HOME / Tutorial


NOTE: The Tutorial is going to be held on October 19, 2022. Full day session. Interested participant need to register separately via this link.  


T1 - Fundamentals of Tensors with applications (click here to zoom)


Multidimensional data can be represented as a tensor, multi-array, or multi-linear functionals without losing any information.

Mathematically, tensors are a generalization of the concepts of scalars, vectors, and matrices. Tensors occur naturally in this world, just like scalars, vectors, and matrices, with many video and machine learning applications. This tutorial focuses on the mathematical fundamentals of tensors and their application in data compression, machine learning, and other areas.


Dr. Hanumant Singh Shekhawat

IIT Guwahati, India

Bio: Dr. Hanumant Singh Shekhawat is working at the Indian Institute of Technology Guwahati as an Assistant Professor in the Department of Electronics and Electrical Engineering. He did his postdoc at the Department of Electrical Engineering, Eindhoven University of Technology, and The Netherlands. His research was on multi-linear data (tensor) reduction techniques, which have applications in video/speech processing, MRI, and (higher dimensional) data analysis. He completed his Ph.D. in Nov 2012, from the Department of Applied Mathematics, the University of Twente, The Netherlands. His research was related to optimization problems in sampling and interpolation. He has completed his master’s (in 2004) from the Department of Electrical engineering and his bachelor’s in Electronics and communication engineering (in 2002) from the Indian Institute of Technology Bombay and Rajasthan University in India, respectively. After the master’s degree, he worked in Texas Instruments India and Sasken Communications India for around four years in the area of electronics, software, and algorithm development. He had a visiting faculty position at the University of Pardubice, the Czech Republic from May to June 2019. Currently, his work is related to problems in radar, tensor, speech, signals, and systems.

Dr. Shodhan Rao

Ghent University Global Campus, South Korea

Bio: Shodhan Rao is an Associate Professor of Applied Mathematics/ Director of Research Center for Biosystems and Biotech Data Science at Ghent University Global Campus (GUGC), Incheon, South Korea. He is also a part-time (10%) Professor with the Department of Mathematical Modelling and Data Analysis at Ghent University’s main campus. He is currently pursuing research on stability, parameter estimation and model reduction of biochemical reaction networks, validity conditions of quasi-steady-state approximations, and the dynamics of competition network models in ecology. His research interests are in the areas of chemical reaction network theory, systems biology, and mathematical ecology.

Sandeep Pandey

Sandeep Pandey Samsung Research and Development Institute, Bangalore, India.

Bio: Sandeep Pandey is working as a Lead Engineer at Samsung Research and Development Institute, Bangalore, and submitted thesis for Ph.D. at IIT Guwahati. His research interest lies in the area of Tensor-based signal processing, deep learning for emotion recognition in speech and text, mental health diagnosis. He has numerous publications in international conferences and reputed journals such as Biomedical Signal Processing and Control, Plos one.




The objective of the workshop is to demonstrate the drone based common object detections which can be applied to satellite images too for common object detections.

Locating common objects like pedestrians, bikers and cars on roads is essential for city surveillance. This workshop presents a way that can help citywide speed up the pedestrian, bikers and car detection process by the use of drones. Any high number of presence of people, or cars can alert the authorities to take further actions. Drone technology is advancing at a rapid pace, improving drone capabilities whilst an increasingly competitive market is driving prices down. Drone capability is having an overall positive impact on society with its focus on usage in crisis and emergency response, search and rescue, automated shipping, the film industry, farming, environmental management, etc. This increases public surveillance and helps the concerned authorities to develop smart cities. Hands on using deep learning techniques of Yolov5 and other object detection techniques can help the system to detect common objects across the cities.

This orientation includes a participatory activity to introduce the concept of deep learning especially Convolution Neural Network for approaching problems in computer vision domain.


Gaurav Tripathi, Ph.D.

Senior Scientist, Central Research Lab,

Bharat Electronics Limited, India

Bio: Gaurav Tripathi received his M. Tech. degree in Information Technology (Specialization: Artificial Intelligence). Indian Institute of Information Technology, Allahabad in 2007 PhD from Delhi Technological University, India. He currently works as senior scientist at Bharat Electronics Ltd. India. His research interest includes Internet of Things, Deep Learning, Convolutional Neural Networks based Computer Vision, Fog computing.


T3 - Introducing deep learning models for human emotion recognition and Analysis (click here to zoom


Human emotions account for a huge factor in their psychological and physical health. Hence, understanding human emotions is very important. Given the complexities of the human brain and its emotions, developing technologies to understand is a very challenging task. This workshop focuses on developing emotion models designed for humans. This workshop would cover topics such as collecting data and exploring emotion analysis tasks and developing novel methods for such tasks. This workshop would start from the very basics of developing deep learning models and solutions to advanced topics such as feature fusion methods. This work is focused on understanding how seq2seq models can help understand hidden patterns in human emotion, and how can such models be used to develop frameworks to build emotion understanding models. This workshop is focused on researchers and machine learning engineers who are interested to pursue research in the domain of emotion recognition. Some basic concept of machine learning and python is expected in this workshop.


Dr. Naagmani Molakatala

School of Computer and Information Sciences,

University of Hyderabad, India

Bio: Faculty in school of CIS teaching AI and Computer science courses, Experienced Communications Specialist with a demonstrated history of working in the computer software industry. Skilled in C++, Java, Management, Software Development, and Leadership. Strong information technology professional with a B.Tech(ECE) focused in Signal processing from SMGH school Anantapur, Govt Polytechnic Anantapur, JNTU Anantapur, JNTU Hyderabad. PhD from University of Hyderabad

Mr. Shankhanil Ghosh

School of Computer and Information Sciences,

University of Hyderabad, India

Bio: Shankhanil is a deep learning researcher who has worked in ML/DL domain for over two years in academia. He is currently working on a multi-modal deep learning solution for mental health issues detection in teens and young adults. He is focused on developing a software deliverable technology stack that can help psychologists and health care professionals track and monitor the emotional parameters of an individual. His team is a part of the India-Korea joint collaboration on battling the mental health crises. In the past, he has worked on building Connet-NoTouch, a B2B product for offering contactless dining solutions, (one of the first to do so) during the height of COVID19. He targetted smaller restaurants in the Tier-2 and Tier-3 cities in Bengal, where technological penetration was low, and thus heavily affected by COVID19.


T4 - Recent Advancement in Deep Learning: Federated Learning and Self-Supervised Learning


Recent advancements in deep learning are solving real-world problems. Deep learning algorithms need a more diverse set of data to generalize and perform in the real world. These algorithms are largely dependent on high-quality labeled data. High-quality varied labeled data is a major obstacle in transmitting these technologies to the end users. Self-supervised learning (SSL) is a developing deep learning technique that aims to address the issues raised by the over-dependence on high-quality labeled data. Internet users generate roughly about 2.5 quintillion bytes of data daily. Most of the data is personal user data. As privacy concerns are increasing among Internet users, the privacy protection use of personal data will establish a trust to interact with user devices. Federated Learning is a technique in which we can train a global deep learning algorithm with protecting the privacy of the user's personal data. Federated learning with self-supervision on user data can help to solve many real-world problems and gain trust to interact with technology.


Dr. Bong Jun Choi

Soongsil University, South Korea

Bio: Dr Bong Jun Choi is an associate professor at the School of Computer Science & Engineering and jointly at the School of Electronic Engineering, Soongsil University, Seoul, Korea. Previously, he was an assistant professor at the Department of Computer Science, State University of New York Korea, Korea, and concurrently a research assistant professor at the Department of Computer Science, Stony Brook University, USA. He received his B.Sc. and M.Sc. degrees from Yonsei University, Korea, both in Electrical and Electronics Engineering, and his PhD from the University of Waterloo, Canada, in Electrical and Computer Engineering. His current research focuses on distributed artificial intelligence, distributed intelligent energy networks, and security. He is a senior member of the IEEE and a member of ACM.

Dr. Ajit Kumar

Soongsil University, South Korea

Bio: Dr Ajit Kumar is a Post Doctoral researcher at Soongsil University, Seoul, South Korea. Currently, he is working in a project titled "Korea-India Joint Network Center (JCN) on Depression Diagnosis and Medication Adherence (우울증 진단 및 약물 순응도 연구 센터)" funded by MSIT, Korea". He completed his PhD in Computer Science and Engineering from the Department of Computer Science, Pondicherry University, in May 2018. His research involves applying machine learning to solve various cyber security issues. Apart from his core research area, he works with other researchers to extend the application of machine learning to other domains. He has published his research works in SCI journals and international conferences. He has won the best paper awards at two conferences for his research works. He is passionate about sharing his skills and knowledge with communities and young researchers.

Mr. Ankit Kumar 

Soongsil University, South Korea

Bio: Ankit Kumar Singh is currently pursuing his Ph.D. in the field of Artificial Intelligence at Soongsil University under the supervision of Prof. Bong Jun Choi. He is a team member of the “Korea-India Joint Network Center (JCN) on Depression Diagnosis and Medication Adherence” project. His research interest includes Privacy Preserving Mental health using Multimodal data. He has worked in ‘Inferigence Quotient’ for a year as a Computer Vision Engineer. He also has experience working as an Android Application Developer. He received his Master's Degree in Computer Application from Pondicherry University. During his master degree, he founded a Student Club ‘HashInclude’ for collaborative learning dedicated to Problem Solving skills and programming.