{"id":864,"date":"2022-05-17T14:54:58","date_gmt":"2022-05-17T14:54:58","guid":{"rendered":"https:\/\/carleton.ca\/ericsson\/?p=864"},"modified":"2025-04-14T11:29:40","modified_gmt":"2025-04-14T15:29:40","slug":"security-in-ml-applications-a-priority-for-ericsson-fellow-xinrui-zhang","status":"publish","type":"post","link":"https:\/\/carleton.ca\/ericsson\/2022\/security-in-ml-applications-a-priority-for-ericsson-fellow-xinrui-zhang\/","title":{"rendered":"Security in ML Applications a Priority for Ericsson Fellow Xinrui Zhang"},"content":{"rendered":"\n
\n
\n\n
\n
\n
\n

\n Security in ML Applications a Priority for Ericsson Fellow Xinrui Zhang\n <\/h1>\n \n \n <\/header>\n\n <\/div>\n\n <\/div>\n\n <\/div>\n<\/section>\n\n\n\n\n\n

By Ellen Tsaprailis
Photos by Lindsay Ralph<\/em><\/p>\n\n\n\n

Ericsson Fellow Xinrui Zhang<\/a> is researching a pattern-driven approach to make Machine Learning Operations (MLOps) safer and more secure while maintaining reasonable efficiency.<\/p>\n\n\n\n

Since the effectiveness of ML has been validated in various fields, there are tremendous needs to develop diverse ML-based systems and maintain their performances during operation time. Consequently, the slow and manual development and operation of ML-based systems does not fit current business environments.<\/p>\n\n\n\n

\u201cWhile creating efficiencies in developing ML-based systems, security is another top priority as many systems are security-critical and\/or safety-critical. Due to the complexity of ML applications and their security requirements, it is essential to integrate security in the early design phase of the ML lifecycle. Meanwhile, the potential threats and countermeasures to ML should be sufficiently understood to develop secure ML-based applications,\u201d explains Zhang.<\/p>\n\n\n\n

Patterns are a systematic way to capture the experience of experts about good or best practices and documents these nuggets of wisdom in an accessible format for designers. As an extension, a security pattern is a way to document security mechanisms to the recurring threats in a certain context.<\/p>\n\n\n\n

\u201cWe have decided to establish a pattern-driven methodology that can be applied to build and operate secure ML applications, aiming to achieve a reasonable tradeoff of efficiency and cost. Once evaluated, this is very likely going to be helpful to a large organization who frequently develops and operates ML applications like Ericsson,\u201d says Zhang. \u201cSecurity should be valued at least the same or even higher than efficiency and cost when developing and operating security-critical and\/or safety-critical ML-based systems.\u201d<\/p>\n\n\n\n

\"\"
Xinrui Zhang is an Ericsson Fellow and PhD student in electrical and computer engineering<\/figcaption><\/figure>\n\n\n\n

Zhang has written and submitted a paper for publication titled, Security Patterns for Machine Learning: The Data-oriented Stages<\/em> to the 25th European Conference on Pattern Languages of Programs (EuroPLoP 2022). In this paper, a collection of security patterns for the data-oriented stages in the ML workflow is documented, including data collection, data storage, and data preparation. Zhang provides a concise guide on how to protect each stage from known threats, as well as a communication vocabulary for different roles to consider security without being security experts.<\/p>\n\n\n\n

Systems and Computer Engineering<\/a> Assistant Professor Jason Jaskolka<\/a> is supervising Zhang\u2019s research in Carleton\u2019s Cyber Security Evaluation and Assurance (CyberSEA) Research Lab<\/a>.<\/p>\n\n\n\n

\u201cMany critical systems such as those found in the transportation, energy, and e-health domains are evolving with the inclusion of 5G networks and ML solutions. With this evolution comes a wide range of security threats that need to be mitigated to assure safe, secure, and reliable system operation,\u201d says Professor Jaskolka. \u201cXinrui\u2019s research is focused on exploring design solutions for building more secure ML-enabled systems and networks. Her results have the potential to reduce security risks at early stages of development when it is less costly to make changes, which is of interest to Ericsson’s software development divisions.\u201d<\/p>\n\n\n\n

Zhang is one of six graduate students who are Ericsson Fellows<\/a> at Carleton University\u2014a unique, talent-building program born out of the Ericsson-Carleton University Partnership for Research and Leadership in Wireless Networks<\/a>.<\/p>\n\n\n\n

Instead of working as a teaching assistant during their graduate studies, Zhang and the other fellows are being supported to focus on their pioneering wireless communications research and get input from both their academic supervisors and Ericsson professionals.<\/p>\n\n\n\n

With a Bachelor of Engineering degree from Carleton, Zhang took the accelerated pathway and moved straight into her PhD program in electrical and computer engineering. Through her PhD, her work as an Ericsson Fellow and in the CyberSEA Lab, Zhang is committed to working on designing security solutions.<\/p>\n\n\n\n

\"\"
Xinrui Zhang<\/figcaption><\/figure>\n\n\n\n

Continuing her research long-term is important to Zhang as she considers her future career goals.<\/p>\n\n\n\n

\u201cI very much appreciate this great opportunity that Carleton and Ericsson have offered. I have gained a lot of industrial insights from this experience. It would be exciting to continue working with Ericsson, whether directly or as an academic collaborator,\u201d says Zhang.<\/p>\n\n\n\n

Ericsson Fellowship<\/strong>
In this prestigious fellowship program<\/a>, Carleton graduate students conduct hands-on research alongside Ericsson experts in state-of-the-art facilities, ensuring students build skills that are in high demand in today’s telecommunications industry.<\/p>\n\n\n\n

\"\"<\/figure>\n","protected":false},"excerpt":{"rendered":"

By Ellen TsaprailisPhotos by Lindsay Ralph Ericsson Fellow Xinrui Zhang is researching a pattern-driven approach to make Machine Learning Operations (MLOps) safer and more secure while maintaining reasonable efficiency. Since the effectiveness of ML has been validated in various fields, there are tremendous needs to develop diverse ML-based systems and maintain their performances during operation […]<\/p>\n","protected":false},"author":2,"featured_media":880,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":"","_links_to":"","_links_to_target":""},"categories":[32],"tags":[],"class_list":["post-864","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"acf":{"cu_post_thumbnail":false},"_links":{"self":[{"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/posts\/864","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/comments?post=864"}],"version-history":[{"count":5,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/posts\/864\/revisions"}],"predecessor-version":[{"id":881,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/posts\/864\/revisions\/881"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/media\/880"}],"wp:attachment":[{"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/media?parent=864"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/categories?post=864"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/carleton.ca\/ericsson\/wp-json\/wp\/v2\/tags?post=864"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}