Research Areas:
1. Applied AI in Industry
2. Machine Learning
3. Quantum Computing
Projects Handled:
AI for Multivariate Time Series Data-Driven Delignification Process in Paper and Pulp Industry. The main objective is to improve pulp-making quality during the phase of delignification in the paper industry using ML and XAI techniques. AI for Multivariate Time Series Data Driven Delignification Process in Paper and Pulp Industry, collaboration with Chalmers University of Technology and Sodra Varo Pulp Mill under WASP-WISE grant and Vinnova, Govt. of Sweden.
Working on the complex molecular structure analysis and finding the thermodynamically stable materials using AI techniques. It is a part of automatic chemical exfoliation using AI in material science. Develop techniques for classifying whether 3D compounds can be structurally edited by a chemical scissor or not. Considering the relatively small amount of data available, both unsupervised clustering techniques and supervised deep learning techniques will be considered. It has a collaboration with the IFM department, funded by the WASP-WASP call and Wellenberg AI, Govt. of Sweden.
How pre-trained models of 3D materials can be used as a foundation model for training targeted models based on the small number of known training examples.
Worked (Consultant) on Toxicity prediction on drug discovery using ML and DL techniques.
A project work is going on deep learning and game-based cognitive assessment for early dementia detection.
One work is completed on developing end-to-end application to detect various objects (deep learning-based object detection models) from an image & live video and provides a description about them along with any valid relationship exists among the objects.
Worked on a dentistry problem of multi-class classification using Convolutional neural network and its various versions along with the implementation of an IoT based smart toothbrush. I am also working on a game development in healthcare industry where developing an AI-based Game to collect patients data through various levels of gaming strategy and make some decision about his/her health and mental conditions through AI-based analysis.
A new deep learning model (DeepPneuNet) for Pneumonia disease detection purposes from chest X-Ray images has been developed. The work is under revision in a Journal.
A work on the detection of potato leaf diseases (multi-objective and multi-class problem) using ensemble deep learning models has been completed and published.
A research work is going on the human and Ebola protein-protein interaction analysis using supervised ML techniques and deep multilayer perceptron. (Required Skill: Python, R, Keras, Tensorflow, Cytoscape etc.). Submitted in a SCI journal.
Successfully implemented an Image Denoising Technique using Quantum Daubechies Wavelet Transform.
Successfully implemented a quantum hybrid edge extraction technique using the Robinson operator.
A research work has been published in the Biomedical Journal, Elsevier (SCI & SCIE) on the human and SARS-CoV-2 protein-protein interaction analysis using supervised ML techniques and deep multilayer perceptron. (Required Skill: Python, R, Keras, Tensorflow, Cytoscape etc.).
Successfully implemented a quantum hybrid feature selection technique (first of its kind) as a PhD research scholar at University of Calcutta in Applied Intelligence, Springer (SCI & SCIE) Journal. This proposed algorithm has used the concept of correlation coefficient based graph-theoretic classical approach initially and then applied the quantum oracle with CNOT operation to verify whether the dataset is suitable for dimensionality reduction or not. If it is suitable, then our algorithm can efficiently estimate their high correlation values by using quantum parallel amplitude estimation and amplitude amplification techniques.
Successfully supervised my B.Tech students to the prediction and analysis of Protein-Protein interaction as Carcinogenic using Deep Learning Techniques. In this final year project, they use the biological database of Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) datasets (https://string-db.org/). This dataset includes the precisely found proteins accounted for assorted cancer diseases (breast cancer, brain tumor, leukemia) on Homo sapiens. The STRING database contains more viable data than other optional databases such as Biogrid and DIP. The information for carcinogenic proteins is extracted from Swiss-Prot “http://www.bmrb.wisc.edu/data_library/Diseases/” and “https://www.uniprot.org”. After removing the duplicate protein interactions and the self-interactions, our dataset contains 466 protein interactions that includes 298 positive interactions and 168 negatives. Our proposed dataset is publicly available now at this link, “https://datappi.000webhostapp.com/”. As a very new technology, deep learning can play a very good role for the above works in classical as well as quantum domain.
Successfully developed a hybrid quantum version of clustering technique as a PhD research scholar at University of Calcutta. IJQI, World Scientific journal (SCIE) has been published our work in the field of Quantum Clustering on regular bipartite graph and 1-d lattice using the concept of discrete time quantum walks.
Successfully implemented the incremental versions of two widely used clustering algorithms (K-means and DBSCAN) that they can easily handle the dynamic databases and consume less time to cluster those dynamic data compare to the existing algorithms at NIT Raipur 2011. (Required Skill: Java and MySQL)
Successfully analyzed and developed as a supervisor, an automatic cursor movement strategy of a single target and multi-target based Brain-Computer interfacing system using unsupervised learning. One efficient multithreaded algorithm had been implemented which tried to reset the cursor position based on peak amplitude of the brain signal so that cursor moves nearer to the target. Along with it, an EEG based emotion detection technique with the help of new dimension reduction scheme (Correlation Based Subset Selection: CSS) was also developed. (Published by two SCI Elsevier Journals (BICA) and also received IEEE Young Professional Best Paper Award in Springer CICBA 2017). (Required Skill: R and SignalPlant)
Successfully completed a work to develop a classical approach of feature selection using graph-theoretic heuristic and hill climbing with the team of some Data Science in University of Calcutta and IEM Kolkata. (Required Skill: R) (Published in Pattern Analysis and Applications Journal, Springer, 2017).
Successfully implemented sentiment-focused web crawling framework to facilitate the quick discovery of sentimental contents of movie reviews and hotel reviews, and analysis of the same using K-Nearest Neighbour (K-NN) and Naïve Bayes classifiers.
Funded Projects:
Working on the application for a European Union project call on ‘AI in Materials Science’, focusing on developing a multimodal foundation model to analyze material structures across diverse datasets. The project aims to build a pipeline for property prediction, classification, and the discovery of new materials.
A grant of Rs. 50,000/- received on setting up with Workshop (hands on Training) and Faculty Development program under IPR cell of Techno International New Town Kolkata from West Bengal Department of Science and Technology (WBDST) 2023-2024 as a co-convener.
Ph.D. Thesis Supervised:
Co-supervising work on 'Disease Diagnosis and Detection in Healthcare 5.0 using AI-Enabled Deep Learning Techniques', at the Indian Institute of Information Technology (IIIT), Dharwad, India. Start Year - 2025.
Ms. Paramita Kundu Maji, Assistant Professor, Techno International, New Town, Kolkata, India and working under MAKAUT.
Co-supervising work on 'AI-Driven Approaches for Forecasting and Classifying Long and Short-Term Time Series Data with Explainability', at Techno India University, Kolkata, India. Start Year - 2025.
Mr. Sudipta Dutta, Assistant Professor, B.P.Poddar Institute of Management and Technology, Kolkata, India and working under MAKAUT.
Co-supervising work on 'Integrating Secure Cryptographic Techniques with AI-Driven Trade Prediction: A Study for Financial Data Protection and Market Forecasting', at Sister Nivedita University, Kolkata, India, Techno India Group. Start Year - 2025.
Ms. Sangeeta Banik, Assistant Registrar, Sister Nivedita University, Kolkata, India, Techno India Group.
PG/Master's level Thesis Supervised:
Year: 2016-2018
Title: Design and Implementation of an EEG Based Brain-Computer Interface System using Supervised and Unsupervised Learning
Name of the scholar: Mr. Debashis Das Chakladar
Organization: Institute of Engineering & Management.
Year: 2013-2015
Title: Weather Forecasting using Convex-Hull and DBSCAN Clustering techniques
Name of the scholar: Mr. Ratul Dey
Organization: Institute of Engineering & Management.
Year: 2013-2015
Title: A New Load Balancing Approach in Cloud Environment
Name of the scholar: Mr. Nilotpal Choudhury
Organization: Institute of Engineering & Management.
Served as a Reviewer :
Worked as a reviewer in IEEE Transactions on Emerging Topics in Computing, Neurocomputing, IEEE Transactions on Neural Systems & Rehabilitation Engineering, Engineering Applications of Artificial Intelligence (Elsevier), IEEE Transactions on Quantum Engineering, Computational Economics, Applied Soft Computing (Elsevier), Journal of Theoretical Physics (Springer), BMC Medical Informatics and Decision Making Journal (2021), Scientific Reports Springer Nature, IEEE Transactions on Computational Social Systems, International Journal of Machine Learning and Cybernetics (Springer, SCIE & SCOPUS), Concurrency and Computation: Practice and Experience Journal (Wiley), Egyptian Informatics Journal (Elsevier), British Journal of Mathematics & Computer Science (2017, SCOPUS), KSII Transactions on Internet and Information Systems (2018, SCIE & SCOPUS), ICEI-2022, CIT-NITJSR 2021 (Springer), ICDMAI 2021(Springer), IEMCON-2014(Elsevier Conference), IC3T 2015(Springer Conference), ICACCI-2015 (IEEE Conference), IEMCON-2015, IEMCON-2016 (IEEE Conference) and CNST-2016 (IEEE Conference).