michaels72 / ai_enhanced_911

AI Enhanced 911

The goal of this project is to make use of ML to make it easier to call for help in an emergency:

Major features

General datasets:

• https://gengo.ai/datasets/the-best-25-datasets-for-natural-language-processing/

• http://deeplearning.net/datasets/

• audio processing

  • a set of audio processing tools
  • Think Digital Signal Processing
  • blog post intro
  • A collection of Audio Files

• natural language processing

  • an introduction to nlp
  • a list of libraries
  • raw text
  • nlp datasets

• web interface

  • an introduction to flask

• geo positioning

• mutli channel call for help

• facial recognition

  • a great set of tutorials on image processing
  • opencv docs
  • image dataset

Engineering Best Practice

• git tutorials
• coding best practices
• testing
• CI
• CD
• containers
• file storage
• documentation

Understanding the phases of a machine learning engineering project

• Defining the system
• Designing the functionality
• Researching the current AI solutions in the space
• Learning (or relearning) the necessary technology
• Classical Statistics Goes Here
  • Exploratory Data Analysis
  • Hypothesis validation / rejection
  • Modeling and feature engineering
  • Model visualization and performance tuning
• Classical Software Engineering Goes Here With Support For the ML
  • Building connections from your model to your application, typically via HTTP or some other protocol
  • building out the rest of the system to support the machine learning portion
  • establishing the infrastructure to ensure latency requirements for the model are met
  • establishing model dashboards to ensure the models results are clear and make sense
  • establishing what qualifies as model drift, and therefore necessitates retraining of the model, or selecting a new model