STEM Student Spotlight: Sumaya Soufi, University of McGill
Welcome to our STEM Student Spotlight series, where we highlight the incredible initiatives and achievements led by STEM students in Canada. Today, we are thrilled to feature Sumaya Soufi, a student at McGill University, which is located in Quebec! Sumaya recently participated in our Work-Integrated Learning Program.
Sumaya has been learning about the fundamentals of data analysis and coding through our Machine Learning and Artificial Intelligence courses. In this blog post, we explore how Sumaya’s new technical skills in data analysis combined with her passion for neuroscience is helping her shape her academic and career endeavors.
Here is what she had to say regarding her course on “Unsupervised Learning”, where she gained hands-on experience with project work.
The M2M program allowed me to gain a more in-depth understanding of unsupervised learning, particularly K-Means clustering. It was fascinating to learn how this model groups data into clusters, and I got hands-on experience implementing it using SKLearn in Python. Additionally, I explored evaluating the model’s performance with the concept of inertia, which measures the distance between data points and their centroids after clustering. Preprocessing techniques like standard scalar, column transformers, and pipelines proved essential for producing meaningful clusters.
In addition, Sumaya learned about how to transfer like OneHotEncorder and SimpleImputer in “Empowering Decision Trees”.
“Delving into the concept of decision trees was a real eye-opener. I not only learned how they operate but also how to use them effectively with SKLearn, preventing issues like data leakage. Implementing pipelines for preprocessing decision tree models using transformers like SimpleImputer and OneHotEncoder was a crucial skill that I honed. Moreover, I discovered the significance of optimizing decision trees to avoid overfitting using hyperparameters.”
Lastly, she gained new insights with the Siholette Analysis Method and automated hyper-parameter selection in order to optimize the performance of her models.
“Another fascinating module was the Silhouette Analysis Method, which allowed me to assess how well data points fit within their clusters (cohesion) compared to other clusters (separation). The project associated with this module enabled me to explore various clustering patterns produced by changing different variables.”
“In the final module, I had the opportunity to work on automating and optimizing hyper-parameter selection through Cross-Validation. This technique not only enhanced the performance of the models but also prevented overfitting, ensuring robust generalization to new data.”
Equipped with valuable technical skills to enrich her portfolio and refined problem-solving abilities for future academic pursuits, here is Sumaya’s reflective journey through the program.
“My experience with the M2M Work-Integrated Learning Program has enriched my understanding of both coding and its applications in my field of neuroscience. I am now more passionate than ever about exploring the potential of machine learning and AI in real-life applications. The knowledge and skills I’ve gained have broadened my horizons and opened new doors for me within the field of machine learning and AI. As I continue my academic journey, I am excited to explore opportunities at the intersection of neuroscience and technology, striving to make meaningful contributions that benefit society.”
By blending her curiosity for neuroscience with the power of data-driven insights, Sumaya was able to learn valuable skills that’ll help her in her future academic endeavors and take autonomy in her learning. Stay tuned for more inspiring stories from our STEM Student Spotlight series!
