Designing AI Solutions for Sustainable Development: Opportunities and Best Practices

Transcript

1. Ayo Kehinde Samuel Designing AI Solutions for Sustainable Development: Opportunities and Best Practices Senior Data Scientist, AI engineer & Software Engineer
2. Background Ayo Kehinde Samuel Key Certifications in Production ML: • Microsoft AI Engineer (Associate) • IBM Certified AI Engineering Professional • GAMES Consort AI Engineering Professional • Advanced ML with TensorFlow on GCP MOOCs:
3. Agenda • Overview of Artificial Intelligence & Machine Learning • Designing AI solutions • Deploying to production • Opportunities for AI in Sustainable Development Ayo Kehinde Samuel
4. • Overview of Artificial Intelligence & Machine Learning
5. 3 reasons to which it is difficult to define AI • Difficulty to clearly define human intelligence • The AI effect • Mix-up between the various stages of AI and types Artificial Intelligence Ayo Kehinde Samuel Simply put, AI is a branch of science studies dealing with the simulation of intelligent behaviour in computers.
6. Artificial Intelligence Ayo Kehinde Samuel AI Wave Domain Brain Power Artificial narrow Intelligence Domain specific, applies AI to specific areas Looks similar but not at the same level (Surpasses mouse brain power) Artificial general Intelligence(AGI) Cross Domain capability, Applies AI to several areas Surpasses human brain power Artificial super Intelligence (ASI) Cross Domain capability, applies AI to all areas Surpasses human brain power Narrow AI vs AGI vs ASI
7. Machine Learning Ayo Kehinde Samuel Machine learning is subset of artificial intelligence (AI) which gives devices the ability to learn from their experiences and improve their self without doing any coding. The goal of machine learning is to develop methods that can automatically detect patterns in data, and then to use the uncovered patterns to predict future data or other outcomes of interest.
8. Machine Learning Ayo Kehinde Samuel Supervised Learning: • Mapping a given set of inputs to a given set of (labeled) outputs (Linear regression, Classification trees, Neural networks) Unsupervised Learning: • Only labeled inputs but unlabeled outputs; Algorithm needs to infer the underlying structure from the data itself (black box) (Cluster analysis)
10. • Designing AI solutions
11. What does it entails? • Understanding the Business Problem • Data collection • Deep dive into Data understanding • Data cleaning • Exploratory Data Analysis • Feature Selection/Engineering/Augmentation • Model training • Model evaluation • Hyper parameter tuning Design Workflow Ayo Kehinde Samuel
12. Model vs Algorithm • Model is what you get when you run the Algorithm over your training data and what you use to make predictions on new data. • A Model is a Function which takes inputs and gives an output (prediction) • You can generate a new Model with the same Algorithm but with different data. • You can get a new Model from the same data but with a different Algorithm or different hyperparameter of same Algorithm • Model is unique to your project and deployed to make predictions. Model or Algorithm Ayo Kehinde Samuel
13. • ML algorithms for designing solutions
14. • ML algorithms for designing solutions
15. • Deploying To Production
16. Batch Processing: • Prediction is done before hand and scheduled Stream Processing: • Prediction is real-time Edge: • model is on the device Type of Deployment Ayo Kehinde Samuel
17. Microsoft set of tools
18. Azure: • Azure is a flexible platform for AI and ML Azure Machine Learning: A set of azure cloud services that enables you to: • Prepare data • Build Models • Train Models Azure & Azure Machine Learning Ayo Kehinde Samuel
19. Advanced Analytics Pattern In Azure
20. Start locally finish up in the cloud
21. • Github provides a code repository for version control • Github actions provides CI/CD capabilities GitHub & GitHub Actions Ayo Kehinde Samuel
22. • Opportunities for AI in Sustainable Development
23. Sustainable Development Goals
24. WATER: • Predictive maintenance of water infrastructure • Monitoring and predicting water demand • Monitoring waste water sources ENERGY: • Smart monitoring and management of energy consumption • Energy supply and demand prediction AI Sustainable Solutions Ayo Kehinde Samuel
25. Ayo Kehinde Samuel Data Analytics for Business: Crafting a Compelling Narrative from Your Data Insights Head of engineering/Analytics, Space Universe