Student Posters:

Improving Prediction Accuracy of Regression Problems with Optimization-based Ensemble Learning and a Two-layer Feature Selection Method 
Authors: Fatemeh Amini, Mohsen Shahhosseini, Guiping Hu, Hieu Pham, Iowa State University
Keywords: Ensemble, Bias-Variance tradeoff, Genetic Algorithms, Feature Selection, Bi-level Optimization
Abstract (pdf) Poster (pdf)

Improving Plant Disease Recognition With Generative Adversarial Network Under Limited Training Set
Authors: Luning Bi and Guiping Hu, Iowa State University
Keywords: plant diseases recognition, generative adversarial network, regularization
Abstract (pdf)

The Effectiveness of Transfer Learning Using a Pre-trained DINET Model for Medical Image Classification
Authors: James Boit, Dakota State University
Keywords: Deep Neural Network, Convolutional Neural Network, Image Classification
Abstract (pdf) Poster (pdf)

An Agent-based Modeling Approach for Predicting the Behavior of Bighead Carp (Hypophthalmichthys nobilis) Under the Influence of Acoustic Deterrence
Authors: Craig Garzella, Joseph Gaudy, Karl Schmitt, Valparaiso University
Keywords: Bighead Carp, Acoustic Deterrence, Agent-Based Model, Invasive Species, Great Lakes
Abstract (pdf) Poster (pdf)

The Effect of APOE SNPs on Brain Tissue Specific Gene Expression
Authors: Blessing Ibe (University of Illinois at Urbana-Champaign), Anastasia Gurinovich (Boston University) and Paola Sebastiani (Boston University)
Keywords: Biostatistics, Differential expression, Disease
Abstract (pdf)

MiReN: An optimization tool for data-driven discovery of global regulatory phenomena during heat stress in rice seed
Authors: Mohammad Mazharul Islam, Jaspreet Sandhu, Harkamal Walia, and Rajib Saha, University of Nebraska-Lincoln
Keywords: Heat stress, minimal regulatory network, transcriptomics
Abstract (pdf) Poster (pdf)

Representing Low-income Households in Building Energy Modelling Tools
Authors: Diba Malekpour Koupaei (Iowa State University), Farzad Hashemi (Penn State University), Vinciane Tabard-Fortecoƫf (INSA Lyon), Ulrike Passe (Iowa State University)
Keywords: Occupant behavior, American Time-Use Survey, Markov-Chain
Abstract (pdf)

Crowdsourcing Wildlife Data from Social Media
Authors: Michelle Ramirez, Saccha Agriel, Mahmooda Ali, Krishna Vamsi Chndu, Kyla Guru, Josephine Huss, Sourav Jayaprakash, Ellen Kidane, Viktor Kirillov, Shirley Li, Varun Mallampati, Jared Manusig, Hannah Mcdougall, Jason Obrycki, Michelle Ramirez, Zoe Wachtel, Tanya Berger-Wolf (University of Illinois at Chicago), Anastasia Gurinovich (Boston University) and Paola Sebastiani (Boston University)
Keywords: Wildlife, Social Media, Images
Abstract (pdf) Poster (pdf)

Invited Posters:

Integrative Materials Design – Connecting the Materials Data Ecosystem
Authors: Ben Blaiszik (UChicago, Argonne), Ian Foster (UChicago, Argonne), Peter Voorhees (Northwestern University), Laura Bartolo (Northwestern University), Dane Morgan (University of Wisconsin-Madison), Dallas Trinkle (University of Illinois at Urbana-Champaign), John Allison (University of Michigan Ann-Arbor)
Keywords: Materials Science, Data, Machine Learning
Abstract: To realize the full value of available data, and to facilitate the application of machine learning to materials science problems will require enhancing the data ecosystem to enable simple discovery and collection of data from many sources, automated dissemination of new data across the ecosystem, and connecting disparate existing services. Here, we present our efforts to connect materials data services and researchers across the Midwest region and beyond.

Digital Agriculture: Unmanned Aircraft Systems, Plant Sciences and Education (UASPSE)
Authors: Aaron Begstrom, University of North Dakota

CADRE – Collaborative Archive & Data Research Environment
Authors: Valentin Pentchev (1), Robert Van Rennes (3), Xiaoran Yan (1), Patricia L. Mabry (1), Jamie Wittenberg (2), Matthew Hutchinson (1), Benjamin Serrette (1). 1: Indiana University Network Science Institute, 2: Indiana University Library, 3: Big Ten Academic Alliance.
Keywords: Scientometrics, Library Science, Big Data, Science of Science
Abstract: CADRE is an IMLS-funded project that provides sustainable, affordable, and standardized text- and data-mining services for licensed big datasets, as well as open and non-consumptive datasets too large or unwieldy to work with in existing research library environments. CADRE offers academic researchers access to these data in a secure cloud-based platform.

Advanced Computational Neuroscience Network (ACNN) Spoke
Authors: Franco Pestilli, Indiana University
Keywords: Neuroscience; Informatics; Cyberinfrastructure

MBDH Initiative: Data Science in Small Colleges
Authors: Lior Shamir, Kansas State University
Keywords: Small colleges, inclusion, resources
Abstract: While large R1 institutions receive the vast majority of the available federal funding, these institutions train less than 15% of the national IT workforce. The initiative of Big Data and Data Science at non-R1 institutions aims at changing the existing reality to provide equal access to all Data Science students. With the Midwest Big Data Hub, we aim at identifying ways of sharing resources and services between small and large institutions, including the needs of non-R1 institutions in decision processes, and designing a new paradigm of federal and academic partnership.
Poster (pdf)

The Engagement and Performance Operations Center: EPOC
Authors: Douglas Southworth (IU), Edward Moynihan (IU), Jennifer Schopf (IU), Jason Zurawski (ESnet)
Keywords: Engagement, Collaborative Science
Abstract: When researchers have better access to their data, their time to science is reduced, they can ask more complicated questions, and can work in larger teams to solve more complicated problems. The Engagement and Performance Operations Center (EPOC) helps researchers routinely and reliably transfer large datasets faster, thereby improving data access and collaboration. Through our targeted partnerships, EPOC has the potential to benefit nearly all of U.S. science, research and education by working across organizational boundaries to address performance issues that can hinder cooperative research. EPOC works with researchers, engineers, and cyberinfrastructure planners to help make sense of performance slowdowns that occur during the end-to-end data transfer process, both in the moment and when long term planning. We help under-resourced sites take advantage of modern cyberinfrastructure approaches so that all researchers can better collaborate.