CMU Group Describes “Superhuman” Poker AI in Science
In a paper published online in December 2017 by the journal Science, Tuomas Sandholm of the CMU School of Computer Science and Noam Brown, a PhD student working with him, detailed how their artificial intelligence program (AI) achieved superhuman performance at Heads-Up, No-Limit Texas Hold’em poker. “Libratus” beat four of the world’s best human players by breaking the game into computationally manageable parts and, based on its opponents’ game play, fixing potential weaknesses in its strategy during the competition. Libratus used PSC’s Bridges system to play and to formulate its strategy.
Earlier that month, another paper by Sandholm and Brown about Libratus took one of three best paper awards at the Neural Information Processing Systems (NIPS 2017) conference in Long Beach, Calif.
Read about the Brains vs. AI Tournament
CMU, PSC and Pitt to Build Brain Data Repository
Researchers with Carnegie Mellon’s Molecular Biosensor and Imaging Center (MBIC), PSC and the University of Pittsburgh’s Center for Biological Imaging (CBI) will help to usher in an era of open data research in neuroscience by building a confocal fluorescence microscopy data repository. The data archive will give researchers easy, searchable access to petabytes of existing data.
The project is funded by a $5 million, five-year grant from the National Institutes of Health’s (NIH’s) National Institute of Mental Health (MH114793) and is part of the federal BRAIN initiative.
The Pittsburgh-based team will bring together MBIC’s and CBI’s expertise in cell imaging and microscopy and pair it with PSC’s expertise in biomedical high performance computing to create a system called the Brain Imaging Archive. Researchers will be able to submit their whole brain images, along with metadata about the images, to the archive. There the data will be indexed into a searchable system that can be accessed using the Internet. Researchers can search the system to find existing data that will help them narrow down their research targets, making research much more efficient.
PSC-Led Flu Vaccine Research Wins International Award
Research on the best strategies for offering flu vaccinations to the public at PSC, the University of Pittsburgh and Soongsil University in the Republic of Korea won a 2017 Innovation Excellence Award from the Hyperion Research User Forum Steering Committee. Hyperion Research, the world’s most respected high performance computing (HPC) industry analyst group for more than 25 years, presented the award at the SC17 HPC conference in Denver, Colo., in November.
Jay DePasse, director of Public Health Applications at PSC and first author of the study, and his collaborators first used PSC’s dedicated public health computing system Olympus to show that offering different types of vaccination—the familiar injected vaccine or two types of “needle sparing” vaccines—would reduce flu cases and make vaccination more cost effective in the Washington, D.C., area. PSC’s larger Bridges system then allowed the team to build a massive, “agent-based” simulation that showed these results also applied to multiple regions of the country with very different populations, including Allegheny County, Pa.; Wayne County, Mi.; Santa Clara County, Calif.; and Salt Lake County, Utah. Bridges’ power also allowed the scientists to test a wider range of assumptions about increased coverage and virus spread, showing that even moderate increases in coverage due to offering more choices can reduce costs and decrease influenza cases by 5,600 to 35,000 people across all five counties. The new work appeared in the journal Vaccine in July, 2017.
PSC Wins a Record Five HPCwire Readers’, Editors’ Choice Awards
Also at the SC17 conference in November, PSC was recognized with a best-year-ever five HPCwire Readers’ and Editors’ Choice Awards. PSC received:
- Readers’ Choice Awards:
- Outstanding Leadership in HPC – Nick Nystrom, Interim Director, PSC.
- Best Use of HPC in Energy: PSC with Texas A&M used OpenFOAM on PSC Bridges & Texas Advanced Computing Center’s Stampede to better understand coolant & heat transfer in high-temperature-jet reactors.
- Best Use of AI: CMU School of Computer Science “Libratus” AI on PSC’s “Bridges” wins the Brains vs. AI competition.
- Editors’ Choice Awards:
- Best Use of high performance Data Analytics – PSC’s Bridges for facilitating assembly and analysis of large-scale genomics data.
- Best Use of HPC in Life Sciences – the University of California San Diego, San Diego Supercomputer Center & PSC identified structural changes activating the gene-splicing technology, CRISPR-Cas9.