Pre-requisite Knowledge
Before we start with the understanding of what is Azure Databricks, we should have:
Background
I would like to give some short information about ‘what is artificial intelligence and machine learning’ before jumping in to Azure machine learning.
Artificial Intelligence
- In simple words ‘Artificial Intelligence (AI)’ is the artificial creation of the system like a human who can observe, react, learn, plan and process instructions and provide intelligence on it.
- It is a rapidly emerging technology and internet enabled technology.
- Sometimes AI is also called Machine Learning.
Introduction
Machine Learning
- Machine learning is not new. It is subset of Artificial Intelligence (AI).
- An algorithm is sequence of activities/actions/steps used to solve a problem.
- Implementing the algorithm and its models is called machine learning in computer world.
- Today, developing a new algorithm to instruct the computer to run it is the cornerstone of the advanced technology.
- Important
- Machine learning is a data science technique that allows computers to use existing data to forecast future behaviors, outcomes, and trends.
- Machine learning works on the mathematical model and is built by using the sample data.
- Machine Learning has the capability to learn and IMPROVE from experience WITHOUT being explicit programmed.
- Examples
- How an email system tracks the spam email.
- How an online shopping system shows the similar product which you are looking for.
- Types
- Supervised Learning – We have trained the model by data sets.
- Unsupervised Learning – Machine learning model learns the data and finds the patterns and relationships in the data. Based on the pattern and relationships the model is trained.
- Reinforcement Learning – Machine learning model will find out the best outcome. It works on a trial and error method. Once the model is trained then it's ready for predicting the new data.
How it works?
- At a high level, a machine algorithm creates one model data based on the existing test data as input.
- Pushes the new input data then the machine learning algorithm makes a prediction based on the model which was prepared in step 1 above.
- This prediction is evaluated and if accepted then an algorithm is deployed.
- If the prediction is not accepted, then machine learning is trained again with bigger training data.
- Azure Machine Learning Service,
- Microsoft Azure provides the cloud-based platform to the machine learning implementation and deployment.
- Using Microsoft Azure ML feature, we can prepare data, train the model, test the model, deploy the model, manage and track the model.
- We can scale out ML to the cloud using Azure ML.
- Azure ML supports the open source technologies like PyTorch, TensorFlow, and scikit-learn.
- This technology can be used in any ML type mentioned above.
- Use the Azure Machine Learning Python SDK with open-source Python packages or use the visual interface.
- It has a visual interface for experimenting and deployment with drag-n-drop.
- Microsoft Azure has Azure Machine Learning Studio to implement, test, train and deploy the ML. Machine Learning Studio is a collaborative place for data science, predictive analytics, cloud resources, and your data meet.
Image Source – Microsoft Docs
- We can implement the ML algorithm and model using tools,
- Visual Interface (Drag-n-drop modules)
- Jupyter notebooks (We can use SDK)
- Visual Studio Code Extension
Image Source – Microsoft Docs
Reference Links
- https://en.wikipedia.org/wiki/Machine_learning
- https://www.internetsociety.org/resources/doc/2017/artificial-intelligence-and-machine-learning-policy-paper/?gclid=EAIaIQobChMIn56dqfXA4wIVENbACh0njwYVEAAYAyAAEgJYRfD_BwE
- https://www.expertsystem.com/machine-learning-definition/
- https://www.edureka.co/blog/what-is-machine-learning/
- https://docs.microsoft.com/en-us/azure/machine-learning/service/overview-what-is-azure-ml
Conclusion
In this article we have learned the overview of Artificial Intelligence, Machine Learning and Azure Machine Learning Service.