Research Experience for Undergraduates

REU Students from ECSU Program at Digital Science Center

REU 2019

REU 2018

Telemetry data plays an important role in the domain of motor racing, which are generated by hundreds of sensors in the cars during the race. Many existing works focus on the telemetry data analysis by improving both the strategy of the drivers and also the mechanical design of the cars. A challenging issue raised however is anomaly detection for alerting events to be worried about (e.g. car crash) in real-time. Hierarchical Temporal Memory (HTM) is an online machine learning technology that aims to capture the structural and algorithmic properties of the neocortex.

Ravi Teja Bingi:

Working with a team, Bingi conducted experiments with a complete anomaly detection system based on HTM Neural Networks and successfully predicted car crash events using Lap Distance, Vehicle Speed and Engine Speed (RPM) from the Indianapolis 500 telemetry data. The knowledge and skills he obtained in the internship program would broaden his views of machine learning and systems in solving critical real-world problems.

REU 2017

  • Gregor von Laszewski, Syed Asad Zahidi: Indoor GPS System for a robot swarm. Manual
  • Gregor von Laszewski, Jon Montgomery: IoT to build a Robot Swarm. Manual
  • Franklin Bettencourt, ExTASY NSF REU Progress Report, Rutgers University, 2017. Slides.
  • Shane Chu, Grid Search Optimization on Viterbi Algorithm for Ice Bottom Detection, Center for Remote Sensing of Ice Sheets, University of Kansas, 2017. Poster.
  • Rishi Shah, Big Data Analytics of Breast Cancer Using Twitter, Columbia University (host institution: Stony Brook University), 2017. Poster.
  • Kathleen Clark, MDAnalysis PCA Tutorial, University of Arizona (host institution: Arizona State University), 2017.

Summer Interns

Machine Learning is a game changer for industry and academia to harness Big Data problems. Traditionally, data centers install hundreds of commodity CPUs in their clusters, which runs Hadoop-like data analytics services. In recent years, small or middle-sized HPC clusters, with Multi/Manycore and GPU architectures become attractive for cost-effective solutions. All five of the student interns learned these technologies with the Harp DAAL framework and implemented various algorithms on an Intel Knights Landing cluster with 68 or 72 cores per node. Specifically,

Anurag Sharma:
Implemented Singular Value Decomposition algorithms with the Harp-DAAL framework. The program ran on an Intel Knights Landing Cluster with 68 or 72 cores per node.

Mayank Jindal:
Implemented Neural Networks, Naïve Bayer classifier, and other analysis and regression algorithms with the Harp-DAAL framework.

Prawal Gangwar:
Implemented Singular Value Decomposition algorithms with the Harp-DAAL framework.

Ravi Teja Bingi:
Neural Networks and PCA algorithms using Allreduce and Broadcast communications with the Harp-DAAL framework.

Manpreet Gulia:
Implemented Neural Network, SVD and PCA algorithms with the Harp-DAAL framework.

REU 2016

REU 2015

  • Sean M. Holloway. University of Kansas. Wideband Radar Simulator for Evaluating of Direction-of-Arrival Processing. Poster and Report.
  • Richmond Adjei. Indiana University Purdue University Indianapolis and Indiana University. Evaluating Techniques for Image Classification. Poster.
  • Tori Wilborn & Omar Owens. Elizabeth City State University, Winston Salam University, and Indiana University. Exploring Learning Algorithms for Layer Identification from Polar Radar Imagery. Poster.
  • Maya Smith & Anthony Scott. Winston Salem University and Indiana University. Analyzing Stock Data Using Multi-Dimensional Scaling. Poster.
  • Daniel Da Silva & Paulo Chagas. Universidade Tecnologica Federal do Parana, Universidade Federal do Para, and Indiana University. Cloud Computing with Cloudmesh. Poster and Report.
  • Carolyn Kiriakos. Stony Brook University. Semantic Annotation of Venues Using Geo-Tagged Social Media Data. Poster.
  • Ian Kenney. Arizona State University. Biomolecular benchmark systems. Report.
  • Nikhil Shenoy and George Chantzialexiou. Rutgers University. Testing, Integration, Development of SPARK for MIDAS. Report.
  • Md Enayat Ullah. Indian Institute of Technology Kanpur. Mentor—Dr. Geoffrey Fox, Indiana University. Approximation in Large Summation Problems using Sampling. Poster.

These REUs were partially supported by the National Science Foundation Grants ACI-1443054: CIF21 DIBBs: Middleware and High Performance Analytics Libraries for Scalable Data Science and PLR-1263061: REU Site: Arctic and Antarctic Project (AaA-REU) with Research Experience for Teachers (RET) Component.