Most Popular Machine Learning Languages and Packages: GitHub
Like “Big Data” before it, “machine learning” has become one of those near-ubiquitous buzz-terms, thrown around not only by experts, but also marketing and PR people who want their company to sound cutting-edge. But for those tech professionals who actually want to break into machine learning as a profession, that immense buzz can make it difficult to discern practical knowledge. Where do you start? What languages should you learn first? Fortunately, GitHub has some good data about which languages and frameworks are dominating the machine-learning discussion. The code repository pulled data on contributions (i.e., pushing code, pull requests, comments, and reviews) to determine the top machine learning languages, which include (in descending order): And that’s not all, as the infomercials say. “We pulled data from the dependency graph to calculate the percentage of projects with machine learning or data science topics that import popular Python packages,” reads GitHub’s blog posting on the matter. In other words, these are the packages and libraries that tech pros use exceedingly often in a machine-learning context: Those interested in machine learning should definitely make a point to explore TensorFlow, which is a notably high-paying skill. Created by Google, TensorFlow is a framework for deep learning that’s existed since 2015, which means it’s a pretty mature platform by the standards of a relatively nascent industry. If you’re interested in learning how TensorFlow actually works, Google has a free, three-hour course (with video and text elements) available on its site. Numpy, an open-source Python library for scientific and numeric computing, has also been popular for a few years (it boasts a number of useful, built-in functions). Understanding Numpy (along with other libraries such as Pandas) is key to figuring out how to frame machine-learning problems. For those just starting out in machine learning, there’s a plethora of free materials online. In addition to Google’s TensorFlow courses, Amazon offers 30 self-paced courses, complete with videos, labs, and text-based lessons. Although those courses are free, Amazon charges tech pros to take the exam for a new AWS certification for machine learning (“AWS Certified Machine Learning—Specialty”). Microsoft’s Professional Program for Artificial Intelligence features 10 online courses that teach 10 skills. And if you want to go a nonprofit route, OpenAI (a foundation dedicated to creating an ethical framework for A.I. development), has lots of materials for more advanced A.I. students, such as tools for “training” artificial intelligence and machine learning platforms. Other companies regularly release their latest research about machine learning, although those materials are often quite advanced. For example, Bloomberg issues a regularly updated list of its researchers’ most recent ML-related publications. Once you’ve gotten your bearings in the machine learning community, make a point of regularly checking out what the experts are up to.