{"id":6346,"date":"2026-04-22T14:15:24","date_gmt":"2026-04-22T18:15:24","guid":{"rendered":"https:\/\/carleton.ca\/sce\/?page_id=6346"},"modified":"2026-04-24T09:34:19","modified_gmt":"2026-04-24T13:34:19","slug":"meng-projects-2","status":"publish","type":"page","link":"https:\/\/carleton.ca\/sce\/meng-projects-2\/","title":{"rendered":"MEng Projects"},"content":{"rendered":"\n
\n
\n\n \n \n \n \n
\n\n
<\/div>\n \n
\n
\n\n

\n 2026 – 2027 MEng Projects\n <\/h1>\n \n <\/header>\n <\/div>\n\n <\/div>\n\n \n\n <\/div>\n<\/section>\n\n\n\n

Multiple Project Supervised by Dr. Mostafa Taha<\/h2>\n\n\n\n

Supervisor:<\/strong> Mostafa Taha<\/p>\n\n\n\n

Website:<\/strong> https:\/\/carleton.ca\/mtaha\/<\/a><\/p>\n\n\n\n

Student Category<\/strong>: M.Eng<\/p>\n\n\n\n

1. NeuroLogic Shield: A Modular Robustness Assessment Platform for Neural and Logic-Based AI Models<\/h3>\n\n\n\n

Project Title:<\/strong> NeuroLogic Shield: A Modular Robustness Assessment Platform for Neural and Logic-Based AI Models<\/p>\n\n\n\n

Project Description:<\/strong> Modern machine learning and AI systems, e.g., whether neural networks and logic\u2011based models, can behave unpredictably when exposed to small, carefully crafted input perturbations. This project focuses on giving students hands\u2011on experience with adversarial robustness, a critical skill for validating AI models before deployment. Students will build a compact \u201cNeuroLogic Shield\u201d platform that evaluates how two types of models: a simple multilayer perceptron (MLP) and a differentiable logic\u2011guided network (DLGN), respond to a single, well\u2011understood adversarial method: the Fast Gradient Sign Method (FGSM). By limiting the scope to one attack and two model classes, the project stays accessible while still offering deep value in understanding model behavior, debugging, and system\u2011level evaluation.<\/p>\n\n\n\n

The student\/team will implement FGSM from first principles, integrate it into a clean and modular evaluation pipeline, and run controlled experiments to assess each model\u2019s robustness. They will design lightweight visualizations and a simple reporting interface that clearly communicates attack impact, accuracy degradation, and model\u2011specific vulnerabilities. The emphasis is on correctness, engineering clarity, and reproducible experimentation rather than building a full adversarial suite. With support in the form of starter code, background training, and close supervision, students can complete a technically meaningful project that teaches practical ML auditing skills and potentially produces a small open\u2011source contribution or workshop\u2011grade technical report.<\/p>\n\n\n\n

2. Explainable Intrusion Detection System (IDS) on FPGA using Differential Logic Gate Networks<\/h4>\n\n\n\n

Project Title:<\/strong> Explainable Intrusion Detection System (IDS) on FPGA using Differential Logic Gate Networks<\/p>\n\n\n\n

In this project, the student will design an Explainable Intrusion Detection System (IDS) using Deep Differential Logic Gate Networks (DDLGNs) and deploy it on an FPGA platform. The main goal is to detect malicious network activity in a way that is both fast and easy to understand. Unlike many traditional machine learning models that act like black boxes, DDLGNs make decisions using logical rules, which allows the system to explain why a network event is classified as normal or malicious. This is especially useful for cybersecurity applications, where trust and clear decision-making are very important.<\/p>\n\n\n\n

The project will involve training a DDLGN-based IDS model, converting it into a logic-gate representation, and deploying it on an FPGA for efficient real-time inference. The system will take selected network features as input, classify the traffic, and provide simple rule-based explanations for its decisions<\/p>\n\n\n\n

3. QuantumVerify: Design and Evaluation of a Post-Quantum Secure Document Signing Platform<\/h4>\n\n\n\n

Project Title: <\/strong>QuantumVerify: Design and Evaluation of a Post-Quantum Secure Document Signing Platform<\/p>\n\n\n\n

As quantum computing advances, many widely used digital\u2011signature schemes\u2014such as RSA and elliptic\u2011curve signatures\u2014face long\u2011term vulnerabilities, raising concerns about the authenticity of documents that must remain trustworthy for years or decades. This project explores that emerging security challenge by building \u201cQuantumVerify,\u201d a functional document\u2011signing platform that relies entirely on standardized post\u2011quantum signature algorithms. Students will integrate CRYSTALS\u2011Dilithium and Falcon into a realistic signing workflow to examine how these quantum\u2011resistant schemes behave in practice. The project blends security fundamentals with applied engineering: instead of studying cryptographic math, students gain hands\u2011on experience with how post\u2011quantum primitives affect file sizes, performance, user workflows, and the feasibility of long\u2011term digital authenticity.<\/p>\n\n\n\n

The student\/team will develop a working document\u2011signing and verification application, run controlled experiments on files of varying sizes, and benchmark key metrics such as signing time, verification latency, key sizes, and signature overhead. The focus is on reproducible engineering evaluation and clear reporting rather than building a full production system. Students will also deploy the platform to a Raspberry\u202fPi or microcontroller to compare desktop\u2011class and embedded\u2011class performance, highlighting real\u2011world constraints faced in IoT, compliance, and low\u2011power environments. The final result offers both practical insight into the trade\u2011offs between Dilithium and Falcon and a hands\u2011on experience with technologies that organizations will increasingly rely on as part of the post\u2011quantum transition.<\/p>\n\n\n\n

4. Privacy and Transparency in Modern AI Services<\/h4>\n\n\n\n

Project Title:<\/strong> Privacy and Transparency in Modern AI Services<\/p>\n\n\n\n

Modern AI services are becoming central to communication, business operations, and software platforms across Canada, yet most users still have little visibility into what happens to their data once it\u2019s submitted to an AI tool. This project explores the emerging \u201cAI Transparency Gap\u201d: a growing uncertainty around who has access to user data, how long it\u2019s stored, whether it contributes to future model training, and where in the world it travels. For organizations required to comply with Canadian privacy principles, especially PIPEDA\u2019s Accountability and Cross\u2011Border Transfer guidelines, this lack of clarity creates practical and legal risks. The goal of this project is to make those hidden processes visible in a way that is understandable to both technical and non\u2011technical audiences.<\/p>\n\n\n\n

The student\/team will investigate how major AI\u2011as\u2011a\u2011Service platforms manage data by creating a structured taxonomy of privacy and access risks, then turning those findings into a public\u2011facing transparency dashboard. Students will analyze documentation, APIs, and platform behaviour to classify services based on access hierarchy, data residency, training ingestion, cross\u2011session memory, and output ownership. Using these insights, they will build a web\u2011based \u201cAI Privacy Pulse\u201d tool that presents simple scorecards, like a nutrition label, for each AI provider. This tool will help Canadian users and small organizations make informed decisions about which AI services align with their privacy expectations.<\/p>\n\n\n\n

Multiple projects on aspects of Software Verification and Validation<\/h2>\n\n\n\n

Project: <\/strong>Multiple projects on aspects of Software Verification and Validation<\/p>\n\n\n\n

Supervisor:<\/strong> Yvan Labiche<\/p>\n\n\n\n

Website:<\/strong> https:\/\/carleton.ca\/squall\/<\/a><\/p>\n\n\n\n

Student Category:<\/strong> UG \/ MASc \/ M.Eng. \/ Ph.D.<\/p>\n\n\n\n

Project Description:<\/strong> The Software Quality Engineering laboratory studies various problems in the field of software verification and validation with the aim to provide sufficient empirical information so that engineers can make informed decisions to use such or such software testing technique. Application domains vary greatly with past work in aerospace, medical imaging, telecommunication, and finance. Problems include the semi-automated construction of tests from plain language specifications, solving the oracle problem (how do we know the outcome of a test execution is what we expect), optimizing software testing from finite state machines and extended finite state machines, and studying the impact of structural coverage principles. Solutions rely on proven, theoretical techniques borrowed from computer science and applied mathematics as well as heuristics, meta-heuristics, machine learning, and AI.<\/p>\n\n\n\n

Pre-requisite:<\/strong> Successful students tend to have background in software engineering, computer science, or computer engineering.<\/p>\n","protected":false},"excerpt":{"rendered":"

Multiple Project Supervised by Dr. Mostafa Taha Supervisor: Mostafa Taha Website: https:\/\/carleton.ca\/mtaha\/ Student Category: M.Eng 1. NeuroLogic Shield: A Modular Robustness Assessment Platform for Neural and Logic-Based AI Models Project Title: NeuroLogic Shield: A Modular Robustness Assessment Platform for Neural and Logic-Based AI Models Project Description: Modern machine learning and AI systems, e.g., whether neural […]<\/p>\n","protected":false},"author":501,"featured_media":6349,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"_acf_changed":false,"_cu_dining_location_slug":"","footnotes":"","_links_to":"","_links_to_target":""},"cu_page_type":[],"class_list":["post-6346","page","type-page","status-publish","has-post-thumbnail","hentry"],"acf":{"cu_post_thumbnail":""},"_links":{"self":[{"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/pages\/6346","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/users\/501"}],"replies":[{"embeddable":true,"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/comments?post=6346"}],"version-history":[{"count":4,"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/pages\/6346\/revisions"}],"predecessor-version":[{"id":6372,"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/pages\/6346\/revisions\/6372"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/media\/6349"}],"wp:attachment":[{"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/media?parent=6346"}],"wp:term":[{"taxonomy":"cu_page_type","embeddable":true,"href":"https:\/\/carleton.ca\/sce\/wp-json\/wp\/v2\/cu_page_type?post=6346"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}