Azure Machine Learning (Azure ML), part of the Cortana Intelligence Suite, makes gaining a deeper knowledge of artificial intelligence easier and faster for computer science students. It also equips them with real skills to succeed in today’s world of advanced automation and data exchange — here’s how.
The Department of Computer Science, Kulliyyah of ICT (KICT), International Islamic University of Malaysia (IIUM) is making improvements in the teaching and learning of Artificial Intelligence (AI) and its related courses. With the increasing specialization of skills needed in AI, existing methods lack supporting tools in the classroom. This may pose a barrier to Bachelor of Computer Science undergraduates gaining sufficient working knowledge of the subject.
In Machine Learning, an upper-level elective module, third-year Computer Science students are expected to understand and apply AI fundamentals to the field of data science. To do this, they had relied on multiple approaches and tools — each performing only a specific process in an experiment. “Without an integrated tool, students had to visualize experiments using their own imagination and often struggled to grasp important concepts,” said Associate Professor Dr. Amelia Ritahani Ismail, lecturer of AI and Machine Learning at IIUM.
Distractions and delays further compromised learning. A portion of lesson time was spent on learning to apply the methods and tools as well as entering large amounts of rudimentary code to function. The consequences were worrying. Students had difficulties meeting learning objectives by the end of the module and faced challenges in applying the concepts for real-life applications. With the university’s standing and students’ future employability at stake, an urgent change was needed to prepare students for Industry 4.0. This is in line with Professor Dr. Abdul Wahab, the Dean of KICT’s aspiration to produce graduates relevant to the industry and society.
Implementing the Azure ML Solution
In 2016, Microsoft first introduced Azure ML, a component within the Cortana Intelligence Suite, to the IIUM Department of Computer Science, initiated by Associate Professor Dr. Normaziah Abdul Aziz as part of KICT Computational Intelligence team’s effort. The cloudbased tool was used to build, test and deploy a wide variety of predictive data analytics solutions.
Dr. Amelia decided to implement Azure ML in the following semester’s Principles of AI module. During the module, second-year students practiced using AI algorithms such as artificial neural networks with Azure ML before they had the option of pursuing the more advanced Machine Learning module in the next academic year.
The response of the students’ exposure to Azure ML was overwhelmingly positive. When enrolment for electives opened the following year, the Machine Learning module saw an increase in subscription. More students had opted to take the 14-week course — the first to be taught using the Azure ML platform.
A Streamlined Learning Process with Azure ML
Students of the Machine Learning module could now perform entire experiments on Azure ML’s integrated, visual-centric platform. “Once students input a dataset or algorithm, Azure ML can visualize it very well and they can understand easily. Even for large datasets, they can see everything at a glance,” said Dr. Amelia. Azure ML’s cloud-based platform further streamlined the learning process. Gone were lengthy installations and technical woes. Hosted on a secure remote server, Azure ML offered a high level of reliability. “In the four semesters using it, I’ve not had any downtime with Azure ML,” said Dr. Amelia.
She noted another plus point: Azure ML had no high-specification hardware demands as the entire service runs on the powerful Microsoft Azure cloud. “As long as there is internet access, I can conduct modules anywhere, using any PC or laptop. Because of this, Azure ML can facilitate mobile learning and let students continue learning outside of lesson time.”
Azure ML as a Platform for Career Excellence
As a result of Azure ML, students of the IIUM’s Machine Learning module saw a significant increase in learning capability. “We used to teach machine learning algorithms and students had difficulty in applying suitable algorithm in their experiment,” said Dr. Amelia.
“Since students started using Azure ML, they’ve been able to understand each phase of experimental design and the experiment as a whole. We were hence able to expand our course outline to teach more algorithms. With a proper experimental design context on the Azure ML platform, students are able to analyze the performances of different algorithms.”
The advantage of learning more algorithms is that students can exercise critical thinking in picking a suitable algorithm to use for a complex project. Already, Dr. Amelia’s students have submitted projects that demonstrate this. Some examples are the Bangla Handwritten Digit Recognition, Sentiment Analysis of corporate videos, weather forecasting on rainfall, and analyzing security data on Botnets, which covered a large variety of outcomes including data preparation, feature engineering, model evaluation and model deployment. “This allows us to expose our undergraduate students working with samples of real industry data analysis,” said Dr. Normaziah.
Many businesses and organizations working with AI and data science today rely on Machine Learning techniques. “Students started getting internships based on their experience with Azure ML in school,” said Dr. Amelia. With her students’ proficiency in using Azure ML leading to a broader understanding of various aspects of artificial intelligence, she believes their future — in light of Industry 4.0 — looks promising.
In addition to Azure ML, several other Computer Science faculty members see opportunities using the tools available in the Cortana Intelligence Suite and the high-quality curriculums provided by the Microsoft Cortana Intelligence Education Program. In doing so, IIUM hopes to enhance its existing courses and offer attractive specialized classes in data science and computational intelligence. “This represents a giant step towards becoming a university of choice in the field of computer science,” added Prof. Dr. Abdul Wahab, the Dean of KICT.