Hi @pierreguillou and @jeremy,
Fastai Study Group Experience
My name is Fernando Melo, Co-organizer of Fastai Study Groups in Brasília-Brazil.
Here are some tips to successfully organize Fastai study groups:
We are the organizers of the largest Fastai community in Brazil and we are located in Brasilia, the capital of the country. We also manage the biggest Meetup group of Machine Learning Brasília.
We have already set up many Fastai courses, completely free and open to the community, about machine learning and deep learning in the last two years.
Based on my experience in organizing 4 face-to-face study group courses(more than 50 group meetings), involving from 30 to 100 students each, I believe that the main points for success in organizing Fastai study groups are:
1 - Organizers
A qualified group of organizers who know the material and are committed to the principles of the Fastai Institute, such as empowering the largest number of people and showing that AI and deep learning can and should be developed by ordinary people, with little experience in coding.
The main organizers of the study groups are myself (Fernando Melo) and Erick Muzart and we have the support of several veterans of the previous courses, among which I highlight Pierre Guillou. It is very important to maintain a team of experienced veterans to support the group, especially in practical classes and extra-class projects.
The organizers have to communicate efficiently and might have a group in whatsapp or telegram to exchange ideas, plan meetings and prepare class material.
It is desirable that the organizers have basic knowledge of didactics and experience in lectures and trainings.
2 - Infrastructure
The physical space: classroom, furniture, air conditioning, big screen, wi-fi network, microphones and bathrooms are a big part of success.
Our meeting place is a training center of excellence, newly built, one of the most complete in the city: Instituto Serzedello Correa- TCU.
People want to meet and, also, have a pleasant experience. Our meetings are 3 hours long and they need to have a minimum of comfort, so that the learning process happens easily with no distractions.
A Good infrastructure will motivate students to attend and continue to strive to learn.
3 - Communication
Communication is key! Communication should be fast and efficient, both between the organizers and the students( here is where they post questions and get help from organizers or classmates).
In our case we use several communication resources: We have a general deep learning group in the Telegram and we create groups in whatsapp / telegram / slack for each course.
We also use the Machine Learning Brasilia meetup to manage the meetings and communicate when will be the next meetings.
4 - Meetings
The meetings are planned in advance, where sections of the video from Jeremy’s class are selected. These sessions usually last 20-25 minutes (based on the Pomodoro technique), followed by 5-10 minutes to answer questions and/or share ideas. Students should be encouraged to pay attention to the video and take note of the doubts, as well as not try to execute the code during the videos.
We control the students presence through QrCode associated with google spreadsheets or, more recently, Kahoot, a gaming-based learning platform.
The three hours of the meeting are divided this way: 90 minutes of class + 10 minute break + 80 minutes of exercise.
Every two video meetings, we have a class that we call a practical laboratory, where we do a step-by-step coding exercise prepared by some volunteer of the course or by an organizer. This practice is highly recommended as it allows students to execute the codes and bring different datasets for application of the theory with their company’s data or study at the university.
5 - Projects
Experienced students organize themselves into groups to execute projects to solve real problems of local public or private companies.
These projects are presented in specific meetings as a way of exercise the practical application of knowledge and recognizing the participants’ efforts.
6 - Marketing
It is important to find talents in the group who have the ability to publicize the existence of the group and its achievements in social media networks, ensuring that the new courses have more and more participants.
The organizers are also expected to publicize the group at seminars and lecture at artificial intelligence events in the city. This is another important form of marketing study group.
I believe these are some important points that are key to success in organizing Fastai study groups. I hope these ideas can help colleagues around the world to create a competent community in the practice of deep learning. It is important to emphasize that much of our success with the community comes from the excellent material created by Jeremy Howard and the entire Fastai team.
I wish everyone a huge success!