2018 Prospective Student Visit Day and Graduate Research Symposium

The Computer Science (CS) department welcomed individuals from around the country interested in graduate programs during our annual Prospective Student Visit Day. We are looking for strong students with diverse backgrounds to join our MCS and PhD programs. A wide variety of research areas are represented by our world-class faculty including algorithms; computer security and privacy, cybersecurity; computational epidemiology; computational logic; distributed computing; human-computer interaction; mobile systems; numeric, parallel, and optimization algorithms; retro-computing and historic computer reconstruction; text/web mining; machine learning; informatics; and virtual environments.

Coinciding with the Prospective Student Visit Day, we also hosted the 4th Iowa Computer Science Graduate Research Symposium (2018). Senior CS PhD students presented talks on their latest research, showcasing a variety of CS research areas including algorithms, big data, machine learning, human computer interaction, and data mining. A "keynote" by a CS faculty member followed. Talks were intended for a wide audience with interest in CS, including CS juniors and seniors. The talks presented by current CS graduate students at the Symposium are excellent examples of the exciting research taking place here in Iowa City!


Friday, Nov 9, 2018

Morning Sessions (in MacLean Hall) for Visiting/Prospective Students


Overview of Graduate Programs, Sriram Pemmaraju, Director of Graduate Studies


CS Faculty Meetings and Research Demonstrations


Visiting students will meet and have lunch with Computer Science Graduate Students

Graduate Student Research Symposium (2520D University Capitol Centre)


Session One


Thamer Al Sulaiman

Detecting Complex Mutational Signatures in Human Genomic Large Data


Shreyas Pai

Distributed Lower Bounds and Communication Complexity




Session Two


Mingrui Liu

Fast Online Learning Algorithms for Large-Scale Imbalanced Data


Shehroze Farooqi

Measuring and Mitigating OAuth Access Token Abuse by Collusion Networks

3:25-4:00pm Reception


Keynote Speaker: Padmini Srinivasan

Text Mining: A field of Opportunities and Challenges


Thamer Al Sulaiman

Title: Detecting Complex Mutational Signatures in Human Genomic Large Data

Thamer Al Sulaiman - UI CS PhD candidateAbstract: All cellular forms of life contain Deoxyribonucleic acid (DNA), a molecule that consists of a sugar backbone and four nitrogenous bases. The order of these bases encode all the information necessary to perform both, basic and complex cellular functions. DNA is replicated prior to create cell division to new tissue/organs, and pass genetic information to future generations. The ideal outcome of this process is an exact copy of the original DNA. However, while replication generally occurs without error, it may leave DNA vulnerable to physical DNA damage (e.g. breaks), chemical adducts formed by exposure to DNA damaging agents, and, rarely  changes via mistakes made during the replication process. All of these can lead to permanent changes in the DNA called mutations. Mutations range in scale from single base changes to long stretches of lost, duplicated, or displaced DNA. Yet, mutations of any magnitude range in their consequence, from no effect on the organism, to disease initiation (e.g. cancer), or even death.

In this talk, we limit our focus to mutations in human DNA, and in particular mutations that result from a process known as Microhomology-mediated break-induced replication (MMBIR). Recent literature in human genetics has identified MMBIR as a potential mechanism for producing complex mutations in DNA. MMBIRFinder is a tool to detect MMBIR mutation signatures in Yeast DNA. Although MMBIRFinder has been used successfully on Yeast DNA, MMBIRFinder is not capable of detecting MMBIR mutations in human DNA. While yeast DNA sequence data do not typically occupy more than 10GB, Human DNA sequence data can easily occupy hundreds of GB, at times approaching a TB. Thus, one major reason for the MMBIRFinder’s deficiency with human DNA is the amount of computations required to process such massive quantities of human sequence data. Our contribution in this regard is two- fold:

1) we utilize parallel computations to significantly reduce the processing time consumed by  the original MMBIRFinder, and address several performance degrading issues inherent in the original design;

2) we introduce a new heuristic to detect MMBIR mutations that were not detected by the original MMBIRFinder, even in the case of Yeast’s DNA.

6th year PhD student | Advisor: Suely Oliveira | Area of research: Clustering and Classification Algorithms for Large Data, Parallel Algorithms for Machine Learning and Scientific Computing, Numerical Analysis

Shreyas Pai

Title: Distributed Lower Bounds and Communication Complexity

Shreyas Pai - UI CS PhD CandidateAbstract: Over the years, computer scientists have developed efficient sequential algorithms for many problems while simultaneously being aware of what cannot be computed efficiently. In the current age of "Big Data", our datasets are so large that even the most efficient sequential algorithms take a prohibitively long time to execute. This problem is exacerbated if the dataset resides across different machines and is so large that it can not fit entirely in a single  machine. One way to address this problem is to process the data in a distributed fashion. In a distributed system we have a network of machines and the input dataset is partitioned across the machines, who communicate with each other in order to process the input. The goal is to design a distributed algorithm that allows the machines to quickly process the entire input without communicating a lot of information.

As distributed algorithms become more and more useful, understanding their limitations becomes equally important. One of the attractions of studying the limitations of distributed algorithms is that we can prove unconditional lower bounds on the amount of time it takes to solve a problem in a distributed manner. In this talk we will explore Communication Complexity which is one of the most useful tools for proving such unconditional distributed lower bounds.

3rd year PhD student | Advisor: Sriram Pemmaraju | Areas of research: Distributed Algorithms, Communication Complexity, Combinatorial Optimization, and Algorithmic Game Theory

Mingrui Liu

Title: Fast Online Learning Algorithms for Large-Scale Imbalanced Data

Mingrui Liu - UI CS PhD CandidateAbstract: Online learning receives tremendous attention since it can handle streaming data. In many applications (e.g., medical diagnostics, spam email detection, malicious URL detection), we are facing with imbalanced data where the number of positive samples is much larger than the number of negative samples. Classical optimization algorithms designed for minimizing the misclassification rate are not suitable for handling large-scale imbalanced data. In this talk, I will present a stochastic optimization algorithm for optimizing AUC (Area under the ROC Curve). Our proposed algorithm improves over the state-of-the-art algorithms in terms of computational complexity, and also shows better performance in real applications.

3rd year PhD student | Advisor: Tianbao Yang | Areas of research: Machine learning, Optimization, and Learning Theory

Shehroze Farooqi

Title: Measuring and Mitigating OAuth Access Token Abuse by Collusion Networks

Shehroze Farooqi - UI CS PhD CandidateAbstract: We uncover a thriving ecosystem of large-scale reputation manipulation services on Facebook that leverage the principle of collusion. Collusion networks collect OAuth access tokens from colluding members and abuse them to provide fake likes or comments to their members. We carry out a comprehensive measurement study to understand how these collusion networks exploit popular third-party Facebook applications with weak security settings to retrieve OAuth access tokens. We infiltrate popular collusion networks using honeypots and identify more than one million colluding Facebook accounts by “milking” these collusion networks. We disclose our findings to Facebook and collaborate with them to implement a series of countermeasures that mitigate OAuth access token abuse without sacrificing application platform usability for third-party developers. These countermeasures remained in place until April 2017, after which Facebook implemented a set of unrelated changes in its infrastructure to counter collusion networks. We are the first to report and effectively mitigate large-scale OAuth access token abuse in the wild.

4th year PhD student | Advisor: Zubair Shafiq | Areas of research: Security and Privacy, Abuse and Fraud, and Online Social Networks

Padmini Srinivasan - Professor of Computer Science at the University of IowaKeynote Speaker Padmini Srinivasan

Title: Text Mining: A Field of Opportunities and Challenges 

Abstract: We live in exciting times with a variety of options for staying connected with each other, for expressing our thoughts and creativity, for recording observations about ourselves and about others and for documenting our advances. As a consequence, there is an abundance of texts of different types from social media posts, technical writings, captions on graphics, emails, notes written by health care providers etc.  When collected, these large and varied text ‘collections’ form rich repositories of human expression that can be processed using automated ‘text mining’ algorithms for either exploratory analysis or for hypothesis driven study. For example, one might analyze text messages between individuals to explore sender - recipient power balance. One might analyze technical writing or poetry from different eras to see if distinct time periods have distinct stylistic signatures. Or one might mine medical records to predict diagnosis. This talk will illustrate these opportunities with projects that we have worked on in recent years. The aim is to convey the excitement and challenges in the field of text mining.

Bio: Srinivasan is a full professor in the Department of Computer Science at the University of Iowa. Her expertise is in text analytics and she has worked on problems such as text retrieval, question answering and summarization and text mining. She has received funding from the NIH, NSF, Rockwell, NIJ (National Institute of Justice). She has been a member and chair of the only standing review board of the NIH/NLM (Biomedical Library and Informatics Review Committee: 2009-2013) and is presently on the Editorial board of BMC Bioinformatics. She has close to 130 refereed publications and recently received a Test of Time Award  from ACM SIGIR for a paper written in 1983. Professor Srinivasan enjoys teaching and mentoring and is happy to have guided several well-placed PhD students.

View at events.uiowa.edu.


Getting to the ConferenceMap of area and symposium location with parking directions (West and South West)

Iowa City lies just off of Interstate 80 in Eastern Iowa. Regardless of whether you are coming from the West or the East, you will want to take exit 244 off of I-80 following Dubuque Street until you intersect with Church Street. Follow Church Street as it turns into Clinton Street and continue straight until you reach the downtown area and the Pentacrest (you will see the Old Capitol to the right/West), which is the heart of the University of Iowa's campus. All of the symposium will take place in that general vicinity in MacLean Hall and 2520D University Capitol Centre (UCC). Check out the map below!