Participants choose their desired topic from the following list:
Cloud is the new way of achieving elastically scalable, self-service computing and applications. Growth of ‘Cloud Infrastructure’ is helping organizations to adopt a ‘Cloud-first’ strategy by moving their on-premise applications safely to the cloud. Cloud is enabling organizations to focus on their core business by minimizing expenditure on computer infrastructure and maintenance, without compromising on the computing requirements of their organizations. It has firmly established itself as the new normal for enterprise IT and business. Ranging from startups to enterprises, it is difficult to find an organization that does not have its workload on the cloud.
Some of the areas of interest for students to work on innovative projects:
The convergence of cloud and mobile computing will continue to promote the growth of centrally coordinated applications that can be delivered to any device. Cloud is the new style of elastically scalable, self-service computing, and both internal and external applications will be built on this new style which employs Software as a Service (SaaS) model. In the near term, the focus for cloud/client will be on synchronizing content and applications across multiple devices and addressing application portability across devices. Over time, applications will evolve to support simultaneous use of multiple devices. In the future, games and enterprise applications alike will use multiple screens and exploit wearables and other devices to deliver an enhanced experience. Example: As the need for cloud based apps and services continues to grow, there will be a need for a development model for these applications and services, which will support the cloud based software development lifecycle.
Microservices
For simple and limited scale applications monolithic architecture is still relevant. Modern cloud application that requires agility scale and reliability, Microservices Architecture offers great promise. A Microservices application is composed of independent components called ‘Microservices’ that work in concert to deliver overall functionality of the application. Unlike monolithic applications, Microservices applications enable the separation of application from underlying IT infrastructure.
Students are encouraged to propose projects on Cloud Applications and Microservices. Projects should be complete in all features and demonstrable while keeping the data secure during rest and during transit.
Some of the areas of interest for students to work on innovative projects:
Depending on the security landscape and attack vectors, security hardening is important for various use-cases of technology implementation. A range of security technologies exist for various domains and their use-cases. Enhancements/improvements and innovations in these security use cases for specific implementations are also required to keep up with the changing landscapes and postures.
Software innovations around security technologies that can handle various use cases are welcome. You can also consider hardware-assisted security models too as part of PoC.
Some of the areas of interest for innovative projects, which also meet the standards/compliance mandates such as, HIPPA, GDPR, ISO 27001, or PCI DSS, are:
Multi-modal biometrics is fast becoming the go-to security option for enterprises that want to offer the most secure and convenient authentication for their customers. Multi-modal authentication benefits the enterprise in specific ways. It vastly improves the customer experience with no cumbersome passwords to remember. It also improves the security of traditional authentication and prevents hacking and data breaches. We are looking for new ideas around it for use cases such as, electronic Know Your Customer (eKYC), user identification/verification, image recognition/verification, or design ideas that can be used across multiple industry/business domains ranging from financial services, transportation & logistics, social services, safe cities, and life sciences & healthcare. All the use cases must also consider meeting the standards/compliance mandates such as HIPPA, GDPR, ISO 27001, or PCI DSS.
Some use cases are as follows:
Fifth Generation Networks (5G), the Internet of Things (IoT), and Edge Computing are essential infrastructure enablers for a range of new business and technology developments. They are together termed as Industry 4.0 that covers areas like autonomous vehicles, smart city grids, e-health, automated factories, mobile content streaming, and data analytics. This marks a significant step on the journey to genuine digital transformation.
5G and Edge computing, along with IoT platforms and frameworks, are key enablers for Industry 4.0. Without these, there will be network problems, not only in providing connectivity for the 20 billion IoT devices but also in transferring and processing the huge volumes of data that will be generated. The problems are not just about bandwidth. Different IoT solutions will have different network requirements. Some devices, like autonomous vehicles and medical equipment demand absolute reliability where low latency will be critical. Other use cases will see networks having to cope with a much higher density of connected devices than seen from traditional 4G networks. The promise of 5G, the development of Network Function Virtualization (NFV), and the ability to process some transactions and store data near to the source of that data, i.e., at the Edge, will be key determinants of the success of Industry 4.0.
Use cases:
Digitization and cheap storage have led to a myriad of data - both structured and unstructured - which we call Big Data. This, along with advances in Machine Learning and Deep Learning, has enabled us in two primary ways. One is to be able to understand the complex interrelationship of data, thus helping in deriving hidden patterns in data. It enables us to make better business decisions through predictive and prescriptive analytics. The second is to automate processes and bring in intelligence into digital things. With faster data management and computing technologies available, it has also become imperative to come up with intelligence based on data and information that is real-time or near real-time. The use of analytics to come up with cognitive (human-like) and intelligent interactions in a B2C scenario is becoming more prevalent. This helps in better personalization and reduces operations costs for businesses. Other than customer-generated data which includes open data sources, and the dark internet, sensors, chips, machines, and other digital devices generate time series/stamped data which also interacts with each other for various outcomes. IoT analytics deals with this stream of analytics primarily.
We are interested in students developing solutions in each of the areas mentioned such as:
Blockchain is a distributed database system - this means, instead of storing files on a single computer, information is stored across millions of computers all over the globe. It gives end-users the power to control their personally identifiable information. We are increasingly observing a wider range of use cases for blockchain across small and large-scale applications.
Blockchain has evolved over the past few years from a very nascent technology to a serious contender for technology disruption. Companies are now at the brink of technological revolution with Blockchain.
Some useful links:
https://research.aimultiple.com/blockchain-applications/
https://thecryptowriter.co/21-top-blockchain-uses-cases-in-2021-with-examples-1f4bbaba474a
Conversational artificial intelligence (AI) refers to technologies, like chatbots or voice assistants, which users can talk to. They use large volumes of data, machine learning, and natural language processing to help imitate human interactions, recognizing speech and text inputs and translating their meanings across various languages.
The world today is evolving towards a universal BoT interface or BoT orchestrator that connects to various BoTs that an enterprise has already invested in and provides one common interface that the user talks to instead of multiple interfaces for multiple assistants. These kinds of platforms help in cost efficiency, increased sales and customer engagement, and easy scalability.
We would be more interested in:
Standard machine learning approaches require centralizing the training data on one machine or in a data center. The downside of this architecture is that all the data collected by local devices and sensors are sent back to the central server for processing, and subsequently returned back to the devices. This round-trip limits a model’s ability to learn in real-time. Also, with the plethora of data privacy laws that have come up, there is a very strong need to localize the data and have a strong governance around it.
Federated learning (FL) in contrast, is an approach that downloads the current model and computes an updated model at the device itself (ala edge computing) using local data. These locally trained models are then sent from the devices back to the central server where they are aggregated, i.e. averaging weights, and then a single consolidated and improved global model is sent back to the devices.
Some of the use cases for federated learning are in :
Mobile devices nowadays contain a variety of personal or even business-related information that is worth being protected from unauthorized access. Owners of such devices should use a passcode or unlock pattern to secure such important assets, but since these techniques are being perceived as annoying barriers, locked devices are not standard. But even if such authentication mechanisms are used, they are very easy to circumvent. The question to answer is “Am I the person using the application, who had logged into the application using Username/Password or MFA?”. Just having PIN lock or phone biometrics is not sufficient in today’s ecosystem.
The expected solution can use behavioral biometrics for continuous authentication. Some ideas in this space are:
Mobile Ad-hoc Network, a.k.a. MANET, is a kind of wireless network, where there need not be any access points or routers, and each mobile node participates in routing data to other nodes. This makes it a decentralized and dynamic network, where these nodes can join and leave at will. Having a constantly changing network topology makes it an interesting area of research. Routing protocols have been developed for different purposes, scale, and performance, like AODV, DSDV, DSR, etc., and many niche use cases have emerged in areas like military, rescue, smart cities. With the advent of 5G, new possibilities have come to the fore, like in AR/VR, drones, vehicles (VANET), IoT, connected devices at home, etc. and this will only intensify in the future. With that comes the valid concern of security, against malicious behavior or compromise of mobile nodes, traffic snooping, man-in-the-middle, etc. Add to that is the use of AI/ML algorithms in preventing, detecting, and fixing the network as well as re-structuring the topology dynamically for better performance. Research on MANETs has again picked up with the onset of 5G, and also WiFi 6. Some of the topics (not limited to) could be:
Problem:
The current pandemic situation forces people to follow certain guidelines, particularly in busy areas, such as stations, airports, and other places or public gatherings. While it is important to allow smooth movement, it is also required to ensure that the guidelines are followed. It is not always possible to have human vigilance at all places. Technology advancement in the field of AI/ML, video analytics, and IoT has made significant progress. The idea proposes using these techniques as a solution to provide a monitoring system that enables administrators to efficiently implement Covid guidelines at public places mentioned above. It will also help in ensuring that the experience is not compromised or at least the impact is minimum.
Solution:
Key features in this application may include:
Problem:
Industry best practices dictate that companies should keep software up-to-date as part of maintaining a healthy security posture. Yet many companies struggle with managing vulnerabilities, especially when it comes to vendor and third-party software. Third party vulnerabilities include vulnerabilities in software dependency packages e.g., Javascript packages, java packages, docker images. Third-party vulnerability management is often a balance between ensuring that critical patches are reviewed and applied quickly, while reducing the risk of downtime due to a potentially unstable patch, also protecting systems from potential security exploits.
The vulnerabilities are identified and published by various organizations such as NIST, GitHub Advisory, mitre. These vulnerabilities are accompanied by information that helps with the remediation.
Examples:
Solution:
The solution designed must have the following
Problem:
Machine learning and deep learning methods can find their use in anti-counterfeiting applications. Artificial intelligence companies use computer vision to recognize fakes. Data scientists design machine learning algorithms to detect details. To a human observer, two shoes can look identical, with one being legit and the other fake. But, data science methods beat the human eye in detecting details. If the fabric pattern is off, deep learning algorithms will see it deviates and report it as a fake.
Why are Anti-counterfeiting solutions required?
Problem:
The banking industry is facing a strong challenge from emerging Fintech companies. Fintech adoption was already growing in 2019 but Covid-19 has rapidly accelerated its adoption globally. In fact, industry data from McKinsey shows that there has been a 4-5% reduction in global cash payments, which is four to five times the recent yearly decrease in cash usage. In addition to influencing consumers to adopt cashless alternatives, Fintech firms have been winning customers through their ability to provide a superior customer experience through hyper-personalized offerings.
However, the battle is not all lost, banks still play a key role in the financial system. They have a large customer base and huge investments in technology infrastructure. This provides them tremendous leverage for any initiatives they could adapt to meet the threat from growing Fintech firms. We believe banks can still win by focusing on building or improving their big data and machine learning investments. Their focus needs to shift from mass-production to mass-personalization of their banking offerings.
Solution:
The solution could ideally focus on one or more of the following capabilities.
End-user Benefits
Problem:
Organizations are struggling to meet the growing demand for software enhancements. Developing and deploying custom software is a critical element of how many companies innovate. With top-performing organizations developing many of their most important software solutions in-house. Recent studies have found, more than half of all software projects are late and over budget, with another 20 percent canceled outright. The application of AI in the software development process promises to mitigate many of these problems.
Solution:
AI can be immensely useful for both new application development and existing applications. In this use case, we will focus on the application of AI to existing applications. The global application modernization services market size is expected to grow from USD 11.4 billion in 2020 to USD 24.8 billion by 2025, at a Compound Annual Growth Rate (CAGR) of 16.8% during the forecast period. AI based solutions can play a significant role in modernizing existing applications and make them cloud native.
AI can be used to refactor applications. Refactoring is defined as “the process of changing a software system in such a way that does not alter the external behavior of the code yet improves its internal structure”. Identifying refactoring opportunities is the first important stage in the refactoring process. The solution should provide developers refactoring recommendations with very less false positives, making developers to confidently use them. As a next step, automatic refactoring should be done to make the solution cloud native. For this project, an existing COM+ application can be refactored and converted to a cloud native application.
Machine learning can be harnessed to predict and automate refactoring operations in COM+ application and refactor them as cloud native applications. ML algorithms should show promising results when applied to different areas of software engineering such as code comprehension, code smells, and automatic refactoring from COM+ application to Cloud native code.
Many organizations still use legacy systems, which create hurdles in the adoption of new cloud native technologies. This AI based solution can help clients cross the hurdle and AI can help them maximize cloud native use with confidence and a higher rate of success.
Participants choose their desired topic from the following list: