Research

My research publications are listed on Google Scholar.
My doctoral research focused on the application of natural language processing and supervised classification techniques to enhance text accessibility in Hindi. Key areas of my research include:

  • Natural Language Processing: Identifying complex words in Hindi text and providing simpler alternatives to enhance readability.
  • Data Science and Machine Learning: Using supervised classification techniques for word sense disambiguation.
  • Educational Technology: Published studies in international journals and conference proceedings.
I earned my Ph.D. from Symbiosis International University, India, with a thesis on Automated Lexical Simplification of Hindi Text. The objective of the research was to improve the accessibility of text in Hindi by automating the identification of complex, that is, difficult to understand words, and providing simpler alternatives to the target complex words. A significant of this research was sponsored by Symbiosis Centre for Research and Innovation, Symbiosis International University, and approved by the Independent Ethics Committee, Symbiosis International University.
The project involved creating language resources, developing models for identifying complex words, and devising and implementing word sense disambiguation techniques for selecting appropriate simpler synonyms.

Key Contributions

  • Designed and executed a comprehensive annotation study, and engineered novel language resources for Hindi, including a stop lemma list, aesthetics corpus, and the CWID-hi dataset, demonstrating expertise in data collection, cleaning, engineering, and corpus linguistics.
  • Devised and tuned classification models for complex word identification and lexical simplification in Hindi.
  • Conducted in-depth data analysis to provide key insights into factors influencing word complexity in Hindi.
  • Used feature engineering methods to identify key predictors of word complexity, with frequency emerging as the most impactful.
  • Leveraged ensemble machine learning models to achieve promising results in Hindi complex word identification.
  • Experimented with and successfully implemented knowledge-based word sense disambiguation for Hindi using BabelNet and ARES sense embeddings.
  • Advanced the field of text simplification for Hindi, with potential applications in education, healthcare, and accessibility.
The resources created independently as well as in collaboration with other researchers are as follows:
The code and data for the studies can be found at: This research represents a significant milestone in language technology and linguistic analysis for Hindi, addressing the need for improved text accessibility and readability. The developed resources and methods have potential real-world applications in various domains requiring simplified Hindi text.

I have served as a program committee member for the 9th Workshop on Speech and Language Processing for Assistive Technologies, 2022.

Besides this, I have conducted research in the field of accessibility as well as education, including education technology and assessment. I have also mentored student research projects in the areas of machine learning and data science, and published papers in the areas of accessibility, education technology, and machine learning.

Key Research Projects

  • Hindi Aesthetics Corpus and Stop Lemma List Creation

    Sponsor:Symbiosis Centre for Research and Innovation, Symbiosis International University

    Developed a comprehensive corpus focused on aesthetics in Hindi and created a novel stop lemma list, addressing the lack of publicly available consistent linguistic resources for Hindi text processing.

    Keywords: Natural Language Processing Corpus Linguistics Text Preprocessing Hindi Language Resources Stop Words

  • CWID-hi Dataset

    Sponsor:Symbiosis Centre for Research and Innovation, Symbiosis International University

    Created an annotated dataset for complex word identification in Hindi, facilitating machine learning approaches to text simplification and readability enhancement.

    Keywords: Dataset Creation Text Annotation Complex Word Identification Dataset

  • Complex Word Identification Model

    Sponsor:Symbiosis Centre for Research and Innovation, Symbiosis International University

    Developed and implemented classification models for identifying complex words in Hindi text, using ensemble machine learning techniques to achieve state-of-the-art results.

    Keywords: Machine Learning Ensemble Models Feature Engineering Supervised Learning

  • Analysis of Word Complexity Factors in Hindi

    Conducted in-depth research to identify and analyze factors affecting word complexity in Hindi, providing valuable insights for text simplification and readability assessment.

    Keywords: Linguistic Analysis Feature Importance Readability Metrics Text Complexity

  • Lexical Simplification System for Hindi

    Designed and implemented a comprehensive system for lexical simplification in Hindi, incorporating complex word identification and word sense disambiguation techniques.

    Keywords: Lexical Simplification Word Sense Disambiguation ARES Sense Embeddings Language Models Word Embeddings

  • Bibliometric Analysis of Studies on Lexical Simplification

    Analyzed articles from reputable journals and conferences, and highlighted the progress made in this area across various languages. Key findings include a notable increase in citations for journal publications compared to conference papers, a call for more simplification studies in languages other than English, and an emphasis on fostering international collaboration among researchers.

    Keywords: Text Simplification Lexical Simplification Natural Language Processing

  • Study of a pedagogy adopted to generate interest in students taking a programming course

    Explored the effectiveness of a blended teaching approach in a mobile application development course for beginners. The research combines traditional, e-learning, and m-learning strategies to enhance student interest and performance. The results demonstrate statistically significant improvements, offering valuable insights into optimizing teaching methodologies for technical subjects in high demand, such as mobile application development.

    Keywords: Blended Learning Education Research Education Technology

  • Study of a pedagogy adopted to generate interest in students taking a programming course

    Studied the adoption of learning management systems (LMS), mobile computing, and blended learning to improve student engagement and academic performance. It was found that using an LMS positively correlates with increased student engagement. Courses that used various LMS modules saw higher usage by students even outside of class. The research provides insights into how emerging educational technologies can enhance learning outcomes when implemented effectively.

    Keywords: Blended Learning Education Research Education Technology Student Engagement Learning Management Systems

  • A Review of Popular Applications on Google Play - Do They Cater to Visually Impaired Users?

    This project aimed to determine whether the popular applications on Google Play, the official store for Android applications, can be used by people with vision impairment. The accessibility of the applications was tested using an external keyboard, and TalkBack, an accessibility tool developed by Google. It was found that several popular applications on the store were not designed keeping accessibility in mind. A framework is proposed that can be used by developers to improve the accessibility of an application.

    Keywords: Accessibility Mobile Application Development Assistive Technology