Data Science Academy Camp 2: Down the Rabbit Hole of Machine Learning
COMPFEST 14, Depok — Camp 2 Data Science Academy (DSA) was held online from August 8, 2022 to August 13, 2022 via the Zoom platform. This camp was mentored by Rendi Chevi, an AI Research Scientist from Kata.ai. Curious about the excitement of Camp 2 of DSA? Let’s see what it’s all about!
Day 1 — The Overview
The first day began with remarks from the MCs, followed by a presentation by Rendi Chevi about “Introduction to Machine Learning”. First of all, Rendi gave an overview of machine learning along with a Roadmap of what will be discussed during the second camp of DSA, then Rendi gave a venn diagram of the relationship between machine learning and artificial intelligence which gave rise to a sub-section, namely deep learning.
Up next, Rendi gave a detailed explanation regarding the various types of learning in machine learning, such as supervised learning, unsupervised learning, and deep learning. According to Rendi, the type of learning that is most often used and crucial for daily life needs is supervised learning. Then, Rendi basically explained the difference between unsupervised learning and supervised learning. After Rendi finished with the types of learning, Rendi explained the three main components of machine learning: dataset, model, and criterion. Next, Rendi gave an overview, detailed explanation, some mathematical models, and examples of each of the main components using everyday problems.
Before the end of day one of Camp 2, Academy presented a question and answer session between the participants and Rendi which was followed by a hands-on session, a photo session and finally, a closure by the MCs.
Day 2 — Unsuperficial Supervised Learning
The second day was also filled with a lecture by Rendi Chevi. On this day, Rendi gave some information related to “Supervised Learning”.
The second day started with Rendi giving an overview of supervised learning, which had been explained by himself on the first day of Camp 2 DSA. Rendi started the lecture session by sharing two tasks of supervised learning, namely: regression and classification. The difference between the two is only in what type of data to work with. For regression, the data that is done is continuous, while classification, the data that is done is discrete.
After that, Rendi explained about the perceptron, an example of a model from supervised learning, whereas the model is one of the three main components of machine learning. Perceptron is a versatile model that forms the foundation of artificial neural networks. After giving an overview of the perceptron, Rendi gave various models, functions, and mathematical examples related to how the perceptron works and how it is used. One of the Perceptron functions that Rendi displays is the activation function, which introduces the model to non-linearity properties that convert the data results from a basic numerical form into probability form.
As usual, after the lecture session was finished, it was followed by a hands-on session. On the second day of the hands-on session, participants were given various examples of using the Perceptron model to visualize various sets of numbers into a collection of black and white pixels that are distinguished by the level of opacity of each pixel. Finally, the second day of Camp 2 ended with a question and answer session, photo session, and closing by the MCs.
Day 3 — Unsupervised Learning and its Intricacies
On the third day of Camp 2, the lecture presented was related to “Unsupervised Learning” brought by Rendi Chevi.
The third day’s lecture session began with Rendi giving an overview of 3 common tasks that can be carried out without supervision, namely: dimensionality reduction, which is a task to change the dimensions of data from high to low while maintaining data quality. Clustering, which is a task to break down and group a dataset into various clusters that are unique to each other. And density estimation, which is a task that serves to estimate the probability of the density of data in the dataset.
After that, Rendi gave illustrations and applications for each task. For example, dimensionality reduction and clustering can be used for: data exploration and analysis, high-dimensional data visualization, feature extraction for downstream tasks, and data compression.
Next off, Rendi gave a hands-on session for each unsupervised learning task where participants were given various simple examples for each task that could be easily illustrated. Finally, the third day of Camp 2 ended with a question and answer session, a presentation session by one of the participants, a photo session, and a closing by the MCs.
Day 4 — Digging Deeper Into Deep Learning
The last day of camp 2 started with Rendi referencing a course from the Massachusetts Institute of Technology (MIT), namely “Introduction to Deep Learning” by Alexander Amini. According to Rendi, this course is very good because this course is one of the free and online courses whose Deep Learning materials include modern examples that are solved in a modern way.
The last day’s sharing session starts with Rendi illustrating the timeline about the development of deep learning, then he reminds the participants of some materials which were learned in previous sessions and will be used in today’s session, such as perceptron and activation functions. After that, Rendi dug deeper about perceptron, especially about how to make a neural network using perceptron. After that, Rendi gave various mathematical formulas that will be helpful in deep learning, namely: quantifying loss, loss optimization, gradient sescent, and backpropagation.
Not only that, Rendi continued with a hands-on session in the form of using a more advanced form of perceptron to hone the participants’ understanding. Then, the participants were given a break and then grouped and given time to do a case study that is related to all the material that has been studied during Camp 2. Finally, Data Science Academy Camp 2 closes with a discussion of case study sessions, photo sessions, and farewells from the MCs.
Interview Time
After Camp 2 event was completed, we had the opportunity to interview Rendi Chev, an AI Research Scientist at Kata.ai. Rendi shared his experience regarding why he chose to pursue a career as an AI Research Scientist. At first, Rendi was very interested in the modeling aspect of machine learning and data science. Then, Rendi also advised that if anyone wants to pursue a career in data science, it is better to look for an interest in a specific field related to data science rather than just being interested in data science as a whole.
Besides Rendi, we also had the opportunity to interview one of the DSA participants, Jechonias from team Kuki. According to Jechonias, one of the most memorable things after joining the DSA was the experience of meeting friends, mentors, and professionals who share the same interests as him. After attending Academy, Jechonias plans to create a personal portfolio of his experiences, then he also wants to hone his skills and develop his potential in the field of data science.
There’s more fun at COMPFEST! Keep an eye on information about the excitement of other COMPFEST events by following our social media at COMPFEST on Twitter, Instagram, Facebook, LinkedIn, and our site at compfest.id. Read the complete articles on our Medium page for more about Academy. (Editorial Marketing/Dylan)
