Please enable JavaScript.
Coggle requires JavaScript to display documents.
Artificial Intelligence, Machine Learning Implementation - Coggle Diagram
Artificial Intelligence
Strong AI
Human like intelligence
We are far from this development
Weak AI
Trained for specific task
Ex.Price Prediction, Autonomous Car
Machine Learning
Input ->
Features
Output ->
Labels
Regression
Labels are numerics with specific range
Classification
Labels are defined set like Yes/No, 0/1, A/B/C
Supervised Learning
Self Supervised Learning
Gen AI
Input ->
Prompts
Output ->
New Content
Facts on Gen AI
Gen AI requires billion of data to train
500 Billion Data used to model GPT 3
Uses Open source community, wikipedia
New Content
Contents are generated by tokens not by words internally
Tokens are based on the type of word, part of speech
Probability is calculated for the words to generate in the sentence
Neural Network is used
Generative AI Journey
Machine Learning Implementation
ML Based API
NLP, Vision, Speech etc
cloud.google.com
Text to Speech API
Speech to Text API
Insights from unstructured text API
Filter user-genrated images
Content moderation for Video
Object Detection & Tracking
Custom Model without ML Expertise
Vertex AI > Auto ML
Custom Based
Image
Text - adding custom labels
Tables - classification, sentiment analysis
Video - Object detection & Tracking
Vertex AI
Foundation Model -> Model Garden
Large Language Model
API -> Vertex Palm API, Palm API
Tools to build apps using APIs(Generative AI Studio)
Mkaer Suite -> Palm API without ML Expertise
Build Complex Custom Models
Vertex AI > Custom Training
with the help of ML expertise