Fly on the wall – Meet our 2018 poster presenters and judges
Our annual poster contest takes place the day before our Community Day Summit every year. We keep all the posters under wraps until after the judging is complete. I love the fact that people are so eager to see the posters that they want to get in to the room early, but there are no sneak previews, that is apart from the professors who come along to give support to their students. But even they have to leave when the judges are ready to start!
I mix up the order in which the posters are displayed so that the judges’ experience is as varied as possible and make a few last minute changes to suit my slight OCD tendencies (don’t put the blue ones together)!
The students come to the room early to get settled and then the judges arrive. We are ready to start. There are 10 posters on display, 8 are presented by students who joined the HPCC Systems intern program this year, and 2 are presented by students who are otherwise involved in our academic program.
Let’s walk around the room as it was laid out on the day. To see more detailed information about each poster, view the abstracts and see a larger version of each poster on our Technical Poster Contest wiki.
Farah Al Shanik PhD Student – Clemson University
Equivalence terms for the text search bundle
Farah’s poster illustrates the work she completed as part of her HPCC Systems internship. Her machine learning related project provides the ability to automatically create equivalents for initialisms and acronyms. Her project also provides a means of applying a table of equivalents and the attributes to build that table from an open source thesaurus such as Moby. In completing this project, she has provided an answer to the question: How similar does a term in a search request need to be to a term in the document to be considered a term match?
Listen to Farah present about this project during our September 2018 Tech Talk.
Notes from the judges’ scorecards:
Farah’s poster was well structured and easy to follow. Judges commented on the clarity with which she presented about her work and her ability to talk around the subject and answer their questions knowledgeably.
Zhe Yu – PhD Student North Carolina State University
How to be rich – A study of Monsters and Mice of American Industry
Zhe’s poster showed how he had used HPCC Systems to find and analyse relationships between public companies, to establish whether there was link between growth and the number of relationships achieved. His results showed that the more isolated companies had higher revenues. So the mice, have it!
Zhe is involved in other projects as part of our collaboration with Tim Menzies and his team at NCSU. Last year, he presented a poster at our contest about the Fast Retrieval of relevant information through HPCC Systems.
Notes from the judges’ scorecards
This poster was very well received and judges commented on the interesting title, images and design used to illustrate Zhe’s fascinating results. His explanation of the ETL process he followed using HPCC Systems was also highlighted as a point of interest.
Nicole Navarro – Masters Student, New College of Florida
Measuring the geo-social distribution of opioid prescriptions
Nicole completed an internship with us this year, working on a business facing team. She was tasked with creating visualizations to demonstrate findings from research into drug socialization across social groups and geographical areas. The idea was to broaden the understanding of the issues facing communities when targeting interventions and treatment/education programs.
Listen to Nicole present about this project during our November 2018 Tech Talk.
Notes from the judges’ scorecards
Judges commented that this poster was extremely well organised and illustrated the power of her visualizations very effectively. They also commented on the clear and professional presentation of her work. It was also noted that this is a very topical issue with judges having seen articles written about the opioid epidemic very recently.
Robert Kennedy – PhD Student, Florida Atlantic University
Distributed Deep Learning with TensorFlow
Robert’s intern project involved combining HPCC Systems with TensorFlow to provide the basis for future neural network research and enhance the capabilities of the HPCC Systems machine learning library. He also spoke about this work on the main stage the following day (Watch Recording / View Slides).
Listen to Robert present about this project during our August 2018 Tech Talk.
Notes from the judges’ scorecards
Judges commented on the innovative nature of this project and said how clearly Robert’s poster explained what they saw as a very clever use of HPCC Systems using pyembed.
Aramis Tanelus – High School Student, American Heritage School
HPCC Systems Robotics Data Ingestion
Aramis is part of a team of high school students working on an autonomous agricultural robot project led by Taiowa Donovan, who presented about this work the following day (Watch Recording / View Slides). Aramis’s intern project achievements support the school’s robotics project by allowing them to ingest and process data collected from various sensors located around the robot. In fact, any robotics project using the most common robotics sensors will be able to use HPCC Systems for this purpose as a result of Aramis’s work.
Watch a recording of the robot in action and listen to Aramis present about this project during our August 2018 Tech Talk.
Notes from the judges’ scorecards
The judges commented on the innovative ideas that went into this project. They were particularly interested in the ability to potentially collect data from a variety of devices, use the HPCC Systems ETL process and perform analysis on the results. They also noted the future potential for extensions to the APIs.
Shah Muhammad Hamdi – PhD Student, Georgia State University
Dimensionality Reduction using PBblas
Feature selection is used in data mining to filter out irrelevant and redundant features from those which are the most predictive. Hamdi’s implementation is likely to be a popular addition to our machine learning library since feature selection promotes a better understanding of the data, reduces the curse of dimensionality, speeds up the learning process, and generally improves the model’s performance. Hamdi’s poster tracks how he completed this work during his internship with HPCC Systems.
Listen to Hamdi present about this project during our August 2018 Tech Talk.
Notes from the judges’ scorecards
Judges commented how this was a much needed feature for the HPCC Systems machine learning library. His excellent comparison of MATLAB and PBBlas in ECL was noted as were his well-crafted ECL algorithms.
Saminda Wijeratne – Georgia Institute of Technology
MPI – Proof of Concept
Saminda’s intern project evaluates the potential performance benefits of switching from our own high-performance cluster computer message passing software, to the industry standard MPI software. Having compared the two solutions, the team decided that there is no compelling reason to make the switch. This project successfully allowed MPI to be used as a pluggable transport layer inside the HPCC Systems platform in some areas. Saminda proved that the transport layer could be replaced by another third party implementation in such a way as to be transparent to the rest of the platform. However, the MPI implementation has some characteristics which make it difficult to cleanly handle exception and stop conditions, so it does not currently fit well into the dynamic HPCC Systems job flow.
Listen to Saminda present about this project during our August 2018 Tech Talk.
Notes from the judges’ scorecards
Judges noted Saminda’s very enthusiastic presentation of what they considered to be project that was a very good fit for the innovation and reinvention theme of the conference. They particularly liked the fact that this project challenged the status quo with clear goals and interesting findings that contributed important knowledge back in to the research and development process.
Itauma Itauma – Kaiser University
Cervical cancer risks factors: Exploratory analysis using HPCC Systems
Itauma is a familiar face at our conference and a member of our academic community. He has entered this competition before in 2016 (Unsupervised learning and image classification in a high performance computing cluster) and supported another student’s poster entry in 2017 (Cerise Walker – Is the secret to longevity eating chocolate?) . In this poster, Itauma shows how his initial analysis of data from a cervical cancer database may be useful in identifying areas with limited access to screening and treatment programs. This exploratory project has the potential to show how it may be possible to reduce mortality by successfully identifying and targeting at risk areas. Itauma used the HPCC Systems Visualizer bundle to generate charts to show the results from his results more effectively.
Itauma presented about another interesting research project on the subject of Predicting College STEM Enrollment using HPCC Systems in Educational Research, at a breakout session during our Community Day Summit (Watch Recording / View Slides).
Notes from the judges’ scorecards
It was clear from the comments that the judges felt Itauma’s poster was well organised and easy to follow. This is a project that is very well suited to HPCC Systems and they were very interested to hear how Itauma had used ECL to analyse his results and produce the visualizations. Judges feel this project has the potential to push the boundaries of research in this field and make a real difference.
Everett Matthew Upchurch Butler – Undergraduate, Kennesaw State University
The Future of Automotive Technology
Matt joined us an intern earlier in the year completing an internship project which involved providing a standard ECL math library. He was also a member of the third place winning team at the KSU Hackathon in 2017. The challenge provided at the hackathon was the inspiration for the poster he presented at our poster contest. This poster focuses on assessing the risks and liabilities associated with autonomous vehicles and how the data collected and stored by different companies and agencies may be kept safe from hackers.
Notes from the judges’ scorecards
The relevance of this poster to today’s world was well recognised. Judges felt this was an innovative piece of work that had wide business appeal and also that it was a great use of the analytical capabilities of HPCC Systems. Judges agreed that the poster was well presented and said that Matt did a great job of explaining his work very clearly.
Lili Xu – PhD Student, Clemson University
Using HPCC Systems machine learning to map public records data descriptions to standardised categories
Lili has interned with us for three years running and each year has presented a poster about her achievements (see her entries for 2017 and 2016). Having almost completed her studies, Lili is now a LexisNexis employee. This poster illustrates the work she completed during her final internship with us this summer. It tells the story of how to address the problem of mapping thousands of disparate public records data descriptions to a corresponding set of standard codes, by exploring driving records data using HPCC Systems machine learning. Lili presented at one of our breakout sessions about her work, alongside our LexisNexis colleague, Gus Reyna (Watch Recording / View Slides).
Listen to Lili present about this project during our September 2018 Tech Talk.
Notes from the judges’ scorecards
Comments showed that judges appreciated the relevance of the data analytics problems presented on Lili’s poster and were impressed by her knowledge and the potential of her findings. They also noted how she told the story very well and that her poster was exceptionally well organised and easy to follow.
Meet the judges
Now you have read about all the posters, let me introduce you to our judges who have the very difficult job of choosing the winners from an impressive collection of posters. There are four judges split in to two teams.
Judging Team 1
Judging Team 1 – LexisNexis colleagues Kunal Aswani and Prabhu Sadasivam
Kunal Aswani – Software Engineer III on the HPCC Systems platform team. Kunal has been heavily involved in the testing of the HPCC Systems platform and works on improvements and features in the ECL Watch user interface.
Prabhu Sadasivam – Director of Software Engineering and leader of the Analytic Technology and Center of excellence for Machine Learning in Applied Research and Development. Prabhu is responsible for the analytics architecture and partnering with modeling leaders and data scientists in implementing revenue-driven analytics solutions including attributes and predictive models.
Judging Team 2
Judging Team 2 – Guest judge Fabian Fier, Humboldt University with Joe Chambers, LexisNexis
Fabian Fier – Fabian is a researcher at Humboldt University in Berlin and has worked on projects supported by HPCC Systems as part of our academic program for a number of years. Fabian also presented in one of our breakout sessions during our 2018 Community Day Summit, speaking about Optimizing Set-Similarity Join and Search with Different Prefix Schemes (Watch Recording / View Slides).
Joe Chambers – Software Engineer, Lead who works on a team that provides data science tools. The focus is on providing products which save development time, decreasing the time to market for new products.
Each judge is given a clipboard which holds a scoring sheet for each student. While they work the room in pairs, we ask them to judge individually using the following criteria:
- Content, including originality and relevance to HPCC Systems and the theme of our summit which was innovation and reinvention.
- Poster design, including organization and visual appeal
- Presentation skills, including communication skills, clarity and flow.
The judges seem surprised when I advise them to spend no more than 7 minutes with each student. But I know from previous experience it takes around 2 hours to complete the judging process. Some interviews inevitably over-run as judges become absorbed in the content and have questions or comments. I also know that students and judges will definitely need to take a break every now and then.
As we get towards the end of the judging period, I am aware that it’s getting rather noisy outside the room. People are assembling to join our academic program meet and greet event which will happen immediately after judging is complete. I’m quite excited about this because there are a number of introductions I want to make. Our student interns often work remotely from their mentors and this will be the first chance they have had to meet each other in person. The same is true of some of the professors involved in our academic program, who generally meet with us via conference call. I can’t think of a more appropriate way to complete the day.
I collect the scorecards, give the presenters and judges a couple of minutes to relax and then open the doors. I hand the scorecards to Trish McCall who squirrels them away for safe keeping, so we can tally the scores in a secret location later!
This meet and greet event is a new addition to the agenda this year and it is having the desired effect. The room is alive with chatter as people talk to the students about their work and I see connections being made all around the room:
Dr Amy Apon, Clemson University with poster presenter Farah Alshanik, meeting her LexisNexis mentors, Kevin Wilmoth and David Miller in person for the first time
Poster presenter Everett Matthew Upchurch Butler supported by Dawn Tatum, IT faculty and Internship Coordinator of CCSE and Jon Preston, Dean of CCSE at Kennesaw State University
Jingqing Zhang, Imperial College London meeting poster presenter Robert Kennedy and his professor, Dr Taghi Khoshgoftaar, Florida Atlantic University
Our youngest 2018 poster presenter Aramis Tanelus, with his teacher Taiowa Donovan, American Heritage School and LexisNexis Mentors David Dehilster (left) and Kevin Wang (front)
Academic Meet and Greet Event featuring Bob Foreman LexisNexis, Tim Menzies North Carolina State University, Dan Camper Lexis Nexis, Fabian Fier Humboldt University, Vijay Ragahavan EVP and CTO, LexisNexis and many more
Vijay Raghavan, EVP and CTO LexisNexis with Saminda Wijeratne, his LexisNexis mentor Jake Smith and Richard Chapman VP Technology, LexisNexis and leader of the HPCC Systems platform team
During the meet and greet, I ask a number of people what they think of the posters displayed this year and which is their personal favourite. At this point, I still have no idea myself who the judges have chosen as the winners, so I am thinking maybe I can pick up a hint. The next day during the awards ceremony, Flavio Villanustre asks the room whether anyone predicted the winners. I’m not sure I could have accurately predicted the outcome even after conducting my mini exit poll!
Answers I heard ranged from comments about the interesting content, people wanting to know more about how to get their hands on the resulting feature or code, the impressive way the student told their accompanying story, how a poster showed an interesting use case for HPCC Systems they hadn’t considered before to the use of images and the appealing presentation. Everyone agreed that the contest was a showcase of impressive achievements.
After the meet and greet, the posters remained on display attracting attention throughout our 2018 Community Day Summit, each student taking the opportunity to talk about and answer a few questions about their poster at the start of the breakout sessions that afternoon.
Congratulations to our winners!
Flavio announces the results, (Watch Recording / View Slides):
Our three winners!
Congratulations to…
Saminda Wijeratne – First Place Winner
Georgia Institute of Technology
MPI – Proof of Concept
Nicole Navarro – Second Place Winner
Masters Student, New College of Florida
Measuring the geo-social distribution of opioid prescriptions
Robert Kennedy Third Place Winner
PhD Student, Florida Atlantic University
Distributed Deep Learning with TensorFlow
Thanks to…
This amazing group of student presenters who worked so hard on their HPCC Systems related project and did such a fantastic job of presenting their results.
Professors and mentors who support the students involved in our academic program and celebrate their achievements.
Our judges, Kunal Aswani, Prabhu Sadasivam, Joe Chambers and Fabian Fier for their time, knowledge and experience.
The whole poster contest team: From the left, Judges Fabian Fier and Joe Chambers, Back row Roert Kennedy, Aramis Tanelus, Everett Matthew Upchurch Butler, Shah Muhammad Hamdi, Itauma Itauma, Saminda Wijeratne and judge, Kunal Aswani. Front row from the right, judge Prabhu Sadasivam, Lili Xu, organiser Lorraine Chapman, Nicole Navarro, Farah Alshanik and Zhe Yu
Join us next year…
Interested in taking part next year? Get in touch and tell us about your HPCC Systems related project. US-based students from high school through to PhD are eligible to enter posters into this annual poster contest.
Interested in interning with HPCC Systems? Find out more about the HPCC Systems intern program and available projects for 2019. The application period is now open and the deadline is Friday 29th March 2019. Contact us for more information.