Google and Coursera have been constantly working together to launch a number of online courses for developers and IT pros over the last few years. Among those courses, one was the Machine Learning Crash course, which provides developers with an introduction to machine learning. Now, after teaming up and working together, they are launching a machine learning specialization on Coursera. It consists of total five courses and has more focus on practical work. The new specialization is known as “Machine Learning with TensorFlow on Google Cloud Platform”. It has students that have built real-world machine learning models. It goes from setting up the environment for their model, to learning how to create and sanitize datasets, to writing distributed models in TensorFlow. It also includes to improving the accuracy of those models and changing them accordingly to improve accuracy with the right parameters. According to Lak Lakshmanan, Google’s Big Data and Machine Learning Tech Lead, students and companies really enjoyed the original machine learning course. But, they also wanted to go deeper into the subject. Besides creating a basic model, Students also wanted to know that how to use those models in production in the cloud.
VP of Coursera’s Enterprise Development, Leag Belsky, considered this feedback from students and said that such specialization with practical models would have more impact. Nowadays, machine learning is playing a vital role in almost every area of tech. So, the target audience for this specialization is everyone who wants to develop skills in this field. It is very hard(nearly impossible) to hire machine learning experts. So, this course will surely be impressive for employers who want to develop and improve the skills of their workforce. According to Lakshmanan, there are many cases in which machine learning models can be implemented. But, these courses will focus more on the “day-to-day models” so that it could bring additional value to many existing products. He also noted that the course should be useful for fresh graduates who are familiar with the theories of machine learning and want to get practical in this field. The course focus more on the real-world problems, so it would have more impact for the beginners.
He also believes that only a few years ago, getting into a course like this was nearly impossible for most of the students as it requires powerful hardware with a dedicated GPU to work productively. But, thanks to cloud platforms that provide GPU access and even specialized hardware like Google Cloud’s TPUs. Cloud platforms have made it easier for everyone to get into a course like this. It is obvious that such courses requires you to be familiar with programming. It has become easier to start with machine learning because of frameworks like TensorFlow. It will take a while when we will come across a course like “get started with programming in Python by building a machine learning model”. According to Lakshmanan, the team is also looking ahead at a next course. The new course will build upon the existing course with a focus on working with unstructured data. It will allow the graduates of the first course to apply the knowledge gained from the course to a complete different set of data.