Centre of Excellence in Artificial Intelligence

The Centre of Excellence in Artificial Intelligence at Mahindra University is a small, active and growing group of researchers that came into effect on August 1, 2018. This Centre is located at the Supercomputer Laboratory with advanced NVidia GPU-based supercomputer DGX-1 and other top-of-the-line CPU-based computing servers. The B.Tech program in Artificial Intelligence offered by the Ecole Centrale College of Engineering in Mahindra University is also academically and technically supported by us.  

Key Focus Areas


  • Novel Convolutional Neural Network Architectures and Algorithms
  • Alternative Neural training algorithms, using Evolutionary techniques and Quantum Concepts
  • Cognitive Modelling
  • Scalable Fuzzy Clustering Algorithms on Big Data
  • Game Theory Concepts.

Application Areas

  • Advance prediction and prognostics using different Neural models including LSTMs
  • Multi-Objective Evolutionary Algorithms in Multiple Disciplines
  • Detection and Prevention against Cyberbullying
  • Analysis of acoustics and Structural health monitoring
  • Accurate, real time modeling of complex physical processes.

Supporting Infrastructure

  • Supercomputer Lab with advanced NVidia GPU-based supercomputer DGX-1
  • Top-of-the-line CPU-based computing servers
  • Dassault-Systemes based 3D Experience with underlying Suite of Models
  • Labs and classrooms for discussions, training and workshops.

Our Research Focus Areas

  • Exploration of evolutionary optimization for training in Deep Learning, and concomitant impact on different aspects of DL functionality developed under the back-propagation paradigm
  • Investigating Transpose CNNs with fractional strides in novel architectures designed for complex process monitoring and control in real time using limited sensor data
  • Novel deep learning architectures for incremental learning
  • Optimal synthesis of associative memories
  • Quantum Inspired neural network training algorithm, and Particle Swarm Optimization Algorithm (PMO)
  • Cognitive modelling, through conceptual spaces and social computing, for generation and comprehension of metaphorical mappings
  • Design of scalable fuzzy based clustering algorithm for big data using High Performance Computing (HPC)
  • Learning in absence of sufficient training data: Deep Transfer and Adversarial Learning.
  • Advance prediction of adverse digressions in continuous time systems, using Extreme Learning Machines (ELM), LSTMs and Koopman Operator
  • Game theory based expert systems for complex industrial reactors
  • Applications of Multi-Objective Evolutionary Algorithms (MOEAs) in diverse sectors
  • Applications of Associative Memories for text/audio/video storage and retrieval
  • Detection and prevention against cyberbullying on social media platforms
  • Modelling of cultural and social contexts for identification of inappropriate texts
  • Classification of disease datasets, and medical images for identifying adverse regions
  • Prediction of agricultural crop yields using ANNs

  • Enhancement of agricultural crop yields by developing seeds resistant to drought, heat and other adverse conditions
  • Fuzzy based federated learning model for Internet of Things
  • Analysis of musical features (Acoustic) and properties, Music Information Retrieval and Recommendation systems
  • Structural Health Monitoring of different classes of structures, incorporating evaluation of crack detection and propagation from images and video footages, vibration analysis
  • Learning robust deep models for industry/healthcare related applications inspired from multimedia and IoT
  • Zero-Shot Learning: Identifying samples from those classes that were not observed during training (in the domain of Computer Vision)


External Projects

Title Amount Organization Duration
Missile Launch Improvement using Evolutionary and Multi-Objective Evolutionary Algorithms INR 9.2 Lakhs Defence Research and Development Organization 3 Years (2016 – 2019)
Development of Real-time, Adaptive, Intelligent Mechanisms for monitoring and control of complex industrial processes within Industrial IoT Frameworks 23 Lakhs Cyber-Physical Systems Division, Department of Science & Technology, GoI 3 Years (2018 – 2021)
Development of Warehouse Resource Allocation System 5.2 Lakhs GROUND INC., Japan 0.5 Years (2020)

Internal Projects

Title Amount Organization Duration
Design of a Novel Machine Learning Algorithms using High-Performance Computing for Next Generation Sequence Analysis of Soybean Genomes 1.5 Lakhs Ecole Centrale College of Engineering, Mahindra University, Hyderabad 2 Years (2019 – 2021)
Novel Deep Learning Architectures 0.5 Lakhs Ecole Centrale College of Engineering, Mahindra University, Hyderabad. 2 Years (2019 – 2021)


International Journals
  1. Jha, ​A. Tiwari, N. Bharill, M. Ratnaparkhe, M. Mounika and N. Nagendra, “A Novel Scalable Kernelized Fuzzy Clustering Algorithm Based on In-Memory Computation for Handling Big Data”, IEEE Transactions on Emerging Topics in Computational Intelligence, doi: 10.1109/TETCI.2020.3016302, 2020.
  2. P. Patel, N. Bharill, A. Tiwari, V. Patel, O. Gupta, J. Cao, J. Li, M. Prasad, “Advanced Quantum Based Neural Network Classifier and its Application for Objectionable Web Content Filtering”, IEEE Access, vol. 7, pp. 98069-98082, 2019
  3. Rai and S. Chakraverty (2020), A Survey on Computational Metaphor Processing. ACM Computing Surveys (CSUR), 53(2), 2020, pp. 1-37. DOI: https://doi.org/10.1145/3373265
  4. Rai, A. Jain, and P. Pandey, (2019). Inclusion of Wikipedia, a language specific knowledge resource to generate and update a synset in WordNet. International Journal of Technology, Policy and Management, 19(4), 405-419, 2019, DOI: 10.1504/IJTPM.2019.104062.
International Conferences
  1. Jha, A. Tiwari, N. Bharill, M. Ratnaparkhe, N. Nagendra, and M. Mounika, “Fuzzy Based Kernelized Clustering Algorithms for Handling Big Data Using Apache Spark, Congress on Intelligent Systems CIS, Sep 05-06, 2020, World Conference in Virtual Format (Accepted)
  2. Thapa, D.K. Jain, P. Singh, N. Bharill, A. Gupta, M. Prasad, “Data-Driven Approach based on Feature Selection Technique for Early Diagnosis of Alzheimer Disease”, Proc. of 2020 International Joint Conference on Neural Networks (IJCNN), Glasgow (UK), 19-24 July, 2020
  3. Siva Raju,  A. Sowmya  and  G. Rama  Murthy,”  Emotion  Detection  using  Visual Information  with Deep  Autoencoders,”   Proceedings  of  2018  IEEE  Symposium  Series  on  Computational  Intelligence     ( SSCI  2018 )
  4. Siva Raju,  A. Sowmya  and  G. Rama  Murthy” Emotion Detection using visual information with Deep Auto- Encoders,”  IEEE, 5th International  Conference for Convergence in Technology, 2019
  5. I. Basha and  G. Rama  Murthy, “Emotion  Classification:  Novel  Deep  Learning  Architectures,” Proceedings  of  IEEE  International  Conference  on  Advanced  Computing  & Communication  Systems  (ICACCS), 15th-16th  March  2019
  6. Rama Murthy,  M. Dileep  and  R. Anil, “Novel  Ceiling  Neuron  Model and  its  Applications,”  IEEE  International  Joint  Conference  on  Neural  Networks  ( IJCNN ),  2019,  July  2019,  Budapest,  Hungary
  7. Rama Murthy,  Vamshi Krishna  Reddy,  Devaki  and  Divya, “Optimal  Synthesis  of  Hopfield  Associative  Memory,” Proceedings  of  ICMLDS  2019,  December  2019, ACM  Digital  Library
  8. Rama Murthy,  Vidya Sree,  Jyothi, Mahalakshmi  and  Manasa  Jagannadham,” Deep  Neural  Networks:  Incremental  Learning,”   Proceedings  of  Intellisys 2020,  Springer publishers
  9. Imthiyaz and  G. Rama  Murthy, “Novel  Deep  Learning  Architectures: Feature  Extractor  and  Radial  Basis  Function  Neural  Network”, IEEE International  Conference  on  Computational Performance Evaluation ( ComPE), 2020
  10. Rama Murthy Aman Singh, GC Jyothi Prasanna, Manasa Jagannadan, Maha Lakshmi Bairaju, Vidya Sree Vankam,”1-D/2-D/3-D Hopfield Associative Memories”,  Proceedings  of  International conference on Artificial Intelligence & Machine Learning (ICAIML), 2020
  11. H. Go, T. Jan, M. Mohanty, O. P. Patel, et al., “Visualization Approach for Malware Classification with ResNeXt”, IEEE World Congress on Computational Intelligence (WCCI), Glasgow (UK), July 2020
  12. Kommireddy, P.R. Pandey and K.N. Raghu, “Detection of Heart Arrhythmia Using Hybrid Neural Networks”, IEEE TENCON, Nov 16-19, 2020, Osaka, Japan
  13. M. Gollapudi, M. Perla, S. Hitesh, R. Kumaran, S. Rai and A. Das, “Understanding User Vulnerability Towards Radicalization on Twitter” 5th International Conference on Computational Social Science IC2S 2 July 17-20, 2019, University of Amsterdam, Netherlands
  14. Rai, A. Garg and S. Chakravarty, “Understanding the role of visual features in Emoji Similarity”, International Conference on Intelligent Information Technologies, pp. 89-97, Springer, Singapore, 2018. DOI: https://doi.org/10.1007/978-981-13-3582-2_7
  15. Arulkumaran, S. Rajesh, A.K. Bhattacharya and V.S.K. Narahari, “Real Time Predictions Of Adverse Digressions In Critical and Noisy Industrial Processes Using LSTMs”, IEEE-HYDCON, Sep 2020, Hyderabad
  16. Sahil and A. K. Bhattacharya, “Accurate Replication of Simulations of Governing Equations of Processes in Industry 4.0 Environments with ANNs for Enhanced Monitoring and Control,” 2019 IEEE Symposium Series on Computational Intelligence (SSCI), Xiamen, China, 2019, pp. 1873-1880
  17. R. Annapureddy, A. K. Bhattacharya and N. R. M, “Adaptive Critic Design for Extreme Learning Machines applied to noisy and drifting industrial processes,” 2018 IEEE Symposium Series on Computational Intelligence (SSCI), Bangalore, India, 2018, pp. 327-334.
  18. Ravinithesh, A.K. Bhattacharya and G. Rishita, “Advance Predictions of critical digressions in a noisy industrial process- performance of Extreme Learning Machines versus Artificial Neural Networks”, IFAC-PapersOnLine, Volume 51, Issue 1, 2018, Pages 98-105, DOI: https://doi.org/10.1016/j.ifacol.2018.05.017also at https://www.sciencedirect.com/science/article/pii/S2405896318301770
  19. Thangeda, A.K. Bhattacharya, G. Rajeswari and Ashok Kumar, “Synthesis of Optimal Trajectories in Aerial Engagements using Differential Evolution”, IFAC-PapersOnLine, Volume 51, Issue 1, 2018, Pages 90-97, DOI: https://doi.org/10.1016/j.ifacol.2018.05.016 or https://www.sciencedirect.com/science/article/pii/S2405896318301769
  20. R. Gautam, M. Jahnavi, P. Thangeda and A.K. Bhattacharya, “Synthesis of optimal trajectories in tactical aerial engagements using Multi-Objective Evolutionary Algorithms”, Advances in Multidisciplinary Analysis and Optimization, Lecture Notes in Mechanical Engineering, Springer Singapore, 2021 (Accepted).
Book Chapters
  1. Rai, S. Chakravarty and D.K. Tayal, “Metaphors in Business Applications: Modelling Subjectivity Through Emotions for Metaphor Comprehension”, in Pinarbasi, F., & Taskiran, M. (Ed.), Natural Language Processing for Global and Local Business (pp. 134-153), 2021. IGI Global. http://doi:10.4018/978-1-7998-4240-8.ch006
  2. Narayan and A.K. Bhattacharya, “Accurate, real-time replication of governing equations of physical systems with Transpose CNNs for Industry 4.0 and Digital Twins”, in Machine Learning in Industry, eds. J. P. Davim and S. Datta, ISSN: 2365-0532, Springer Switzerland, 2021 (Accepted).

Computing Facilities in Supercomputer Laboratory

The Supercomputer Lab of Mahindra University is created out of the baseline requirements for supporting high intensity computations for Artificial Intelligence research and applications, that incorporate Machine learning, Deep learning and Data Science.

The core composition of this lab is the DGX-1 supercomputer platform, at whose kernel is a dual-core CPU server with 20 processors and 8 Tesla V100 GPU cards made up of 40,960 Nvidia CUDA cores, all connected through NVLink which minimizes internal communication overheads. On this kernel is built the platform that is a complex stack of components and software including AI Deep Learning frameworks, libraries and drivers.

This software stack is supported by DGX-1 cloud management services which continually provide updates and additional inputs. The software stack is composed of the most popular deep learning frameworks, as well as Nvidia DIGITS deep learning training application, third-party accelerated solutions, the Nvidia Deep learning SDK, Docker and drivers.

Deep Learning tasks, particularly training of complex Artificial Neural Network (ANN) architectures with many layers, hundreds of thousands of parameters, and tens of thousands of data samples, take enormous computing times if the task is performed serially on a single-thread. Instead, when launched on parallel threads, the training data can be split into multiple subsets and each launched on one thread so that the ANN training can be speeded up, ideally, by a factor equal to the number of threads. The GPU-based architecture of DGX-1 works on this principle to speed up AI applications.

The DGX-1 heart of the Mahindra University Supercomputer Lab is reinforced with a number of top-of-the-line CPU Servers that can be seamlessly scaled up in numbers. A fast Infiniband-based network connectivity is in the process of installation to connect these servers so that CPU-based parallelism can also be attained for simulations not attuned to GPU processing.

A Dassault-Systemes based 3D-Experience package is also installed in this lab. This package is an overlay on a suite of software that includes CATIA, DELPHI and other related modules, and facilitates Augmented Reality – Virtual Reality based immersive experience. A set of 30 Workstations coupled with a Server are installed and linked to the other platforms of this laboratory for facilitation of the complete 3D-Experience package. This will enable the attainment of a 3D-Experience Centre of Excellence at this lab.

Thus from the above matter it will be clear that the Supercomputer Laboratory, that is  home to our Centre of Excellence in Artificial Intelligence – also supports 3D-Experience and other facilities for high-intensity computing in a spectrum of domains like Mechanical, Aerospace, Civil, Electrical, Communications and Natural Sciences. At the core of this is of course, massive computing power, enabled by the DGX-1 Supercomputer, multiple powerful Servers, and a set of 30 Workstations, all interconnected through an extremely fast data communication network. What is not so apparent, however, is that these co-located facilities can themselves be the trigger for a new and unique experience in an Indian Research Institution, namely, the coupling of AI with the trans-domain world of engineering in multiple new and creative ways of which visionaries can see possibilities that have yet to be manifested in the mundane world.

The Center will bring together the following Stakeholders