SOUTH ASIAN UNIVERSITY
FACULTY OF MATHEMATICs &COMPUTER SCIENCE
DEPARTMENT OF COMPUTER SCIENCE
Proforma for Six-Monthly Progress Reports of PhDStudents
1. Name of the PhD student:
MOHIT DEY
2. Enrollment no. and date of admission:SAU /CS(e)/2020/02
3. Name of the Supervisor(s): Dr ehachi shasrna
4. Period of the progress report:
From Februay 2025 To Sebleng 2026
Adetailed progress report summarizing the work done during the evaluation period and duly
signed by the PhD student must be attached.
Date: (Sign. of the student)
Remarks of the Supervisor:
gts eoctension
(Sign. of the supervisor)
Recommendation ofRAC
Mr/ Mehit dag pYesented his/her progress report before the RAC on
Recommendation of RPC:
(Supervisor) (Member) (Member)
Recommendation of BOS
(Dean)
Progress Report
lam Mohit Dey, Enrolment No. SAU/CSPY2020/03, Ph.D. scholar from the
Department of Computer Science and Engineering, enrolled on 26th October
2020, completing the Sth year of my Ph.D. in October 2025. As per the
university guidelines,I am submitting this progress report along with a request
for extension of my Ph.D. tenure for 6months I have completed four technical
research problems. The summary of thw work is given below. While, I am
waiting for the final acceptance of the submitted papers.
Research Problem 1
Identification of DDoS Attack Using Activity Pattern of loT Devices Preserving
Data Privacy
Summary:
The paper presents a method to identify DDoS attacks in loT networks by
analysing the activity patterns of devices in a distributed environment while
preserving the privacy of the data exchanged, using the philosophy of Federated
learning.
Status:
Paper was presented at the 3rd ASIANCON, Title" Identification of DDoS
Attack Using Activity Pattern of IoT Devices Preserving Data Privacy" held in
Pune from 25th to 27th August 2023. Available on IEEE Xplore, Conference
proceeding at DOI: 10.1109/ASIANCON58793.2023.10270584
Research Problem 2
Impact of Various Entropy Measures on Federated Hoeffding Tree Based
Learning for Intrusion Detection in loT
Summary:
The paper presents adecentralized intrusion detection framework integrating
Hoefiding Tree-based incremental learning with federated learning, enhanced
by generalized entropy measures including Rényi and Tsallis entropies.
Status:
Security and
Revised version of the manuscript has been submitted to Wiley's
Privacy(E-sci) Journal on 16th September 2025.
Research Problem 3
A Secure Protocol for Detecting Intrusions in loT Traffic Data Streams Using
Federated Learning
Summary:
learning to detect
The paper presents a secure protocol utilizing Federated
intrusions in IoT traffic data streams, which is essential for safeguarding the
network of loT devices. The proposed approach integrates Hoeffding Tree
both scalability
based learning within a federated learning framework to ensure
and privacy preservation.
Status:
on 22nd
Paper submitted to Springer's Cluster Computing(E-sci, Q1) Journal
May 2025 for peer review.
Research Problem 4
Detection
Secure Federated Maximum Entropy Classifier for IoT Intrusion
Summary:
ASecure Federated Maximum Entropy Classifier (SFMEC) is proposed for loT
intrusion detection. The approach leverages maximum entropy principles within
a federated learning setting to improve detection accuracy while maintaining
data privacy across distributed IoT devices.
Status:
Manuscript is ready and is to be submitted to IEEE transactions on Information
Forensics and Security.