microsoft / ML-For-Beginners

12 weeks, 26 lessons, 52 quizzes, classic Machine Learning for all
https://microsoft.github.io/ML-For-Beginners/
MIT License
68.27k stars 13.98k forks source link

[Content]: Real World Applications for Classic Machine Learning - Select an area #28

Closed jlooper closed 3 years ago

jlooper commented 3 years ago

To complete the lessons, we want to add an overview of ML as it is used in the real world. This will focus o classic ML only! So don't worry about Neural Networks for this curriculum. Pick a domain and write a paragraph about it as a reply to this issue, and I'll add it to the lesson and credit you!

AskingAlexander commented 3 years ago

Retail - Inventory management

You can think about Retail as the industry where you would like to offer the best experience to your customers regardless if they are buying the chocolate since it's their break, they are getting their weekly groceries, or remodeling their living room. With this in mind, Inventory Management will mean you will have to make sure that: A (very) large number of items (classes) have at least a decent number of each (samples/individuals) Those items are placed properly You won't spend eternity or a big number of people making sure this process is done properly Luckily, this is an ML lesson so we'll solve this using Computer Vision, more detailed Instance Segmentation, which means that we will identify "things" in a picture, label those "things" and count them. The process will become: The store clerk goes to the shelf Takes a picture of the shelf Our amazing app will analyze the image, get all the items, add them to the database and also show the clerk the results for possible corrections Done in just a walk around the store We can elevate this operation, of course, by putting cameras in all the places where we can check for inventory. If you want an out-of-the-box solution, Microsoft offers Custom Vision where we'll have just to label the data and they will help us with the model, cool, right?

AskingAlexander commented 3 years ago

Waiting for the evaluation of the solution so I can adjust the weights for the next ones :)

jlooper commented 3 years ago

hi, Let me clarify a bit! I'm looking for actual case studies that we can link to. Here's an example:

Energy sector: This article discusses in detail how clustering and time series forecasting help predict future energy use in Ireland, based off of smart metering: https://www-cdn.knime.com/sites/default/files/inline-images/knime_bigdata_energy_timeseries_whitepaper.pdf

So the task at hand is to find a real-world whitepaper, case study or use case that was written up, link to it, and write a paragraph about how CLASSIC ML (not neural networks like the custom vision topic above) helped solve a problem in the real world using Regression, Classification, Clustering, Time Series, or any of our other topics WITHOUT using NN.

jlooper commented 3 years ago

@ornelladotcom if you want to 'adopt' a chunk here let's post it here :)

ornelladotcom commented 3 years ago

Working on the Finance section now

jlooper commented 3 years ago

Here's a good one for health care: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7979218

AskingAlexander commented 3 years ago

For Retail: Inventory = https://www.pymnts.com/news/retail/2021/farmstead-brings-ai-inventory-doordash-delivery-to-grocers/ Personalization (Amazon OFC) = https://www.emarketer.com/content/do-retailers-have-firm-understanding-of-their-shoppers

What do you think?

jlooper commented 3 years ago

For Retail: Inventory = https://www.pymnts.com/news/retail/2021/farmstead-brings-ai-inventory-doordash-delivery-to-grocers/ hi, it's promising, especially the prediction aspect. Can you find a whitepaper that details exactly what strategies they used? (regression, clustering, time series...) Personalization (Amazon OFC) = https://www.emarketer.com/content/do-retailers-have-firm-understanding-of-their-shoppers On this one, I don't think we have enough detail. We really need whitepapers, if we can get them!

What do you think?

jlooper commented 3 years ago

Let's look at Wayfair https://www.aboutwayfair.com/tech-blog?q=&s=0&f0=00000178-3b97-d9a4-a37a-fbf7460c0001

jlooper commented 3 years ago

This is perfect: https://www.aboutwayfair.com/tech-innovation/how-we-use-machine-learning-and-natural-language-processing-to-empower-search