Machine Learning for more than Sales and Fraud Detection-- which just sounds so excruciatingly tedious I have to forge my own path!
Recipes
|
ML task |
Example |
|---|---|
|
Classification |
Is this recipe runtime-ready? |
|
Multi-label classification |
Is it beans, greens, grains, soup, salmon, salad? |
|
Feature extraction |
Ingredient count, cooking method, pantry overlap, effort words |
|
Embeddings |
“chickpea bowl” and “white bean rice thing” are semantically close |
|
Recommendation |
“You have white beans, bok choy, rice, and broth. Here are likely matches.” |
|
Clustering |
Find natural recipe families without predefining categories |
Plants
|
Observation |
Possible features |
|---|---|
|
Plant type |
basil, lavender, poppy, pepper |
|
Environment |
indoor/outdoor, shelf, window, grow light |
|
Soil |
Promix, clay pot, drainage |
|
Watering |
last watered, water need |
|
Growth status |
sprouted, leggy, stalled, thriving |
|
Photo evidence |
leaf color, posture, density |
|
Weather |
humidity, temperature, cloudy/rainy |