How do you remove predisposition from the artificial intelligence models and ensure that the predictions are fair? What are the 3 stages in which the bias mitigation option can be applied? This code pattern responses these questions to assist you make notified choice by consuming the outcomes of predictive designs.
If you have concerns about this code pattern, ask them or try to find answers in the associated forum.
Fairness in information and artificial intelligence algorithms is vital to developing safe and responsible AI systems. While precision is one metric for evaluating the accuracy of a machine finding out model, fairness gives you a way to understand the useful implications of releasing the design in a real-world situation.
In this code pattern, you use a diabetes data set to anticipate whether a person is susceptible to have diabetes. Youll utilize IBM Watson ® Studio, IBM Cloud Object Storage, and the AI Fairness 360 Toolkit to create the information, apply the predisposition mitigation algorithm, then evaluate the outcomes.
After completing this code pattern, you understand how to:
Visit to IBM Watson Studio powered by Spark, start IBM Cloud Object Storage, and produce a job.
Publish the.csv data file to IBM Cloud Object Storage.
Load the data file in the Watson Studio notebook.
Set Up the AI Fairness 360 Toolkit in the Watson Studio notebook.
Evaluate the results after applying the predisposition mitigation algorithm throughout pre-processing, in-processing, and post-processing phases.
Produce a project utilizing Watson Studio
Use the AI Fairness 360 Toolkit
Find the detailed actions for this pattern in the readme file. The actions will reveal you how to:
Create an account with IBM Cloud.
Produce a new Watson Studio task.
Produce the notebook.
Place the data as DataFrame.
Run the notebook.
Examine the results.
This code pattern belongs to the The AI 360 Toolkit: AI models described use case series, which assists developers and stakeholders to comprehend the AI design lifecycle entirely and to help them make informed decisions.