Selected Tutorials

S.No.Tutorial Title
1Malware Analysis and Detection
Mohit Sewak (Microsoft, India), Hemant Rathore (BITS Pilani, Goa)
2Neuro-Symbolic Techniques for XAI and Logical Reasoning
Raghava Mutharaju (IIITD, Delhi)
3Federated Learning in the Real-World: From Theory to Practice
Tushar Semwal(Microsoft, India), Madhusudhanan Krishnamoorthy(Microsoft, India), Rajeev Gupta(Microsoft, India)
4Self-Supervised Learning to Process Labeled and Unlabeled Medical Image Data
Mayuri Mehta (SCET, Surat)

1. Malware Analysis and Detection

Speakers:

  • Mohit Sewak (Microsoft, India)
  • Hemant Rathore (BITS Pilani, Goa)

Brief outline of the tutorial:

Today computing devices like laptops, mobile phones, smart devices, etc., have penetrated very deep into our modern society and have become an integral part of our daily lives. Currently, more than half of the world's population uses computers/mobile devices for their professional/ personal needs. However, these computing devices are targeted by malware designers encouraged by profits/gains associated with the attack. According to a recent report, monetary losses due to cybercrime are expected to reach 10 trillion dollars annually by 2025. The primary role in providing defense against malware attacks is designed and developed by the anti-malware community (researchers and the anti-virus industry). Traditionally anti-viruses are based on the signature, heuristic, and behavior based detection engines. However, these engines are unable to detect next-generation polymorphic and metamorphic malware. Thus researchers have started developing malware detection engines based on machine learning to complement the existing anti-virus engines. However, there are many open research challenges in these models like adversarial robustness, explainability, fairness, etc., which we are going to discuss in detailduring the tutorial.

Speakers Bio

Mohit Sewak is an Artificial Intelligence and Cybersecurity researcher with over 15 years of experience in designing innovative AI software and solutions. Mohit holds more than a dozen patents across the US, India, and worldwide for innovative AI solutions that empower many international products. Mohit is the author of multiple AI book titles on topics including technologies like Deep Reinforcement Learning and Convolutional Neural Networks. Mohit's research is focused on designing AI-based malware and other advanced threat detection and protection systems. Currently, Mohit serves as a Principal Data Scientist for Security & Compliance Research at Microsoft R&D.
Hemant Rathore is a cyber security expert with more than ten years of experience in industry and academia. His current work focuses on the topic of Adversarial Robustness and Explainability in Malware Detection Models. His research interests are in the area of Malware Analysis, Network Security, Machine Learning, and Operating Systems. He has guided several undergraduate and postgraduate students in their independent research projects and published many research papers in reputed journals/ conferences.

2. Neuro-Symbolic Techniques for XAI and Logical Reasoning

Speakers:

  • Raghava Mutharaju (IIIT Delhi, Delhi)

Brief outline of the tutorial:

Neuro-Symbolic AI brings together the neural and symbolic aspects of AI. Symbolic techniques are transparent with provable guarantees for correctness. On the other hand, neural techniques are robust to noise and can easily pick up the patterns from the data. By combining the complementary strengths of these two approaches, it is possible to build AI systems that are robust and transparent. In this tutorial, we will discuss the use of neuro-symbolic techniques for explainable AI (XAI) and logical reasoning over Knowledge Graphs and ontologies.

Speakers Bio

Raghava Mutharaju is an Assistant Professor in the Computer Science and Engineering department of IIIT-Delhi, India and leads the Knowledgeable Computing and Reasoning (KRaCR; pronounced as cracker) Lab. He got his PhD in Computer Science and Engineering from Wright State University, USA, in 2016. He has worked in Industry research labs such as GE Research, IBM Research, Bell Labs, and Xerox Research. His research interest is in Semantic Web and in general in Knowledge Representation and Reasoning. This includes knowledge graphs, ontology modelling, reasoning, querying, and its applications. He has published at several venues such as ISWC, ESWC, ECAI, and WISE. He has co-organized workshops at ISWC 2020, WWW 2019, WebSci 2017, ISWC 2015 and tutorials at ISWC 2019, IJCAI 2016, AAAI 2015 and ISWC 2014. He is/has been on the Program Committee of several (Semantic) Web conferences such as AAAI, WWW, ISWC, ESWC, CIKM, K-CAP and SEMANTiCS. More information is available on his lab's homepage at https://kracr.iiitd.edu.in/.

3. Federated Learning in the Real-World: From Theory to Practice

Speakers:

  • Tushar Semwal (Microsoft, India)
  • Madhusudhanan Krishnamoorthy(Microsoft, India)
  • Rajeev Gupta(Microsoft, India)

Brief outline of the tutorial:

With the advent of the Internet of Things (IoT), there has been a huge surge in the volume of data collected by the devices at the edge of a network. This data is often collected and stored in the remote cloud servers to gain useful insights by training a model on this data. As an alternative, Federated Learning has been proposed where instead of learning a single global model centrally at the cloud server, each participating client device trains a model on its own local data and only share the weight gradients with the shared global model. Thus, in contrast to the sharing of raw data, the weights of the model are shared and distributed across the federation of client devices. One important use case could be an IoT in medical and health environments, Federated Learning (FL) can enable other organizations, which have similar data and have similar modelling requirements, to train in a single better global model which can then be distributed to each participating institute. In this tutorial, we will begin by providing a formal definition of FL, basic terminologies, architectures, and overview of challenges associated with centralized machine learning paradigms. We will then describe the federated learning framework through its various flavors such as horizontal federated learning, vertical federated learning, and federated transfer learning. In addition, the tutorial will also cover our recent published work on Federated Transfer learning. In this full day tutorial, after the first half of discussing the theoretical aspects of FL, the second half will begin with a hands-on introduction to python-based programming of a simple FL algorithm and testing on a benchmark dataset. In the tutorial, we will also introduce a fresh domain on Federated Graph Learning where the different components of FL are adapted for graph datasets.

Speakers Bio

Tushar Semwal is an Applied Scientist at Microsoft Search Assistant & Intelligence (MSAI), India. He got both master’s and PhD degrees in Computer Science and Engineering from IIT Guwahati. Before joining Microsoft, Tushar served as a Research Associate at the Soft Computing Labs in the University of Edinburgh, Scotland. He is a recipient of the prestigious 4-year research fellowship award from TCS India, for his industry applicable research. He has won several travel grants from Microsoft, SERB India, SIAM, and TCS. His research interests include privacy-aware ML, Graph Representation Learning, and large-scale distributed systems.
Madhusudhanan Krishnamoorthy is a Senior Data Scientist at Microsoft Search Assistant & Intelligence (MSAI), India. He got his master's in Data science and Engineering from BITS Pilani. Before joining Microsoft, Madhu served as a Chief Data Scientist in Bank of America. He has more than 7 publications and 72 patents in the areas of cybersecurity, Mixed Reality, LiFi, Information extraction and cellular automata. His current work focuses on Graph representation learning and serving of embeddings at a larger scale.
Rajeev Gupta is a Principal Applied Scientist at Microsoft Search Assistant & Intelligence (MSAI), India. He got his PhD from Indian Institute of Technology (IIT) Mumbai (Bombay) in the area of distributed data management. He has more than 30 publications and 20 patents in the areas of data management, information extraction, and distributed computing in reputed conferences and journals.

4. Self-Supervised Learning to Process Labeled and Unlabeled Medical Image Data

Speakers:

  • Mayuri Mehta (SCET, Surat)

Brief outline of the tutorial:

Medical imaging plays a significant role in developing automated clinical applications for early detection, monitoring, diagnosis, and treatment evaluation of various medical conditions. Deep learning is essential for in-depth and accurate analysis of medical images. Specifically, deep convolutional networks are most appropriate for extracting meaningful features from medical images. A huge amount of labeled data is required to train these deep convolutional networks. However, manually labeling medical images is time-consuming and expensive for medical experts. In addition, the major issue with the manual labeling of the huge dataset is the bias among human annotators. Therefore, applied deep learning is essential to process the dataset having a few labeled and largely unlabeled data. Applied deep learning includes semi-supervised learning, Self-Supervised Learning (SSL) and reinforcement learning. Among them, SSL has been widely used in recent years to process medical data to reduce the data labeling cost and leverage the unlabeled data pool. SSL attempts to learn the visual representations of the data using proxy tasks perceived as pretext tasks. Pretext tasks are responsible for learning the prominent visual representations of data to use the learned representations or model weights obtained in the process for the downstream task. The first half of this tutorial will comprise the emergence of AI in Healthcare, the significance of applied deep learning to process big healthcare data, various self-supervised learning frameworks, different types of pretext tasks, and how to design or select a suitable pretext task for processing medical images. In addition, medical image processing with labeled and unlabeled datasets will be discussed. In the second half of the tutorial, various SSL-based healthcare solutions (use cases) will be discussed. The discussion of each use case will include motivation, precise problem statement, the proposed solution, dataset, experimental results and challenges faced. Subsequently, the functioning of these healthcare solutions will be demonstrated. Finally, challenges and enormous future research opportunities will be discussed.

Speakers Bio

Dr. Mayuri Mehta is a passionate learner, teacher and researcher. She received a doctorate in Computer Engineering from the National Institute of Technology, Surat, India. Her areas of teaching and research include Data Science, Healthcare Informatics, Machine Learning/Deep Learning, Computer Algorithms and Python Programming. Her 22 years of professional experience includes several academic and research achievements along with administrative and organizational capabilities. She is awarded the "Researcher of the Year Award (Engineering, Female)" by the 3rd International Business and Academic Excellence Award (IBAE-2021) committee for her Exceptional Calibre and Outstanding Performance as an Academician, Researcher, Mentor, Advisor, and a Thought Leader. She has 11 patents and 1 copyright to her credit. She has published two books: (1) Tracking and Preventing Diseases with Artificial Intelligence and (2) Knowledge Modelling and Big Data Analytics in Healthcare with Springer and CRC Press, respectively. Her books on "Explainable AI: Foundations, Methodologies and Applications" and "Recent Advances in Data and Algorithms for e-Government" with Springer will be published this year. She is the author of 33 research papers and 3 book chapters. She has worked on several academic assignments in collaboration with professors of universities across the globe. She has visited Germany, France, Switzerland, Oman, Dubai, Hongkong, Macau and Thailand for professional and personal purposes. She is an adjunct professor at Gujarat's largest private university- Parul University. Her AI-powered Healthcare project was approved for funding by the Multidisciplinary Research Unit of Surat Municipal Institute of Medical Education and Research (SMIMER). She has also received funds several times from Gujarat Council on Science and Technology (GUJCOST). She has received funds from Student Start-Up & Innovation Policy (SSIP), Government of Gujarat, India, for filing 2 patents. She has served in several International Conferences in different positions. She has conducted 80+ sessions in International Conferences, Short Term Training Programs (STTPs), Faculty Development Programs (FDPs), etc. With the noble intention of applying her technical knowledge for societal impact, she is working on several AI-powered research projects in Healthcare in association with doctors doing private practice and doctors of Medical Colleges. She is a member of professional bodies such as IEEE, ISTE, CSI.