April 21, 2021- Artificial intelligence aka machine learning is embedded in almost every electronic device we interact with and is helping increase efficiencies and efficacies within everything we do. Why? Because computers can process data much faster and with far greater accuracy than our ape based minds.
What is Machine Learning?
In my personal opinion Artificial Intelligence (AI) gets a bad wrap from the naysayers. And I think it's almost entirely derived from ignorance, I state that with the most innocence as possible. I mean, if you really pay attention today to how machine learning is assisting us,,,,, it's everywhere, and it's improving our lives and efficiencies.
First off, computers are in no way going to go all terminator on us and they are programmed by humans to execute on an event by doing A or B. All machine learning really is a choice based on conditions and over time the computer starts to see patterns and executes upon those patterns. Take cars that drive themselves, they have been shown, signs, road markings, visible objects, etc and have been programmed to execute actions or non actions based on those objects. It's the job of the computer to put it together in real time and at a rate that is far greater than a human's minds ability to process. It's certainly not perfect, but it's proven to be far more reliable than humans.
How Machine Learning is Helping
Machine learning is quickly becoming integrated into just about everything. Just the other day I found a smart bird feeder, which will ID bird types and notify you of a new bird species at your feeder via a mobile app, where you can view the bird remotely. Crazy right!?........ I find all of these devices as amazing. Think about it, now bird ID's in everyone's backyards can help scientists keep tabs on bird types around the world and their migration patterns. Powerful and fun!
Machine learning is all about patterns within data points and with more and more data, the device has the potential of becoming much more accurate. And if done correctly can actually become more proficient than an expert or master beekeeper. A smart beehive will contain sensors such as;
Interior Temperature
Interior Humidity
Exterior Temperature
Exterior Humidity
Hive Weight
Hive Microphone
CO2 Sensor
Accelerometer
VOC Sensor
GPS
Currently a simple sound recording device alone can detect disease, queen presence, swarming tendency, among others all off of a 120 second recording at any moment at the entrance of a hive. Now consider adding all the other signals the bees give off such as pheromones, hive temp, CO2 levels, etc and start data mapping them based upon known bee tendencies. I would challenge that not a single master beekeeper could take that data with a lifetime of knowledge and still even come close to what 40,o00 connected hives with minute based data points interconnected, matched with observations and billions of data points, done with near 100% data accuracy while compared with bee tendencies based on geographical locations and species.
Currently connected hives like the Hyper Hyve have the ability to take onsite data and compare that data to other historical hives data and compare it with known followed actions of the bees or hive. Such as, is there high levels of varroa, did the hive swarm, is the hive being attacked by other bees or predators, etc. The list is almost endless with the correct data points and proper machine learning algorithms.
Where is AI in Beekeeping Going?
The future of AI or machine learning in the beekeeping industry is simply limited on our own imagination and limitations on sensor data. The next big leap will be with the development of a chip that can detect the pheromones within the hive. Bees use pheromones to signal so many things on top of sound, temp, CO2, and others which we have no problem detecting now.
Besides hardware development challenges, the largest leaps will be in the code integrated into the machine learning programs and how they can effectively predict hive actions with high levels of efficacy.