Research

Present Research

Research Topics :

Suicide Prevention: A Structured Approach for Analyzing Social Media Data (PI)
- sub-topics: Natural Language Processing, LLM, Clinical Data Analysis, Social Media Data Analysis, Explainable AI, Prompt Engineering, Survey, Interviews, Human Computer Interactions
- Funded by IRI and AccessComputing
- Link to Publications:
  - https://ieeexplore.ieee.org/abstract/document/10460058 [10]
  - https://ceur-ws.org/Vol-3649/Paper5.pdf [9]
Anxiety in Academia: Analyzing Anxiety Patterns and Interventions for Student Success
- sub-topics: Natural Language Processing, Survey, Interviews, Statistical Analysis, Curriculum Development and Adjustment
- Link to Publications:
  - https://library.iated.org/view/SUDHAKARAN2024NAV [13]
  - https://dl.acm.org/doi/abs/10.1145/3585059.3611432 [15]
Integrating Ethical and Technical Proficiency in Education by Bridging the Interdisciplinary and Practical Gaps
- sub-topics: Natural Language Processing, LLM, Web Crawling, Explainable AI, Responsible AI, Survey, Interviews
- Link to Publication:
  - https://library.iated.org/view/NUR2024NAV [12]
Beyond Black Boxes: Unraveling Text Summarization with Explainable AI
- Project: A collaborative project with Birla Institute of Technology and Science, Pilani
- Link to Publication:
  - https://jai.front-sci.com/index.php/jai/article/view/1331 [19]
Assured Knowledge Graph-based Learning for Formal System Specification
- sub-topics: Formal Method, Natural Language Processing, Knowledge graph, Graphical Inference
- Link to Publications:
  - https://ieeexplore.ieee.org/abstract/document/10005530 [16]
  - https://ieeexplore.ieee.org/abstract/document/10316143 [17]
Fake News Analysis on Social Media Platforms
- senior project: Developing a Chrome Extension for Fake News Detection
- sub-topics: Web Crawling, Natural Language Processing, Social Media Data Analysis, Software Engineering
- Link to publication:
  - https://dl.acm.org/doi/abs/10.1145/3529190.3534755 [14]
Discovering Insights from Heterogeneous Student Data: Data-driven Early Interventions for At-risk Students
- sub-topics: AI in Education, CS Education, Applied Machine Learning, AI assisted Human Decision making, Explainable AI, Visualization, Interactive Tool Development, Storytelling

- Link to publications:
  - https://library.iated.org/view/NUR2022EXP [8]
  - continuation of dissertation project

8. Independent Research Projects:

Sentiment Analysis on Bangla Texts: https://aclanthology.org/2023.banglalp-1.38/ [11]
Survey on XAI Techniques: https://www.mdpi.com/2079-9292/12/5/1092 [18]

Research @UNCC

1. Knowledge Graph-Embeddings based Explainable Artificial Intelligence for Enterprise Performance Management (2021, NSF Funded)

- Project website: https://www.gopeaks.org/nsf-i-corps-team-gopeaks

2. The Role of Information Scent in Information Retrieval Behavior (2020, NSF Funded)

3. Focal Point Analysis and Explainable AI for Trusting Machine Learning(2019, DOE Funded)

- Project: A collaborative project with Pacific Northwest National Laboratory
- Link to Publication: https://dl.acm.org/doi/10.1145/3377325.3377514 [4]

4. Anomaly Detection in Transactive Control (2018, DOE Funded)

- Project: A collaborative project with Pacific Northwest National Laboratory
- Link to Publication: https://dl.acm.org/doi/10.1145/3307772.3331028 [2]

5. Graphical Inference for Cybersecurity and Community Detection (2017)

- Publication URLs:
  - 10.1109/BigData47090.2019.9005966 [3]
  - https://arxiv.org/abs/1810.01547 [6]

6. OWASP ASIDE Project (2015-2016)

- Project URL: https://www.owasp.org/index.php/OWASP_ASIDE_Project#Source_Code

Dissertation Research:

Title: DEVELOPING TEMPORAL MACHINE LEARNING APPROACHES TO SUPPORT MODELING, EXPLAINING, AND SENSEMAKING OF ACADEMIC SUCCESS AND RISK OF UNDERGRADUATE STUDENTS

The purpose of this NSF-funded dissertation project is to develop analytical tools and techniques for analyzing the behavior of students at risk at an earlier stage of their education. Besides the three analytical models, I also applied explainable AI to the deep learning model to explain i) the behavior and the training process of the model itself and ii) the association/correlation of different features with predictive variables. Three main components of the project-

Student Data Model and Database Design - Implemented a student data model in MongoDB to accommodate the heterogeneous student data.
Analytical Models: Developed three analytical models - 1) network analytics, 2) unsupervised machine learning, and 3) deep learning-based analytics on the student data model and analyzed students' behavior and performance.
Prototype design for an interactive tool: Designed a prototype of an analytical tool for analyzing students' behavior with temporal analytics and explainable artificial intelligence. This interactive tool involves the student advisers in the knowledge discovery process along with the data scientists. In addition to data scientists, domain experts are involved in the knowledge discovery process to leverage their expertise
Publications:

Google Sites

Report abuse