2020-01-09: Kritika Garg (Computer Science PhD Student)

I am Kritika Garg, a first-year Ph.D. student at Old Dominion University. My research interests are in the fields of web archiving, social media, and natural language processing. I joined the Old Dominion University in the fall of 2019 under the supervision of Dr. Michael L. Nelson and Dr. Michele C. Weigle. I work with Web Science and Digital Libraries Research Group (WS-DL) where our focus is in the fields of web archiving, digital preservation, social media, and human-computer interaction. My current research work is in the field of web archiving, including analyzing access patterns of robots and humans in web archives and studying whether the patterns prevalent with the Internet Archive are present across different web archives.

I completed my undergrad from Guru Gobind Singh Indraprastha University in June 2019. During my undergrad, I started attending various tech events by tech-groups such as Google Developer GroupPyDelhiWomen Techmakers, Women Who Code, etc. These events acquainted me with various computer science technologies and tools such as python language, machine learning, natural language processing. Attending these events and meeting notable people in science inspired me to pursue a career in research.

Before Joining the WS-DL group, I worked as a research intern at the Indian Institute of Technology (IIT) in Delhi, India. I worked in the fields of natural language processing, social network analysis, machine learning, and information retrieval, under the supervision of Dr. Saroj Kaushik and Mr. Kuntal Dey. We worked on topic life-cycle analysis based on a novel idea where each topic is defined by a cluster of semantically related hashtags. Furthermore, we analyzed the hashtag and topic life-cycle with respect to the communities, about how topics morph over evolutions of hashtags within and across communities. This resulted in "Topic Lifecycle on Social Networks: Analyzing the Effects of Semantic Continuity and Social Communities" (ECIR 2018, with Kuntal Dey, Saroj Kaushik, and Ritvik Shrivastava).  We also conducted a thorough study of topic life-cycle by assessing the influence of users and frequently-used hashtags on the life-cycle, which lead to the publication of "Assessing the role of participants in evolution of topic life cycles on social networks"(Computational Social Networks 2018, also with Dey, Kaushik, and Shrivastava).
Topic Lifecycle on Social Networks: Analyzing the Effects of Semantic Continuity and Social Communities
The four of us also published "Assessing Topical Homophily on Twitter" (Complex Networks 2018) where we investigated the relationship between the familiarity of users and textual similarity of their social media content at the user, peer-group and community granularity. During my internship, I also developed a novel socio-temporal hashtag recommendation system using machine learning and the NLP-based approach. I re-implemented the emtagger model based on word vectors and deeply ingrained socio-temporal techniques within it. The social aspect of the system aims to make use of the hashtags generated by familiar users and the temporal aspect aims to age the tweets.  We also published "A Socio-Temporal Hashtag Recommendation System for Twitter" in Complex Networks 2018. Before coming to ODU, I also worked on developing an information retrieval system based on insights gained from topical life-cycle analysis.

I joined ODU in August 2019 and I recently completed my first semester. I took three courses this semester which helped me to enhance my technical skills. In the Data Visualization course, I learned how to create effective visualizations in R. The Web Server Design course taught me how to run my own RFC-compliant HTTP webserver using python. The Introduction to Emerging Technologies course acquainted me with research work in modern emerging technologies and taught me how to review the work.

During this time period, I worked on analyzing the access logs of the Internet Archive to understand how users access a web archive. This can help gain insights on how to design web archives and how to tailor to their holdings to their respective user bases. This work is an extension of published work by Dr. Yasmin AlNoamany. We presented this work in the Trick or Research event on Halloween in our lab to undergrad students.

-- Kritika Garg (@kritika_garg)