To arrange for an interview with a researcher, please contact the Communications staff member identified at the end of each tip. For more information on ORNL and its research and development activities, please refer to one of our media contacts. If you have a general media-related question or comment, you can send it to [email protected].

MEDICINE – Deep learning for cancer research… Newswise — The development and maturation of automated data tools for cancer research, part of the objectives outlined in the White House’s Cancer Moonshot initiative, could give medical researchers and policymakers an unprecedented view of the U.S. cancer population—at a level of detail typically only obtained for clinical trial patients (less than 5 percent of the overall cancer population). Using the Titan supercomputer, a team led by Oak Ridge National Laboratory’s Georgia Tourassi is making progress towards this goal by employing deep learning techniques to extract useful information from text-based cancer pathology reports. So far, the team has established deep learning’s advantages in multi-task learning, using nearly 2,000 cancer pathology reports to train a neural network to identify a cancer’s primary site and laterality from text. In another study, Tourassi’s team deployed deep learning to match the cancer’s origin to a corresponding topological code, a more specific classification than primary site or laterality. The promising performance trends measured in these early studies will guide the team as they scale up deep learning to tackle larger datasets and move towards less human supervision. [Contact: Jonathan Hines, (865) 574-6944; [email protected]] Image: https://www.ornl.gov/sites/default/files/news/images/01%20-%20MEDICINE_Deep_Learning_for_Cancer_Research_1.jpg

Caption: A graphic representation of how deep learning is used to extract information from text-based cancer pathology reports. By developing and employing a neural network, a web of weighted calculations designed to produce informed guesses on how to correctly carry out tasks, researchers are working to automate cancer surveillance and accelerate research.

HYDROPOWER – On the map …

Oak Ridge National Laboratory researchers have produced the next generation of the National Hydropower Map – a visualization tool that provides updated statistics on overall capacity and performance on the nation’s hydropower fleet. The map is part of the lab’s National Hydropower Asset Assessment Program, which synthesizes and extends vast national datasets used to evaluate existing and potential hydropower resources in the United States. The ORNL-led program advances integrated energy, water and ecosystem research that informs efforts to boost U.S. hydropower strategic initiatives. In 2015, the Department of Energy reported that hydropower accounted for 6.2 percent of net electricity generation, 48 percent of all renewable power generation and 97 percent of electricity storage in the United States. The newly updated map is available on the NHAAP website. [Contact: Sara Shoemaker, (865) 576-9219, [email protected]]

Image: https://www.ornl.gov/sites/default/files/news/images/02%20-%20HYDROPOWER_On_the_Map.jpg

Caption: Oak Ridge National Laboratory’s hydropower research team creates and maintains the National Hydropower Map that illustrates the importance of hydropower to the nation’s energy system. (Photo credit: TVA/Douglas Dam)

BUILDINGS – Calibrating energy savings …

Researchers at Oak Ridge National Laboratory have further validated the accuracy of the ORNL-developed Autotune building energy model calibration software, an open source system that can inform contractors to make smarter retrofit decisions. Through two case studies, the research team analyzed Autotune’s accuracy, beating the industry’s energy efficiency standards while automating the equivalent of about 45 man-hours of calibration effort per building. Using Autotune, the team analyzed whole-building energy consumption and multiple parameters such as lighting, electricity loads, heating, ventilation, and air conditioning use and temperatures of existing buildings and corrected, or tuned, the models to yield optimal energy efficiency. Results of the case studies were published in Applied Energy. [Contact: Sara Shoemaker, (865) 576-9219, [email protected]]

Image: https://www.ornl.gov/sites/default/files/news/images/03%20-%20BUILDINGS_Calibrating_Energy_Savings.jpg

Caption: ORNL’s Autotune building energy model calibration software, used to inform contractors to make smarter retrofit decisions, created models that beat the industry’s energy efficiency standards.

COMPUTING ̶ Faster calculations …

A multi-institution team led by Jefferson Lab’s Robert Edwards has been using Oak Ridge National Laboratory’s Titan supercomputer to drastically improve modeling subatomic particles using lattice quantum chromodynamics methods. The Jefferson Lab investigators partnered with NVIDIA high-performance computing researcher Kate Clark, who helped create a multigrid algorithm to improve simulations studying the smallest measurable units of matter—quarks, gluons and mesons. The team has seen 7- to 10-fold increases in code performance since implementing the algorithm to take advantage of Titan’s hybrid architecture. The innovation benefits more than just the Edwards team—other research teams are already implementing the algorithm into their own codes. [Eric Gedenk, (865) 241-5497, [email protected]]

Image: https://www.ornl.gov/sites/default/files/news/images/04%20-%20COMPUTING_Faster_Calculations_1.jpg

Caption: By leveraging Titan’s hybrid architecture, a multi-institution research team achieved a 7- to 10-fold increase in code performance for LQCD calculations.

NUCLEAR – Molten salt reactor workshop…

Oak Ridge National Laboratory’s ongoing research in molten salt reactors—along with decades of renowned efforts in the field—is bringing nuclear experts together for its second workshop on the new generation of MSRs. The October 4-5 event is organized around the theme of “Moving MSRs Forward—The Next Steps” and will highlight research and development progress in MSRs, along with technological and commercial updates. The lab’s 2015 workshop took place in conjunction with the commemoration of the 50th anniversary of the startup of the Molten Salt Reactor Experiment at ORNL. This year’s meeting will again include government representatives, U.S. and international researchers, academics, regulators, utilities and reactor design firms. “The response from last year’s attendees was extremely positive and a motivating factor to have ORNL host this second workshop as progress is made on the advancement of MSRs,” ORNL’s David Holcomb said. [Contact: Jason Ellis, (865) 241-5819, [email protected]]

Image: https://www.ornl.gov/sites/default/files/news/images/NUCLEAR_MSR_test_loop_workshop.jpg

Caption: ORNL researchers will discuss a high-temperature fluoride salt pumped test loop project during the 2016 Molten Salt Reactor Workshop.

MATERIALS – Quickly gaining deep insight …

Scientists at Oak Ridge National Laboratory are harnessing big data capture and analytics to quickly develop deep insight into materials and their dynamics. Their new technique builds on Kelvin probe force microscopy, long used to gather information about electronic properties by “feeling” a material’s surface with a probe. The new approach, called General Mode KPFM and described in Scientific Reports, fully captures information about the probe motion to provide surface voltage data on nanometer length and microsecond time scales. “New insights can be gained into the relationship between structure and electrochemical function in complex systems including batteries, solar cells and even biological systems,” said first author Liam Collins. [Contact: Dawn Levy, (865) 576-6448; [email protected]]

Image: https://www.ornl.gov/sites/default/files/news/images/06%20-%20MATERIALS_Quickly_Gaining_Deep_Insight_3_0.jpg

Caption: General Mode KPFM uses advanced signal processing and analysis methods to extract local electronic properties directly from a noisy cantilever deflection signal. Unlike a traditional method that filters and averages the signal, General Mode KPFM preserves the time component of the response. SENSORS – Quantum leap …

By exploiting some exotic quantum states, researchers have conceptually designed a sensor that features unparalleled sensitivity. In a paper published in Physical Review A, Ali Passian of Oak Ridge National Laboratory and George Siopsis of the University of Tennessee describe a photon-sensing scheme for a detector that would behave much like a miniature trampoline. The concept is based on a suspended micro-bridge, but other shapes can be considered. “Photons hitting its surface would cause a tiny displacement that can be detected because of the quantum mechanical effect of squeezing, allowing us to minimize the movement of the sensor prior to arrival of the photons,” Passian said. This approach reduces the inherent limitation caused by all objects having a motion even in their lowest energy state. The next step will be validation through laboratory experiments. [Contact: Sara Shoemaker, (865) 576-9219, [email protected]]

Image: https://www.ornl.gov/sites/default/files/news/images/sensor.jpeg

Caption: A conceptually designed sensor could detect low-intensity light via a squeezed quantum state, that is, a minimum uncertainty state, offering a potential pathway to build a better infrared detector for both research and industrial applications.