Deploying a Machine Learning Model without a single line of Code
On the 29th of January, Aoun Lutfi and Mahitab Hassan from the Developer Experience team at IBM hosted the ‘Deploying a Machine Learning Model without a single line of Code’ workshop at AstroLabs.
Aoun kick-started the workshop by defining machine learning. He clarified the difference between two frequently confused terms, machine learning, and artificial intelligence.
Artificial intelligence(AI) focuses on creating machines that are capable of carrying out asks in a way considered “smart”. It is human intelligence exhibited by machines.
Machine learning is a way of achieving AI, where the machine is trained to parse large amounts of data and use algorithms that give it the ability to learn how to accomplish/perform certain task.
He went over key terminology and illustrated the different types of machine learning algorithms. He went into detail to explain supervised, unsupervised, semi-supervised and reinforcement learning and gave examples and algorithms for each.
He further went on to explain the methodology behind building a machine learning model. How you start by obtaining data to clean and manipulate it, then you train the model, test the data and improve your model.
Aoun then gave an overview of Data Science Experience (DSX), a collaborative platform for data scientists and summarized its functions.
Next, he outlined the capabilities of Watson Machine Learning (WML), how it can be used to create, train and deploy self-learning models.
When building a machine learning model, traditionally, the developer would have to prepare the code by themselves and use a service, such as containers, to create the environment for doing the classification and training.
With WML, everything is taken care of, all the developer has to do is upload the code. Also, WML provides the computational power needed to do the training and classification, which could otherwise take a painfully long time with larger data sets.
Mahitab then took over to conduct the hands-on lab and went over the steps to building and deploying a machine learning model. She demonstrated how effortless it is to do so using DSX and WML and finally deployed it to the cloud, all in a matter of minutes!
If you missed this workshop, don’t worry! More are on their way, join our meetup group at https://www.meetup.com/IBMCloud-Dubai/ to stay updated.
Resources:
Slides: https://ibm.ent.box.com/v/MLinMinutes
Signup/Login to IBM Cloud: https://ibm.biz/BdZB3w
Hands-On Lab guide: https://developer.ibm.com/code/howtos/ml-in-minutes
Signup/Login to Data Science Experience(DSX): https://datascience.ibm.com/
