This repository has been archived by the owner on Aug 2, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 160
01. Roadmap for eShopOnContainersAI (Artificial Intelligence features)
Miguel Veloso edited this page Feb 19, 2019
·
6 revisions
-
Product recommendations based on Azure Machine Learning Studio model
-
Products search based on image recognition through multiple approaches:
- Azure Cognitive Services (Easy/basic Computer Vision approach)
-
TensorFlow + TensorFlowSharp (Advanced Deep Learning approach)
- Using a pre-trained model
- Using a custom model to be trained with Azure ML Workbench (Vienna)
-
CNTK (Advanced Deep Learning approach)
- Currently, the CNTK microservice needs to be deployed on Windows / Windows Container since CNTK still doesn't support .NET Core.
-
BOT as client app for eShopOnContainers
- Based on Microsoft Bot Framework
- Surfacing AI features like product search based on image recognition
- Language understanding (based on L.U.I.S.) to access eShop features like Order's status
- Internal Dashboard app with Orders Predictions based on a model trained with Azure ML Workbench (Vienna)
- Logging in to Identity Server
-
Recommendation systems
Product recommendation (ML Studio, C#) -
Computer Vision
Image classification-
Cognitive Services - Mobile applications (Xamarin, C#)
-
TensorFlow - Custom Model (ML Workbench, CNTK, TensorFlow, C#)
-
Natural Language Processing
Skype Bot (Microsoft Bot Framework, LUIS, C#) -
Regression
Sales Forecasting (ML.NET, C#)