page contents
University of Wisconsin–Madison

Research

Focus on new faculty: Dimitris Papailiopoulos, speeding up machine learning pipelines

papailiopoulos_dimitris-825pixelswide

 

The world of science and technology is currently being reshaped by machine learning. At the heart of machine learning are algorithms whose aim is to mimic the complex processes of human learning, but on a near-unimaginable scale.

Although computers already are learning our preferences, enhancing our online experiences, and allowing for seamless automation to an unimaginable extent, for assistant professor Dimitris Papailiopoulos, who joined the electrical and computer engineering faculty in fall 2016, it is about how we can enable computers to learn concepts even faster.

In several modern applications, it can take up to several weeks for a computer to “learn” a certain concept. For instance, in driverless vehicles, the machine needs time to learn the difference between humans and cars, and intersections and cul de sacs—intuiting the same automatic connections that any human needs to make while driving. However, if researchers like Papailiopoulos can streamline this computational learning process, these novel machines will become much more efficient, accurate, and easier to interact with.

“In contemporary applications, learning algorithms are required to go through millions of examples to learn a concept—this is called the training phase,” he says. “For several modern machine-learning pipelines, this is a major computational bottleneck. The goal of my research is to significantly speed up this training phase.”

The current focus of his research is to reduce the typical duration of training from approximately a week, to only a day or two. Essentially, he’s trying to get computers to learn concepts that are in some cases as complex as those learned by humans, but at an extremely larger scale and rate.

Papailiopoulos, who grew up in Greece, developed an early interest in mathematics and computers, so electrical and computer engineering offered a perfect fit. He received his engineering diploma and master’s degree from the Technical University of Crete in Chania, Greece. He spent seven years there before moving to the United States, where he attended USC as a graduate student. He then followed his advisor to the University of Texas at Austin, where he received his PhD. After his PhD, he spent two years as a postdoctoral researcher at the University of California at Berkeley, where he was a member of the AMPLab.

During his PhD, Papailiopoulos mainly worked on large-scale data processing and data storage problems, which has led to several theoretical and practical innovations in the field of coding theory.

In many ways, his work with coding theory and data storage has been the foundation of his research. He started working on machine-learning problems a little more than three years ago, and is excited about the potential impact of developing and implementing new ideas in the field.

An academic at heart, Papailiopoulos moved to UW-Madison because of its strong research environment.

“I chose UW-Madison because I needed a place that would give me the academic and intellectual freedom to work on the things that I wanted to pursue,” he says. “I was also impressed by how excited people here were about my work and what I have to bring to the table.”

Like many new faculty members, he was also impressed by the highly collaborative and open environment on campus. “I think it’s one of the things that makes UW-Madison special, and isn’t necessarily true for many places; there’s a lot of opportunities for interdisciplinary and cross-disciplinary research here,” he says. “We have, for example, the Wisconsin Institute for Discovery, that brings together faculty from many disciplines like computer science, biology, electrical and computer engineering, statistics, with all sorts of backgrounds—this offers a fantastic space for researchers to collaborate with one another. You’ll find people from the social sciences and engineering working on a joint project, and I think that’s a very beautiful thing that you don’t find in many places.”

Papailiopoulos was hired through the college’s Grainger Institute for Engineering and is affiliated with the Wisconsin Institute for Discovery (WID) on campus. He’s part of the optimization group at the WID, which allows him to interact with people who use the same mathematical tools that he does, but often for entirely different applications.

“The optimization group offers a lot of intellectual diversity. You are exposed to concepts that may have been applied to, say, industrial and systems engineering, and you can get ideas about how you can apply those same concepts to some of the computational work in the field of machine learning,” Papailiopoulos says. “There are also all sorts of events organized by the group, which is another way to get your ideas out there, and get feedback from people of various scientific backgrounds.”

He is currently teaching a special topics class, ECE 901: Large-scale Machine Learning and Optimization. The class delves into algorithmic challenges associated with machine learning, specifically when applied to large data sets, which is very close to Papailiopoulos’s research interests.

“Machine-learning is a cutting-edge field right now, so there’s a lot of fascinating open problems,” he says. “I enjoy getting to share them with the students, and hearing interesting feedback and ideas on how we can resolve some of the existing algorithmic challenges.”

Madison has become increasingly high-tech, rated No. 5 in the country for high-tech jobs. For Papailiopoulos, Madison is a “hidden gem,” integrating both Midwestern charm with a growing tech scene.

“There’s a lot of tech industry here, and lots of young professionals inside and outside of the university, which creates a very lively and highly intellectual environment,” he says. “There are exciting things happening in Madison.”

Author: Lexy Brodt

Focus on new faculty: Benjamin Peherstorfer, developing mathematical tools to make tractable large-scale numerical simulations in engineering

 

benjamin-peherstorfer-img_2399_crop-825w

 

 

 

How do scientists and engineers solve numerical problems that would require years upon years to solve on today’s fastest supercomputers?

They don’t.

Instead, with the assistance of mathematics, scientists and engineers can rely on approximations that provide answers that are sufficiently accurate for many applications and that can be computed in seconds on laptop computers.

Reducing the complexity of numerical simulations with mathematical tools is what keeps Benjamin Peherstorfer busy at his desk all day. Peherstorfer, hired through the college’s Grainger Institute for Engineering, is the newest faculty member in the Department of Mechanical Engineering. Peherstorfer takes complex computer models from engineers and scientists, develops mathematical tools to reduce them to only the very essential components, and performs the mathematical analysis to ensure the error introduced by this approximation is within an acceptable tolerance.

While approximations may seem not ideal, Peherstorfer says the computer models in engineering and science have become so complex that simply buying a larger computer is not an option. “Using reduced models is really not about making it fast, but about making it possible at all,” Peherstorfer says.

So, what are these incredibly complex problems Peherstorfer is interested in?

One is related to numerical simulations of liquid rocket engines (LREs). Peherstorfer uses his computer modeling methods to account for the uncertainty inherent in these problems: “In virtually any engineering system, uncertainties are introduced because of incomplete data, measurement errors, or tiny variations in the manufacturing process,” Peherstorfer explains. “So how do these small variation affect the system? To estimate the effects, we perform millions of numerical simulations to make statistical statements on how likely it is for the rocket engine to fail, for example. Reduced models are essential to make these computations tractable.”

To complicate Peherstorfer’s research, he works in the context of inverse problems, as opposed to forward problems.

“In forward problems, you have the inputs and you feed them into your numerical simulation and to get your output. In inverse problems, you have the output and you would like to know what inputs gave rise to that output.” Peherstorfer says.

A common real-world example of an inverse problem is x-ray computed tomography, where an object is imagined based on how it scatters incoming x-rays. A computationally very demanding inverse problem is imaging subsurface Earth based on seismic waves. Roughly speaking, inverse problems require more computing power and are more difficult to answer than forward problems, Peherstorfer says.

“When you now think that you additionally want to quantify uncertainties in inverse problems, then you can again see the need for developing cheap and certified reduced models,” he says.

Peherstorfer has long been interested in computational problems in science and engineering and harnessing the power of computers to help find solutions—or at least approximations—to them. He received his bachelor’s, master’s and PhD in computer science at the Technical University of Munich (TUM) in Germany. There, he was a member of the scientific computing group and worked on machine learning to detect patterns in data streams in real-time.

Peherstorfer comes to UW-Madison from a postdoctoral position at Massachusetts Institute of Technology (MIT) in its aerospace computational design laboratory. While at UW-Madison, he plans to work on multi-fidelity modeling to combine multiple computational models for uncertainty quantification in inverse problems. In addition, he’s establishing research collaborations that go beyond the department and college.

“I am absolutely excited about the incredible collaboration opportunities at UW-Madison,” Peherstorfer says. “UW-Madison provides a very collaborative environment, which is a great opportunity for me and my interdisciplinary research that cuts across math, computer science, and engineering.”

Author: Will Cushman

Advanced nano-cutter to boost emerging materials research at UW-Madison

The University of Wisconsin-Madison College of Engineering is the new home of a unique machine that is capable of 3D milling precise to one nanometer. The machine, called the ROBONANO α-0iB, is the first of its kind in North America, and it brings extremely advanced technological capabilities that could represent the future of advanced manufacturing.

The ROBONANO, which is on a multi-year loan from the Japanese robotics manufacturer FANUC, arrived on Sept. 1, 2016, and is housed in the laboratory of Sangkee Min, an assistant professor of mechanical engineering at UW-Madison and a faculty member in the Grainger Institute for Engineering. Officials from FANUC traveled from Japan for a ribbon-cutting ceremony and open house for the ROBONANO, held Sept. 11, 2016. The ROBONANO’s extremely precise capabilities offer Min and colleagues exciting new research opportunities, which he hopes will open up improved and novel approaches to the manufacturing of everything from semiconductors to toys and mobile devices to scientific instruments.

The ROBONANO’s superiority over previous generations of similar machines is obvious: Its ability to cut at the nanoscale is two orders of magnitude more precise than most machines used in advanced manufacturing today.

Image of Sangkee Min and ROBONANOThe ROBONANO is a 5-axis machine that uses non-contact air bearings, which gives it nearly limitless configurations for cutting, scribing and milling materials. Where it’s truly exceptional, however, is in its nano precision. Many materials have different properties at the nanoscale, meaning the ROBONANO can potentially handle emerging and existing materials in new and useful ways.

Min will use the machine’s unique capabilities to explore its suitability for manufacturing emerging materials, as well as currently available materials like synthetic sapphire, which is a promising shatter-proof alternative to glass for screens on devices such as tablets and smartphones. Synthetic sapphire—which is made from heating aluminum oxide to extremely high temperatures—currently is difficult to manufacture at large scales because it is very brittle and difficult to handle. However, Min has already conducted initial research on synthetic sapphire with the ROBONANO machine in Japan and discovered that the material sometimes behaves ductile when handled at the extremely tiny nano level. “Many materials have different properties at the nanoscale that create all sorts of different possibilities that aren’t possible with conventional machines,” he says.

It’s primarily these differences in the physical properties of materials at the nanoscale that Min wants to explore, both in emerging materials and in materials like sapphire that require alternative handling methods to become truly manufacturable.

Min also hopes to explore how the machine can help open up new possibilities for manufacturing design. Most designers are constrained by manufacturing limitations that can choke creativity and slow innovation. Min points to smartphone design as a prime example of this “design for manufacturing” paradigm leading to stale product lines.

“The design of the Apple iPhone has not changed very much since from the first one to the latest iPhone that was just announced,” Min says. “It’s the same for a lot of products. Vehicles are the same. A Ford looks like a Ford.”

That’s because manufacturers have long-term investments in supply chains that are difficult and costly to switch on a dime. The capital risk for changing a manufacturing process is often too high. Min says he hopes his research with the ROBONANO will identify ways to speed up the process and becomes one of the enabling technologies for a new manufacturing paradigm—what Min calls “manufacturing for design.”

“I want to be able to ask the manufacturer, ‘what is your perfect design?’ And be able to provide that,” Min says.

Among successes he’s already had: He recently helped a national laboratory vastly improve the imaging capabilities of a microscopic instrument it manufactures.

The ROBONANO has existed more than for 10 years in Japan, where the semiconductor industry is already using it to improve its products. Min says that the semiconductor industry is one among many industries that can benefit from the ROBONANO’s capabilities. He’s also been approached by the toymaker Lego and other well-known brands to help improve their products. “The opportunities are almost limitless for improving products and manufacturing processes with this machine,” Min says.

 

Focus on new faculty: Sangkee Min, manufacturing the future of innovative design

sangkee-min

 

Sangkee Min doesn’t merely want to push the envelope of possibility for manufacturing; rather, he hopes to redefine the envelope entirely, ushering in a new era of truly innovative designs.

Product development involves an iterative dialogue between designers and manufacturing engineers. Oftentimes, the brains behind creative new concepts find themselves at odds with the operators of the machines that must produce the finished object. Cost, material and manufacturing limitations require designers to repeatedly rework their original visions before the final product leaves the assembly line. This back-and-forth postpones progress and stifles innovation.

During the late 1980s, a new paradigm called “design for manufacturability” emerged. Engineers sought to educate designers about manufacturing processes. Designers began conceptualizing products with an eye toward maximal yield for minimal cost. Under this framework, the average time for an innovation to come to market has drastically decreased. However, seeing design as simply another cost-reduction strategy stifles innovation and constrains creativity. Designers reign in their ideas to conform to existing manufacturing constraints, as industrial engineers resist implementing changes that could allow for pioneering products.

“Manufacturing engineers are conservative,” says Min, who in fall 2015 joined UW-Madison as a professor in the Department of Mechanical Engineering and in the Grainger Institute for Engineering. “They don’t want to invest capital in changing machinery to embrace new techniques because they are constantly under pressure to reduce costs.” He says that the industrial environment offers engineers scant opportunities to explore or adapt new technologies in their daily practice.

The resistance to change on the production side, combined with designers restraining their ideas to conform to existing limitations, leads to a spiral of stagnation.

Min wants to break this cycle. “Everyone is stuck with outdated concepts about what is possible. The market isn’t competing with new ideas,” he says.

Rather than framing design as one more opportunity to cut costs, Min aspires to flip the script. “If you can deliver something good that the consumer will buy, then you are the value-added process,“ he says. He envisions a future where designers, unencumbered by manufacturing constraints, have ultimate freedom to produce their most avant-garde ideas.

With that mantra in mind, Min comes to the Grainger Institute to advance a new theory of concept-creation called “manufacturing for design.”

Manufacturing for design dictates that rather than revising a concept due to manufacturing hurdles, engineers should work together to overcome traditional challenges standing in the way. Min aspires to bring together experts across a panoply of manufacturing processes to attack any problems from multiple angles. Drawing from his experience in industry and the expertise of UW-Madison faculty, he hopes to create a network of innovation from a huge knowledge base, including specialists in photolithography, metal cutting, injection molding, etching, additive manufacturing and advanced manufacturing in order to produce truly innovative designs.

The applications of manufacturing for design range from high-performance consumer products to scientific instruments. Min recently applied this idea to producing components for electron microscopy at Lawrence Berkeley National Laboratories. The scientists working within the group said that they typically met a wall of rejection from engineers when they asked for high-performance precision parts within their instruments. Rather than telling the physicists that their vision for improvements to the reactor was impossible, Min applied his own expertise in nanoscale cutting to find a way to fabricate the requested parts. His open-mindedness not only advanced the field of manufacturing, but fostered discovery by providing a new tool to researchers.

Min comes to UW-Madison with extensive experience ranging from his affiliation with Keio University, one of Japan’s centers for engineering excellence, to his time as executive director of the Manufacturing Institute for Research on Advanced Initiatives (MIRAI) in Berkeley, California. He hopes to contribute to society by helping with innovative designs, and sees the Grainger Institute—and the university overall—as the perfect arena to further his vision.

“In an industry setting, manufacturing for design is a huge risk because I cannot guarantee a result. It is research-focused manufacturing at this moment,” says Min.

While working within the university, he plans to interact with designers and engineers to identify current challenges, which he will translate into research projects. “Wisconsin is the only place I’d be able to do this kind of work,” he says. “It requires a huge amount of collaborative effort, not just one professor driving research.”

With the support of the Grainger Institute and the entire engineering college as resources, Min hopes to explore unprecedented frontiers in manufacturing. He doesn’t merely hope to push the envelope; he wants to refold the envelope into an innovative, unorthodox, entirely new shape.

Author: Sam Million-Weaver

Focus on new faculty: Xin Wang, untying knotty systems

Xin Wang1 (1).jpg

Where others see hopelessly foundering logistics, Xin Wang sees the possibility of creating sustainable and resilient systems for things like developing biofuels and keeping cities operating when disaster strikes.

Wang, assistant professor of industrial and systems engineering and an affiliate of the new Grainger Institute for Engineering, brings an interdisciplinary approach to solving these large-scale problems.

“This is a very exciting opportunity,” says Wang. “My focus and background is interdisciplinary. I plan to collaborate with colleagues in the Grainger Institute, those in industrial engineering and other departments.”

When Wang was a doctoral student at the University of Illinois at Urbana-Champaign, his research—mainly funded by the National Science Foundation—honed in on the creation of biofuels, which is wrought with competing factors and uncertainty.

For example, Wang developed mathematical models to examine ways in which the government can better design policies to answer challenging questions in the development of corn-based and other cellulosic biofuels, such as how to maintain food security and develop new energy sources in ways that are environmentally sustainable.

“Our framework can be applied successfully to solve issues when there are competition, reliability and interdependence issues,” he says.

To validate his research, Wang employed a multi-user, web-based simulation game in which players assumed the roles of various stakeholders who make individual decisions on use of farmland, biofuel investments and government mandates and subsidies.

“This software can help us collect data that simulates reality, and help show how various decisions by stakeholders affect the entire system,” Wang says. “We have to consider not only the economic impact, but the social and environmental impacts. That will help develop the biofuel industry in a sustainable way.”

Wang says the model can be applied to a variety of complex systems, such as infrastructure or manufacturing systems. Recently, Wang began working with the U.S. Army Corps of Engineers to use the mathematical model to analyze the reliability of urban systems.

Specifically, Wang researches how the government plans to protect critical urban infrastructure. He hopes the research will help the government evaluate adverse impacts and enhance preparedness and reliability of key urban systems.

During a natural or man-made disaster, Wang says events may cascade—causing a potentially disastrous effect. Some of the effects hit the physical infrastructure and some affect the supply of resources.

For example, a power outage could cause electricity-dependent water systems to shut down. In turn, people might travel to seek water, pinching fuel supplies and resulting in gridlocked traffic.

“When you consider the problem at first, it may not seem complicated, but when you factor in people’s behavior, it can become significant and disaster could happen,” Wang says. “If the disruption caused by people’s behavior is at a critical infrastructure, it may amplify the disruption.”

The research will help the government know the social impact of infrastructure breakdowns and respond accordingly.

“We hope to tell, based on the infrastructure disruption, what is the social impact? Then the government can have an idea about the reliability of a city and which infrastructure is the most critical to protect,” Wang says.

Inherent in Wang’s research is a depth of knowledge in logistics systems and supply chain management, key components in advanced manufacturing, one of the focuses of the Grainger Institute. The institute, created in 2014 with a $25 million gift from The Grainger Foundation, serves as an incubator for transdisciplinary research.

“Advanced manufacturing is not only using the innovative technology to improve production,” says Wang. “But it also needs to use successful management methodologies to enhance supply chain efficiency and reduce supply uncertainty or energy usage and environmental impact.”