From seed to store: AI to optimize the indoor farms of the future
summary
Bowery runs warehouse-scale indoor vertical farms using hydroponics and LED lighting, claiming 100x more productive land use and 95% water efficient recirculating systems while using zero pesticides. Farms instrument every input (temperature, CO2, light spectrum, nutrients) and collect imagery with custom cameras; computer vision + deep learning produces growth curves and defect detection. Machine learning treats recipe tuning as a multi-armed bandit over trillions of permutations, runs ABC... on 100,000 crop cycles/year per farm, and schedules harvesting and packing to minimize loss across 200+ cultivars.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 0:28About Bowery
- 1:31Modernizing farming
- 1:59Challenges and opportunities
- 2:33Bowerys solution
- 3:18LED technology
- 3:53Key technological trends
- 4:21Looking ahead
- 4:39Farmer Tom
- 5:27The 1900s
- 6:24Bowery OS
- 8:02What is an operating system
- 9:53Learning faster
- 11:54Multiarmed bandit problem
- 13:29Tom Colicchio
- 14:09The Perfect Recipe
- 16:15Deep Learning
- 19:31Optimization Problems
- 24:37Thank you
related talks
19mPlant Factory: Sensor, Data, Machine Learning
07mAI Infrastructure Transforming Modern Agriculture
26mAI, big data, digital and future of agriculture
10mUsing Machine Learning to Reduce World Hunger
22mAgri-Tech Talk
27mScalable Data Flow Management for Smart Farming Digital Twins
15mAI Gardening: Harvesting and Pruning the AI Tree
20mRobots will be running the farm