Lopata Hall, Room 101
"Cyber Security For Next-Generation Healthcare"
Adviser: Raj Jain
In the healthcare systems ensuring the privacy and integrity of medical data is highly crucial. Next-generation healthcare frameworks in IoT domain collect medical data from patients at home/hospitals that can be wirelessly transmitted by distributed sensor networks through an access network to multiple clouds where it can be further processed and analyzed for real time diagnosis. Though researchers have adopted many machine learning based intrusion detection systems to detect malicious attacks in various domains, yet, these suffer from high false positive rates, and are unsuitable in catering to the needs of medical domain. However, exploring distributed deep learning approaches on IoT, coupled with the innovative physical layer security at the device level is a novel effort. The transformational value of the approach lies in integrating the work in device and device-to-device communication security in the IoT domain with the multi-cloud hierarchy in the cloud domain that would tilt the balance in favor of benefits as against the risks of IoT in healthcare domain.
"Extending Blockchains for Collaborative Decision Making and Risk Assessment Applications"
Adviser: Raj Jain
A blockchain provides a secured paradigm to achieve consensus using a distributed and peer-to-peer network in which no trusted central party is required. As a result, it has the potential to resolve many challenges that are faced with current centralized controllers in globally distributed applications. To date, the technology has been used as a distributed database to record transactions and track objects.
This talk introduces probabilistic blockchain, a novel extension of the current blockchain that allows an enclosed conversion of blockchain data to useful knowledge. The proposed approach can combine blockchains and artificial intelligence to facilitate building next-generation intelligent and knowledge-based blockchains. It extends blockchain’s applications to building efficient risk assessment and collaborative decision-making applications, where a group of decision-makers participates in an event analysis from different perspectives. Applications include (but not limited to) intrusion and malware detections, stock market predictions, insurance, and recommendation systems. A blockchain-based intrusion detection will be presented to show the feasibility of the proposed approach.