and Outlines Aspirations for oneAPI

November 17, 2019

Intel today revealed a few more details about its forthcoming Xe line of GPUs – the top SKU is named Ponte Vecchio and will be used in Aurora, the first plann Read more…

By John Russell

SC19: Welcome to Denver

November 17, 2019

A significant swath of the HPC community has come to Denver for SC19, which began today (Sunday) with a rich technical program. As is customary, the ribbon cutt Read more…

By Tiffany Trader

Supercomputer-Powered AI Tackles a Key Fusion Energy Challenge

August 7, 2019

Fusion energy is the Holy Grail of the energy world: low-radioactivity, low-waste, zero-carbon, high-output nuclear power that can run on hydrogen or lithium. T Read more…

By Oliver Peckham

Using AI to Solve One of the Most Prevailing Problems in CFD

October 17, 2019

How can artificial intelligence (AI) and high-performance computing (HPC) solve mesh generation, one of the most commonly referenced problems in computational engineering? A new study has set out to answer this question and create an industry-first AI-mesh application... Read more…

By James Sharpe

Cray Wins NNSA-Livermore ‘El Capitan’ Exascale Contract

August 13, 2019

Cray has won the bid to build the first exascale supercomputer for the National Nuclear Security Administration (NNSA) and Lawrence Livermore National Laborator Read more…

By Tiffany Trader

DARPA Looks to Propel Parallelism

September 4, 2019

As Moore’s law runs out of steam, new programming approaches are being pursued with the goal of greater hardware performance with less coding. The Defense Advanced Projects Research Agency is launching a new programming effort aimed at leveraging the benefits of massive distributed parallelism with less sweat. Read more…

By George Leopold

AMD Launches Epyc Rome, First 7nm CPU

August 8, 2019

From a gala event at the Palace of Fine Arts in San Francisco yesterday (Aug. 7), AMD launched its second-generation Epyc Rome x86 chips, based on its 7nm proce Read more…

By Tiffany Trader

D-Wave’s Path to 5000 Qubits; Google’s Quantum Supremacy Claim

September 24, 2019

On the heels of IBM’s quantum news last week come two more quantum items. D-Wave Systems today announced the name of its forthcoming 5000-qubit system, Advantage (yes the name choice isn’t serendipity), at its user conference being held this week in Newport, RI. Read more…

By John Russell

Ayar Labs to Demo Photonics Chiplet in FPGA Package at Hot Chips

August 19, 2019

Silicon startup Ayar Labs continues to gain momentum with its DARPA-backed optical chiplet technology that puts advanced electronics and optics on the same chip Read more…

By Tiffany Trader

Crystal Ball Gazing: IBM’s Vision for the Future of Computing

October 14, 2019

Dario Gil, IBM’s relatively new director of research, painted a intriguing portrait of the future of computing along with a rough idea of how IBM thinks we’ Read more…

By John Russell

Leading Solution Providers

ISC 2019 Virtual Booth Video Tour

CRAY
CRAY
DDN
DDN
DELL EMC
DELL EMC
GOOGLE
GOOGLE
ONE STOP SYSTEMS
ONE STOP SYSTEMS
PANASAS
PANASAS
VERNE GLOBAL
VERNE GLOBAL

Cray, Fujitsu Both Bringing Fujitsu A64FX-based Supercomputers to Market in 2020

November 12, 2019

The number of top-tier HPC systems makers has shrunk due to a steady march of M&A activity, but there is increased diversity and choice of processing compon Read more…

By Tiffany Trader

Intel Confirms Retreat on Omni-Path

August 1, 2019

Intel Corp.’s plans to make a big splash in the network fabric market for linking HPC and other workloads has apparently belly-flopped. The chipmaker confirmed to us the outlines of an earlier report by the website CRN that it has jettisoned plans for a second-generation version of its Omni-Path interconnect... Read more…

By Staff report

Kubernetes, Containers and HPC

September 19, 2019

Software containers and Kubernetes are important tools for building, deploying, running and managing modern enterprise applications at scale and delivering enterprise software faster and more reliably to the end user — while using resources more efficiently and reducing costs. Read more…

By Daniel Gruber, Burak Yenier and Wolfgang Gentzsch, UberCloud

Dell Ramps Up HPC Testing of AMD Rome Processors

October 21, 2019

Dell Technologies is wading deeper into the AMD-based systems market with a growing evaluation program for the latest Epyc (Rome) microprocessors from AMD. In a Read more…

By John Russell

Rise of NIH’s Biowulf Mirrors the Rise of Computational Biology

July 29, 2019

The story of NIH’s supercomputer Biowulf is fascinating, important, and in many ways representative of the transformation of life sciences and biomedical res Read more…

By John Russell

Xilinx vs. Intel: FPGA Market Leaders Launch Server Accelerator Cards

August 6, 2019

The two FPGA market leaders, Intel and Xilinx, both announced new accelerator cards this week designed to handle specialized, compute-intensive workloads and un Read more…

By Doug Black

Intel Debuts New GPU – Ponte Vecchio – and Outlines Aspirations for oneAPI

November 17, 2019

Intel today revealed a few more details about its forthcoming Xe line of GPUs – the top SKU is named Ponte Vecchio and will be used in Aurora, the first plann Read more…

By John Russell

When Dense Matrix Representations Beat Sparse

September 9, 2019

In our world filled with unintended consequences, it turns out that saving memory space to help deal with GPU limitations, knowing it introduces performance pen Read more…

By James Reinders

  • arrow
  • Click Here for More Headlines
  • arrow
Do NOT follow this link or you will be banned from the site!
Share This