AI@Ashoka
The AI@Ashoka initiative represents a significant step in Ashoka University’s commitment to cutting-edge research and interdisciplinary innovation. Led by Prof. Partha P. Chakrabarti (Advisor to Computer Science, Ashoka University, and Professor of Computer Science and Engineering at IIT Kharagpur), this initiative is driving advancements in Artificial Intelligence (AI) and Machine Learning (ML) across diverse fields, from environmental studies and psychology to biology, chemistry, and computer science.
The eleven ongoing projects demonstrate AI’s transformative potential—enhancing fire detection accuracy, biodiversity monitoring, and cognitive assessments in infants, while also advancing cancer research, antiviral drug discovery, and carbon capture solutions. Researchers are leveraging ML-driven computational models to refine diagnostics, optimise therapeutic strategies, and develop scalable solutions to pressing scientific challenges.
By integrating AI into interdisciplinary research, AI@Ashoka is not only expanding the frontiers of knowledge but also addressing critical questions about AI’s ethical, legal, and societal impact—equipping students and researchers to navigate the future of this rapidly evolving field.
Our Projects
Applications of Artificial Intelligence and Machine Learning in inferring the past demographic history of a species
Led by Dr. Balaji Chattopadhyay ( Trivedi School of Biosciences) and Dr. Kritika M. Garg (Ramalingaswami Fellow, Department of Biology), this project aims to address critical challenges in biodiversity conservation.
Human activities have severely disrupted natural habitats, driving the current phase of climate change and triggering the sixth mass extinction event. Understanding how species have historically responded to climatic shifts is essential for predicting their future trajectories. While existing computational methods can reconstruct past population dynamics, they are often inaccessible, have limited resolution in complex demographic scenarios, and require significant computational resources and expertise.
This project seeks to develop an AI/ML-based program capable of predicting population size fluctuations—capturing complex demographic events and subtle variations—by analysing genomic data in past climatic conditions.
Quantifying Fire at High Resolution across India
Led by Dr. Meghna Agarwala (Department of Environmental Studies), this project is working to improve the accuracy of satellite-based fire detection, which currently identifies less than 30% of fire events. Global fire databases, such as MODIS fire products, report even lower accuracy—under 20%—for detecting small-scale fires, leading to significant omissions.
Fires play a crucial role in local ecological processes, such as vegetation dynamics and soil erosion, as well as global climate systems, including hydrological cycles and atmospheric change. Better fire mapping is helping forest departments manage fires more effectively, but current detection methods continue to struggle with misclassification.
The research team is building on previous work with Landsat data, which increased fire detection accuracy to approximately 78% but still faced misclassification challenges. To address this, the project is applying time-series analysis and machine learning (ML) techniques to improve classification by analysing fire trends over time rather than relying on absolute pixel values. This approach is proving effective in agricultural monitoring and forest fire detection in China, demonstrating its potential for broader application.
By refining these techniques, the team is developing a robust, nationwide fire detection model for India, strengthening fire monitoring, management, and climate resilience efforts.
Patterns of gaze fixations to social and non-social stimuli in high-risk and low-risk infants
Under the guidance of Dr. Madhavilatha Maganti, this project is developing objective measures to assess infants’ neural functioning and cognitive development, aiming to create tools for the early identification of cognitive deficits in high-risk infants.
For such assessments to be effective, they must enable deep data collection while being affordable, portable, quick, culturally neutral, and easy to administer without requiring extensive training for experimenters. Eye tracking is emerging as an ideal solution, particularly for infants younger than nine months, as it relies on visual attention rather than motor skills, which can be underdeveloped in high-risk populations.
The study is analysing gaze fixations in both high-risk (preterm and small-for-gestational-age) and low-risk (typically developing) infants at 3, 6, and 9 months of age, using data from 450 infants collected at the Infant Lab at Dr. RML Hospital, Delhi. By examining patterns of eye movements, researchers are working to identify early markers of cognitive differences and developmental delays, offering critical insights for timely intervention and improved neurodevelopmental outcomes.
Cancer Stem Cells Profiling
Under the supervision of Dr. Debayan Gupta (Department of Computer Science), researchers are exploring the role of tumour-resident stem cells, or cancer stem cells (CSCs), in driving treatment resistance and tumour recurrence. These cells, a highly adaptable subgroup within tumours, pose significant challenges for conventional therapies, making their identification a critical area of study.
Existing methods for detecting CSCs struggle with scalability, reproducibility, and consistency across cancer types. To overcome these limitations, the team is employing machine-learning techniques to analyse bulk RNA sequencing (RNAseq) data. In a study of patient samples across 22 cancer types, their approach has uncovered clinical, transcriptomic, and immunological signatures associated with distinct CSC states. These findings offer new insights into cancer progression and potential therapeutic interventions, marking a step forward in precision oncology.
AI-powered analysis of multiparameter flow cytometry data for Indian clinical laboratories
Under the leadership of Dr. Rama Akondy (Department of Biology) and Dr. Debayan Gupta (Department of Computer Science), researchers are developing an AI-based approach for analysing flow cytometry data.
Flow cytometry, a widely used laboratory technique, allows for the simultaneous analysis of multiple cellular properties. While essential in immunology research, it also serves as a diagnostic tool for conditions such as hematologic malignancies. However, diagnosis currently relies on manual interpretation by clinical pathologists, requiring specialised expertise and time-intensive analysis.
By integrating AI-based methods, this project seeks to enhance the speed, objectivity, and efficiency of flow cytometry analysis while also uncovering novel disease patterns and associations. This approach has the potential to transform diagnostic processes, making them more precise and accessible.
Investigating the therapeutic relevance of novel antimicrobial peptide variants against fast-evolving multidrug-resistant pathogens
Under the direction of Dr. Imroze Khan (Department of Biology), this project addresses the growing challenge of antimicrobial resistance (AMR)—a major global health threat caused by microbes developing resistance to widely used antibiotics.
Antimicrobial peptides (AMPs) have emerged as promising alternatives due to their broad-spectrum activity and reduced resistance rates, attributed to their cocktail mode of action. While recent advances in deep learning models have expanded the repertoire of novel AMPs, their clinical success remains limited due to resistance evolution against individual peptides.
To bridge this gap, the study examines the evolutionary basis of AMP diversity and integrates these parameters to design optimised AMP cocktails for therapeutic use. By leveraging evolutionary insights, this approach aims to enhance the efficacy and durability of antimicrobial treatments in the face of rising drug resistance.
AI-driven analyses of confocal micrographs for biomarker discovery to enable targeted cancer therapy.
Supervised by Dr. Kasturi Mitra (Department of Biology), this project focuses on developing computational approaches to identify mitochondria-related quantitative biomarkers for cancer energy metabolism.
The research aims to create AI-ML models capable of predicting mitochondrial function based on mitochondrial structure, leveraging data from the quantitative mito-SiM method developed in the lab. Preliminary models have already been tested for prediction accuracy, and further refinements are underway.
Automated long-term passive acoustic monitoring for biodiversity estimation
Under the leadership of Dr. Bittu Kaveri Rajaraman (Department of Biology and Psychology), researchers are advancing a non-invasive, long-term biodiversity monitoring system using automatic acoustic sampling methods. This approach aims to track the impact of climate change on ecosystems while also testing conservation interventions.
Acoustic sampling allows for continuous, high-resolution data collection, offering insights into biodiversity patterns across daily and seasonal cycles. Such granularity is crucial, as annual or large-scale biodiversity assessments often miss the nuances of ecological variation. Researchers also plan to integrate this method with camera traps and automated image classification, enhancing monitoring accuracy and broadening the scope of observation.
By leveraging automation, the project seeks to create a scalable, data-driven system for tracking biodiversity shifts over time, providing critical information for conservation efforts in a rapidly changing climate.
High Throughput COF Discovery to Capture CO₂ using ML
Under the direction of Dr. Vidya Avasare (Department of Chemistry), researchers are exploring the potential of covalent organic frameworks (COFs) as efficient materials for capturing atmospheric CO₂ and reducing carbon footprints.
COFs have emerged as promising candidates for carbon capture, with their inherent structural features playing a key role in determining their effectiveness. To better understand and enhance their performance, researchers are employing computational methods to analyse the molecular properties that contribute to CO₂ absorption. Additionally, machine learning techniques are being integrated to accelerate the design and optimisation of high-performance COFs through high-throughput screening.
This research aims to develop next-generation carbon capture solutions, leveraging computational and AI-driven insights to address global climate challenges.
Ensemble learning framework to find potential small molecule inhibitors against NS3 serine protease
Dr. Sourav Chatterjee (Department of Chemistry) and Dr. Rintu Kutum (Faculty Fellow, Department of Computer Science, Ashoka University) are leading a research initiative to develop computationally driven antiviral therapeutics against mosquito-borne flaviviruses such as Zika, Dengue, Yellow Fever, and West Nile viruses.
At the core of this study is the NS3 protein, a multifunctional enzyme that, with the NS2B cofactor, functions as a serine protease essential for viral replication. Disrupting this protease’s activity is lethal to the virus, making it a critical target for drug development.
To identify potential inhibitors, the researchers are developing a computational framework that integrates machine learning and deep learning-based algorithms to screen small molecule libraries. This approach aims to discover active and allosteric site inhibitors of the NS3 serine protease, offering a pathway toward broad-spectrum antiviral therapeutics against dengue and related flavivirus infections.
Assessing AI Tools for Augmenting Synthesis in Research Writing in the Humanities and Social Sciences
The advent of AI language tools like ChatGPT has drastically altered various writing tasks, including ideation, refinement, creativity, and synthesis. However, quantifying and benchmarking the extent to which ChatGPT (and other such tools) satisfactorily execute these tasks, and how they accommodate discourse patterns and implement discipline-specific academic styles in the social sciences and humanities, remains a challenge.
This study aims to design a comprehensive research framework to address these objectives, ultimately providing insights into the effective integration of AI in academic writing across disciplines in the social sciences and humanities.
