The culture and growth of scientific research in various domains are interesting to observe. For many decades traditional scientific research has been accessible or done in academia or research institutes. Each decade or century witnessed various sciences rising to the limelight. Once it was physical sciences, then it was biological science, the age of computers, etc.
In this early twenty-first century, the star in this is Artificial Intelligence (AI) and Machine Learning (ML). We can see phenomenal interests from academicians, people from the industry, and even enthusiasts in this field. According to my opinion, the key reasons for the interest are the general applicability of the field and open-access resources.
The Wide Applicability
The access to the high quantity of data and computing made it a potential solution for many areas where it is possible to collect data in a digital form. Nowadays, artificial intelligence and machine learning are applied to several fields including healthcare, education, automobile, art, entertainment, retail, finance, etc. This brought people from many fields to flock to machine learning and AI. In addition to that fresh perspectives and culmination of ideas, it gave birth to several sub-fields also.
The Open Research Culture
Regardless of all these the most positive thing in the field is the open research and development culture. Unlike many other fields where most of the publications are behind paywalls and are only accessible by people with money and academic institutions, machine learning research papers, almost all of them are freely available in preprint form from Arxiv.
Reproducible research is expected as a mandatory standard in machine learning. I often had the experience of asking authors from other fields for code that supports their paper and they were often reluctant to share. But in machine learning, almost all papers share their code to reproduce the results. The platform GitHub provides an excellent opportunity to do distributed collaborative coding of large projects across a large group of developers. It has the provision to mark beginner-level bug issues reported publicly. You can be the contributor of a popular machine learning library from Google via GitHub. Nothing is stopping you from doing that except your interest.
Open Access to Resources
There are numerous lecture videos on machine learning and its subtopics made freely available by top-notch institutions. Podcasts do an amazing job of bringing the latest news to our convenient time and/or place, be it in the commute or kitchen. Organizations like Coursera are providing high-quality free education (not only in machine learning) through MOOCs intending to democratize higher education. Top tech companies are competing against each other to make free access to computing resources and software tools. If you take the top three programming languages in the area all of them (Python, R, and Julia) are open source and sets commercial software way back. Most of the industry standard and learning machine learning software libraries are open source and pushed forwarded by volunteers across the world.
Open Research Groups
Among all these factors, the machine learning and AI community take a revolutionary next step by forming volunteer groups across the world. The platforms like discord, slack, mattermost are used as a meetup place and organize the collaboration. Many such groups organize weekly or even daily virtual events to discuss research papers, new tools, software libraries, etc.
The following is a tiny list of such open research groups which I am aware of.
The collaborative work resulting from such groups has been published in top conferences in the field. I agree with the fact that solving research problems is hard and it takes months or years of efforts to crack a problem. But despite these facts, such open research initiatives are growing slowly but steadily. I am sure that this is going to make remarkable outcomes in the coming years and more such groups will emerge.