Arka Ghosh Email & Phone Number

Guelph, Ontario, Canada

Halifax, Nova Scotia, Canada

• Worked towards the implementation of an anomaly-based Intrusion Detection System (IDS) to effectively detect malicious traffics from normal traffic using Machine Learning and Deep Transfer Learning under the.
• Curated and pre-processed publicly available datasets for intrusion detection research, ensuring cleanliness and relevance, and extracted significant features to reduce the curse of dimensionality and enhance.
• Implemented a chunk-based approach to transform tabular data into RGB images to make it suitable for CNN based image models.
• Developed a weighted ensemble learning strategy based on misclassification rates of three different variants of a pre-trained CNN model for the effective detection of malicious network flows.
• Implemented additional state-of-the-art IDS frameworks for result comparison with the proposed framework.

CSCI 3110: Algorithms I (Summer 2022)CSCI 3120: Operating Systems (Fall 2023)

CSCI 3110: Algorithms I (Summer 2022, Summer 2023 & Summer 2024)CSCI 3171: Network Computing (Fall 2022)CSCI 4174/6708: Advanced Topics in Network Security (Winter 2023 & Winter 2024)CSCI 4146/6409: Process of Data Science (Summer 2023)CSCI: 3120: Operating Systems (Fall 2023)CSCI 4171/6704: Advanced Topics in Networks (Fall 2023)

Halifax, Nova Scotia, Canada

• Worked on a project in collaboration between Scient Analytics Inc. and DeepSense on real-time quantitative analysis of piston core sediment samples combining expertise from Computer Vision and Machine Learning on.
• Delivered weekly project report and collaborated with the advisors from both Scient Analytics Inc. and DeepSense to conduct thorough analysis of challenges and ensure outcomes of the project.
• Utilized OpenCV’s image processing capabilities, i.e. Edge and Contour Detection, Masking, Morphological Operation, to detect and extract Region of Interests (ROI) from the core samples.
• Applied SciPy’s statistical function on the extracted ROIs to derive insights about the properties of the core sample such as length and fracture of the core samples.
• Investigated and implemented popular DANN architectures, i.e., U-net, with large scale imagery training data for semantic segmentation to create binary masks for the core samples.

• Organized tutorial sessions for two consecutive semesters on topics of a second year undergrad course titled Digital Logic Design.
• Collaborated with course instructor to ensure smooth execution of lab sessions and enhance student learning experiences.
• Graded assignments, lab reports, and exams promptly and provided constructive feedback to help students improve their performance.

Master'S Degree, Computer Science, 4.08/4.30

Thesis: ME-IDS: An Ensemble Transfer Learning Framework Based on Misclassified Samples for Intrusion Detection Systems Courses: CSCI 6708.

Bachelor'S Degree, Computer Science And Engineering, 3.62/4.00 (Distinction)

Undergrad Thesis: Urban Sound Classification Using Convolutional Neural Network and Long Short Term Memory Based on Multiple Features