Projects and Theses
Advisor: Dr. Anirban Sen and Dr. Dipanjan Chakraborty
Students: Ansh Madan and Anagha Bhavsar
This capstone develops and validates a psychologically grounded framework for understanding why misinformation is persuasive. Rather than focusing only on surface linguistic cues, the project operationalises the Elaboration Likelihood Model (central vs. peripheral routes) alongside key cognitive biases (availability bias,...
Misinformation
Computational Social Science
Bias Detection in Book Reviews: Understanding LLM-Based Bias Detection and Finding Objective Reviews
Advisor: Prof. Lipika Dey
Students: Kunal Singh, Kenisha Chandak
This project investigates seven types of biases—gender, genre, rating, author, popularity, and cultural—in book reviews across multiple datasets using an ensemble of local LLMs (LLaMA, Mistral, Phi, Qwen, and Gemma). The project begins with data collection through web scraping and...
Machine Learning
Natural Language Processing
Advisor: Professor Anirban Sen, Professor Debayan Gupta, Professor Aalok Thakkar
Students: Ananya Basotia, Kashyap J
This project develops India-specific AI safety benchmarks and risk databases focused on education and financial lending. It constructs domain-specific AI Safety Risk Databases and a unified ontology grounded in real-world Indian deployments. In addition, it introduces an education-focused bias benchmark...
AI Safety
Responsible AI
Advisor: Prof. Anirban Sen
Students: Kartikeya Agrawal
This capstone project implements an end-to-end computational pipeline to analyze and quantify media bias within the Indian news ecosystem. Adapting the framework of Saez-Trumper et al. (2013), the project introduces a novel Event Threading algorithm using SBERT embeddings and Time...
Natural Language Processing
Computational Social Science
Advisor: Professor Partha Pratim Das, Professor Anandjit Goswami
Students: Diya Tripathi, Jyotirmay Zamre
This project investigates systematic price distortions faced by Indian electricity Distribution Companies (DISCOMs) arising from divergences between regulated Average Revenue Requirement (ARR) and real-time market prices on the Indian Energy Exchange (IEX). Using granular hourly data from Delhi’s power distribution...
Energy Systems
Machine Learning
Advisor: Dr. Lipika Dey, Dr. Mayank Garg
Students: Himangi Parekh, Soham Shah
Electronic health records (EHRs) contain rich information about a patient’s intensive care unit (ICU) course, but this information is spread across multiple structured tables and is difficult to interpret quickly. This project investigates whether large language models (LLMs) can generate...
Natural Language Processing
Healthcare AI
Advisor: Prof. Debayan Gupta
Students: Chirag S
This project implements and benchmarks two Private Set Intersection (PSI) protocols---OT‑PSI and Laconic PSI---under realistic security parameters. It analyzes the trade‑off between computational cost and communication bandwidth, and empirically verifies the constant‑size receiver communication property of Laconic PSI.
Cryptography
Systems
Advisor: Ms. Lipika Dey
Students: Aryan Verma, Avik Mittal
vid2kg is an end-to-end automated system for extracting structured recipe knowledge from multilingual Indian cooking videos on YouTube. The project converts unstructured, speech-driven video content into machine-readable formats suitable for knowledge graphs and downstream applications such as nutrition analysis, semantic...
Natural Language Processing
Knowledge Graphs
Advisor: Mahabir Prasad Jhanwar, Saravanan Vijayakumaran
Students: Aryan Nath
This project designs a post-quantum secure and privacy-preserving age proof system for India’s Aadhaar identity infrastructure. It replaces the classical RSA signature used in Aadhaar QR codes with the lattice-based Falcon signature scheme and integrates this verification into an incrementally...
Cryptography
Zero-Knowledge Proofs
Privacy
Advisor: Dr. Rintu Kutum
Students: Anurag Moyde
This project designs and evaluates an expert-in-the-loop clinical document digitization platform for IVF medical records in the Indian healthcare context. The system integrates Document Visual Question Answering principles with structured annotation workflows to convert heterogeneous, unstructured IVF documents into schema-validated,...
Artificial Intelligence
Healthcare
Advisor: Prof. Debayan Gupta
Students: Aaryan Talreja
This project presents a comprehensive AI/ML-based deal valuation intelligence system for mergers and acquisitions in India's pharmaceutical and healthcare sector. We developed a systematic framework to classify 150 M&A transactions (totaling $69.5 billion) as undervalued, fairly valued, or overvalued. A...
Machine Learning
Financial Analysis
Advisor: Subhashish Bannerjee; Venkatesh Potluri
Students: Vaanee Tripathi
This capstone explores the challenge of providing real-time accessibility to classroom diagrams for students with visual impairments, focusing specifically on flowcharts in computer science education. Through formative research, grammar design, and dual technical pipelines (traditional computer vision and vision-language models),...
Accessibility
Computer Vision
Advisor: Dr. Rintu Kutum
Students: Tejasdeep Singh
This project develops a scalable, privacy-preserving autonomous prompt optimization framework for clinical documentation tasks. By adapting the ProTeGi (Prompt Optimization with Textual Gradients) methodology to medical summarization, the system leverages local open-source large language models and an LLM-as-a-Judge evaluator to...
Machine Learning
Healthcare AI
Advisor: Prof. Aalok Thakkar
Students: Prabhpreet Singh Setia, Pranav Jayanandan
Transformer architectures have become foundational across computer vision and multimodal learning, yet they remain highly vulnerable to adversarial perturbations. This project integrates global Lipschitz certification with local Jacobian-based attention regularization to improve robustness in Vision Transformers. By combining CertViT’s architecture-level...
Machine Learning
Computer Vision
Advisor: Aalok Thakkar
Students: Kudakwashe Mavis Chakanyuka
This project explores how Human–Computer Interaction (HCI) principles can be applied to improve the usability and accessibility of security protocol verification. Formal verification tools provide strong guarantees about properties such as secrecy and authentication, but their outputs are often difficult...
Human-Computer Interaction
Cybersecurity
Advisor: Prof. Lipika Dey; Prof. Prartha Pratim Das
Students: Ziv Barretto, Jacob Mathew
This project develops a truly agentic AI system for conversational data analytics and forecasting. Instead of running a fixed pipeline, the system autonomously reasons about a user’s natural-language request, routes it to specialized agents (e.g., exploratory data analysis vs. forecasting),...
Artificial Intelligence
Data Analytics
Advisor: Ms. Lipika Dey
Students: Aryan Verma, Avik Mittal
vid2kg is an end-to-end automated system that extracts structured recipe information from unstructured, multilingual Indian cooking videos on YouTube. The project focuses on converting informal spoken recipes into machine-readable formats suitable for food knowledge graphs, search, and downstream analytics.
Machine Learning
Natural Language Processing
Advisor: Professor Sudheendra Hangal
Students: Anirvaan Kar, Myra Malik
Sudoku puzzles have evolved into numerous variations, including colour-based and symbol-based versions. While cognitive load theory suggests that tasks imposing significant information processing demands strain working memory, few studies have examined how representational changes—such as using colours versus numbers—affect cognitive...
Cognitive Science
Human-Computer Interaction
Advisor: Prof. Anirban Sen, Prof. Debayan Gupta, Prof. Aalok Thakkar
Students: Ananya Basotia, Kashyap J
This project focuses on developing AI safety benchmarks and structured risk databases grounded in the Indian socio-technical context. It addresses the gap between global AI risk frameworks and local realities by systematically collecting, categorising, and analysing AI-related harms relevant to...
Artificial Intelligence
AI Safety
Advisor: Dr. Lipika Dey & Dr. Mayank Garg Department of Computer Science Ashoka University Monsoon 2025
Students: Himangi Parekh, Soham Shah Advisors
Electronic health records (EHRs) contain rich information about a patient’s intensive care unit (ICU) course, but this information is spread across multiple structured tables and is difficult to interpret quickly. This thesis investigates whether large language models (LLMs) can generate...
Healthcare
Natural Language Processing