Aditya, Evan (ML101)
Aditya, Zhi Yi (ML102)
In machine learning (ML), statistical methods are used to empower machines to learn without being programmed explicitly.
The field focuses on letting algorithms learn from the provided data, collect insights, and make predictions on unanalyzed data based on the gathered information. In recent times, technologies such as Computer Vision, Deep Fake and Recommendation Systems have been created with the help of ML.
In our first session, ML101, we gave an introduction about ML, how it differs from the typical programming, and what ML can do. Afterwards we shared about Supervised and Unsupervised Learning, and gave an example each (K-Means and Random Forest). The participants had a chance to better understand the basics of ML, and try them out with a quick activity as well.
An example of Unsupervised Learning
Aditya giving explanations about Supervised Learning
In our next ML Series, we covered about Deep Learning, a subset of ML and explained how it is different compared to ML. Similarly to how humans learn from experience, the deep learning algorithm would perform a task repeatedly, each time tweaking it a little to improve the outcome. With Deep Learning, new applications of Computer Vision have been introduced and have been slowly integrated into our daily lives as well.
We also shared about how open source libraries such as Keras and Tensorflow have been used in both Deep Learning and Machine Learning, where we shared some examples of how we could implement these libraries in our code.
Participants also got a chance to try out how Deep Learning works with an activity involving the Fashion-MNIST dataset, where they tried predicting what fashion category an item belonged to based on it’s image.
An example of how Computer Vision can be done
How models are being trained in ML