EXPERIENCE
• We developed a metadata-based malicious and benign package detection model for the NPM repository.
• We introduced two sets of features, easy to manipulate (ETM) and difficult to manipulate (DTM),
where manipulating DTM depends on long-term planning and monotonic properties. Then, we verify
our feature selection effectiveness based on four well known machine learning and one deep learning
techniques.
• We also analyse algorithms’ performance using metadata manipulation and recommend better metadata adversarial attack-resistant algorithms.
• Experiment analysis shows that our proposed model reduces 97.56% False Positive and 80.35 % False
Negative number.
• Achievement 1: Sajal Halder et al. ”Malicious Package Detection using Metadata Information”, In
submission (A* Conference).
• Achievement 2: Sajal Halder et al. ”Install Time Malicious Package Detection on NPM Repository”,
In process.
Skills: Feature Engineering, Data Scraping, Open Source Software, Machine Learning Model, Deep Learning,
Python, Report/Article Writing, Data Visualization and Teamw
PhD Researcher, RMIT University Melbourne, Australia.
i. Project Title: Capacity aware Fair POI Recommendation using Users Satisfaction.
Innovation: Design a novel fair POI recommendation model considering POI capacity and users preferences.
Language and Tools: Python, Actor-critic network, GCN, Tensorflow, Neural Network, Graph Allocation and Optimization
Duration: July 2021 to date.
ii. Project Title: Deep reinforcement learning of dynamic POI generation and optimization for itinerary
recommendation
Innovation: Propose and implement deep actor-critic network based itinerary recommendation model
named DRLIR.
Language and Tools: Python, AWS, Actor-critic network, GCN, Tensorflow, Neural Network
Duration: January 2021 to June 2021.
iii. Project Title: POI Recommendation with Queuing Time and User Interest Awareness
Innovation: Improve users interest based multi-tasking based transformer architecture models (TLR U
and TLR-M UI) for best POI recommendation.
Language and Tools: Python, PyCharm, Jupyter, Tensorflow, Keras, Matplotlib, Transformer, Lucid
Chart, Overleaf
Duration: March 2020 to December 2020.
iv. Project Title: Transformer-based Multi-tasking for Queuing Time Aware Next POI Recommendation
Innovation: Design and implement multi-tasking based transformer architecture models (TLR and
TLR-M) for better POI recommendation.
Language and Tools: Python, PyCharm, Jupyter, Tensorflow, Keras, Matplotlib, Transformer, Lucid
Chart, Overleaf
Duration: August 2019 to February 2020.
v. Project Title: Efficient Itinerary Recommendation via Personalized POI Selection and Pruning
Innovation: Develop efficient personalized itinerary recommendation model (EffiTourRec).
Language and Tools: R-studio, Lucid Chart, Overleaf, MatLab.
Duration: August 2018 to July 2019.
Research/Project Member
i Research Group: VLSI Research Group, Department o f Computer Science and Engineering, University of Dhaka, Bangladesh
Supervisor: Professor Dr. Hafiz Md. Hasan Babu
Duration: July 2016 to June 2018.
ii Project Name: Establishment of Basic Campus Network in Jagannath University
Source of Fund: HEQEP, UGC, Government of Bangladesh.
Duration: January 2017 to July 2018.
iii Project Name: Establishment of Basic Campus Network on Bangabandhu Sheikh Mujibur Rahman
Science & Technology University,
Source of Fund: HEQEP, UGC, Government of Bangladesh.
Duration: June 2015 to July 2016
Journal/ Conference Reviewers
• Engineering Applications of Artificial Intelligence (Q1 Ranked Journal, IF: 7.8)
• Information Process and Management (Q1 Ranked Journal, IF: 7.47)
• Applied Soft Computing (Q1 Ranked Journal, IF: 8.7)
• Scientific Reports (Q1 Ranked Journal, IF: 4.6)
• IEEE Transactions on Computational Social Systems (Q1 Ranked Journal, IF: 4.74)
• Transactions on Knowledge Discovery from Data (Q1 Ranked Journal, IF: 4.2)
• IEEE Access (Q1 Ranked Journal, IF 3.367)
• Expert System with Applications, (Q1 Ranked Journal, IF 6.954)
• World Wide Web (Q1 Ranked Journal, IF: 3.00)
• Knowledge and Information Systems (Q1 Ranked Journal, IF: 2.82)
• Neural Processing Letters (IF: 3.1)
SELECTED PUBLICATIONS
1. Sajal Halder, Kwan Hui Lim, Jeffrey Chan, and Xiuzhen Zhang.A Survey on Personalized Itinerary
Recommendation: From Optimisation to Deep Learning. In Applied Soft Computing Journal, 111200,
2023, Q1 Ranked Journal, (Impact Factor – 8.7). Publisher: Elsevier.
2. Sajal Halder, Kwan Hui Lim, Jeffrey Chan, and Xiuzhen Zhang. Capacity-aware fair POI recommendation combining transformer neural networks and resource allocation policy. In Applied Soft
Computing Journal, pages 110720, 2023, Q1 Ranked Journal, (Impact Factor – 8.7). Publisher: Elsevier.
3. Sajal Halder, Kwan Hui Lim, Jeffrey Chan, and Xiuzhen Zhang. POI Recommendation with Queuing
Time and User Interest Awareness. In Data Mining and Knowledge Discovery, 2022, Q1 Ranked
Journal, (Impact Factor – 5.406). Publisher: Springer.
4. Sajal Halder, Kwan Hui Lim, Jeffrey Chan, and Xiuzhen Zhang. Efficient Itinerary Recommendation
via Personalised POI Selection and Pruning. In Knowledge and Information Systems, Volume 64, Pages
963–993, 2022. Q1 Ranked Journal, (Impact Factor – 2.53), Publisher: Springer.
5. Sajal Halder, Kwan Hui Lim, Jeffrey Chan and Xiuzhen Zhang. Transformer-based Multi-task
Learning for Queuing Time Aware Next POI Recommendation. In the 25th Pacific-Asia Conference
on Knowledge Discovery and Data Mining (PAKDD-2021), A Ranked Conference.
6. Sajal Halder, Md. Samiullah and Young-Koo Lee. Supergraph based Periodic Patterns Mining in
Dynamic Social Networks. Expert Systems with Applications-An International Journal. volume 72,
pages 430-442, 2017, Q1 Ranked Journal, (Impact Factor – 6.954). Publisher: Elsevier.
7. MS Rahman, Sajal Halder, MA Uddin, UK Acharjee, An efficient hybrid system for anomaly detection in social networks. In Cybersecurity, Q1 ranked Journal, (Impact Factor – 6.27), 2021, Volume 4,
Pages 1-11, Publisher: Springer.
8. Iram Fatima, Sajal Halder, Muhammad Aamir Saleem, Rabia Batool, Muhammad Fahim, YoungKoo Lee, Sungyoung Lee. Smart CDSS: Integration of Social Media Interaction Engine (SMIE) in
Healthcare for Chronic Disease Patients. Multimedia Tools and Applications, pages 1-21, 2013, Q1
Ranked Journal, (Impact Factor – 2.757), Publisher: Springer.
9. Yongkoo Han, Kisung Park, Donghai Guan, Sajal Halder, and Young-Koo Lee. Topological Similarity
based Feature Selection for Graph Classification. In the Computer Journal, bxt123, 2013, Q2 Ranked
Journal, (Impact Factor – 0.755), Publisher: Oxford University Press.