Keynote Speakers

Many people assume that they are alone and untrackable in this connected world – they are totally wrong! Many companies, individuals, private bodies and government entities are all watching us and constantly too – without ever getting our permission and/or giving us notice. We ourselves carry many devices that reveal our location, the kind of work we do, all about our friends and relatives and many more.

An approach called, Get, Know and Act All (GKA), which collects many types of datasets and information from many devices and sources in order to provide answers to queries. The underlying rich knowledge base can also capture health related datasets and deep personalized datasets that an individual himself/herself may not know or store in their memory. The approach also employs deep recovery of information so that a person’s possible medical condition, sexual fantasies and other such highly personal information can also be gleaned. This keynote will also explore algorithms relating to computational linguistics, Data Science and large-scale statistics, their architectures and machine learning techniques. While this keynote is technical in nature, will also discuss legal, ethical, moral and professional debates abound this arena, however note that all the datasets are derived from publicly available sources and cradles.

Geospatial technology, often referred as Geographical Information System (GIS), has emerged as an enabling technology for today’s society. In such an information system, locational information plays an important role in spatio-temporal analysis. Geospatial information is increasing playing a crucial role in both societal and economic developments of any nation, including India, impacting the life and livelihood of the citizens. It is a critical tool for making informed decisions on key economic, environmental and social issues, and helps in both short-term and long-term development processes.

In a spatially-enabled society, the activities and events have a geographical and temporal context, make decisions through the effective and efficient use of spatial data, information and services. It is an evolving concept where location, place and other spatial information are available to governments, citizens and commercial organizations for better decision-making. The major challenge in such a system is seamless sharing of diverse geospatial information. This is addressed by adopting a Spatial Data Infrastructure (SDI), which enables the sharing and effective usage of geographic information through standardized formats and protocols for data access and interoperability. SDI, along with Geo-AI, Geo-Cloud and various location-based services, is the backbone for realization of the spatially-enabled society.

Abstract: As data-driven technology advances, more data is generated, placing greater demands on centralized computing and increasing the concern of data breaches. Recent developments in distributed computing have addressed this problem through federated learning. Conventional federated learning consists of various edge devices connected to a single central server. Dependence of federated learning on a single central device for accumulating and aggregating local models makes it vulnerable to single point failures. Additionally, the lack of participation incentives discourages edge devices from using resources for processing data locally. In this context an incentive mechanism can be used to motivate edge devices to contribute. Both challenges can be addressed simultaneously by leveraging blockchain within a federated learning framework. In beyond 5G wireless transaction systems, blockchain has proven to increase privacy and security. Blockchain also initiates a federation of trustworthy devices by validating the local models. Blockchain-enabled federated learning comprises of edge devices and miners. Hence the framework is constrained by limited resources of edge devices and miners. Also, with the increased utilization of edge computing, more and more edge devices are being deployed. In blockchain enabled federated learning, an increase in edge devices can increase the burden on miners to verify a considerable number of local updates and may also require the deployment of more miners. Additionally, increasing miners may improve data latency and security. However, it also increases the forking probability. In addition to these challenges, the involvement of slow devices with insufficient or redundant data can slow down the learning process. Devices may also have limited processing power, bandwidth, and storage resources. There are several such challenges in Blockchain Enabled Federated Learning which present opportunities to researchers working in the area of Beyond 5G Networks and Edge Computing, for providing effective solutions. This talk will introduce the audience to the area of Blockchain Enabled Federated Learning and discuss various Challenges and Opportunities in Beyond 5G Networks and Edge Computing.

Abstract: The paradigm of cloud computing has bought about many benefits especially with respect to cost , ease and speed of development. This has been possible due to democratization of server computing and has been an enabler for thousands of projects, studies and companies worldwide. However, this paradigm becomes constrained when regulatory, security and performance requirements - such as data residency, low latency applications come to the fore.

This talk introduces the concept of a satellite location which can be used to host IBM Cloud services using one’s own compute resources. The talk delves into details of how the management of these resources is performed and utilized for the purpose of hosting a Cloud service. The overall architecture and the security framework along with the shared responsibility model is discussed.