Exponential Thinking: Artificial Intelligence & Machine Learning – Part 6


Welcome Back to Exponential Thinking: Part 6

1. Introduction, 2. Renewables, 3. Smart Ag, 4. Online Learning, 5. Vertical Farming

In the last few months, we’ve covered renewable energy, smart agriculture, online learning, and vertical farming. This week we’ll be talking about something even more revolutionary and mind-bending.

Continuing Our Series on Exponential Technologies.  

If you’re reading this, maybe you’re like me and follow technology news. If so, you’ve seen a rising trend recently, especially in the last few years. This is beyond the scope of what we’ve discussed in the past few articles. In fact, this might seem like an article better fit for Science Fiction Quarterly or a similar publication. But I assure you, as certainly as fish poop grows great plants, the future of food will depend on this emerging technology.  

So what is this trending technology changing the way we interact with our world? I’m talking about artificial intelligence or A.I.

If you’ve used any of the most popular websites in the world, including Facebook and Google, or have used an Apple or Microsoft product, you have used artificial intelligence. Many of us rely on artificial intelligence multiple times every day, whether or not we know it. Recently, Google revealed that an artificial intelligence nicknamed “Rankbrain” was the 3rd most important factor in the Google Search algorithm that controls the ubiquitous search function from google.com. So the next time you look up, “how do I compost fish solids,” “what are the different steps of the nitrogen cycle,” or some other aquaponics question, you’ll get your answer from a real-life thinking machine.

First, I’ll describe three prominent trends in artificial intelligence: Machine Learning, Deep Thinking, and the Digital Assistant. Then we’ll explore how these and other artificial intelligences will change how we grow food, and sooner than you might think.   

Machine Learning and Artificial Intelligence

So what is Artificial Intelligence, exactly? AI comes in many forms, but at its most fundamental level, it’s a computer that is capable of learning from its environment and past experiences to create new programming for itself. This is akin to a human child learning as she grows up from a child to an intelligent woman, and why we call it artificial Intelligence. Think of it like learning your times tables. When you learn 3×3 equals nine and 4×3 equals twelve you can infer that 6 x 3 will be eighteen. AI works like that. Once it learns the rule is to count by three it will know the value for any amount of three’s you want to know. 

One form of AI that has been generating major buzz lately is known as Machine Learning. Machine learning is when an AI is capable of analyzing and recognizing patterns and then making predictions based on that data. Google’s Rankbrain represents machine learning with a predictive search function. You know how when you’re typing something into the search bar, it finishes the word or phrase for you? That’s Rankbrain’s way of helping you out!

Google Server Pipes
Google Server Pipes: an inside look at Rankbrain’s…brain.

 

Rankbrain and other AI like it are remarkably similar to microbes. Just as the tiny Nitrobacter tirelessly devours nitrite and produces nitrate in a joyous cycle of self-fulfillment, Rankbrain’s day is happily spent serving over 3.3 billion searches per day (that’s billion with a B).  

Deep Learning, or Why The First AI to See Youtube Watched Cat Videos

Deep Learning about Cats
When AI stares into the kitty, the kitty stares back

A subset of Machine Learning AI is known as Deep Learning. Deep learning algorithms allow computers to analyze images and videos and come to conclusions and describe what they see. When Google unleashed their deep learning AI on Youtube, it spent 3 days analyzing 10 million randomly selected videos. The result? This adorable AI ended up watching and cataloging over 80,000 cat videos!

In addition to Rankbrain, Google recently announced another machine learning based AI.  A new feature for their two email apps Gmail and Inbox now serves the user automatically generated responses to emails based on your own language. This new feature uses your previous correspondence to craft replies that actually are something you would say! I think of it as the next evolution to autocorrect during typing. The beauty of this new AI feature is that it can analyze contextually, meaning it can understand what is happening between the people in an email exchange. It knows when you buy a plane ticket that it should track your flight info for you. It knows when someone asks you and your significant other to come to a party, it will prep yes, no, and maybe responses and it will make them plural because you and your significant other were invited together.

Ever wonder how Facebook recognizes your face and your friends in photos? Well, the answer lies in Deep Learning. In the last two weeks, Facebook announced their in-house AI is able to observe images and describe them in detail. For now, it can only observe objects and living things in the photos, but in time it will also recognize contextual information like what action is happening in a given photo. So not only will it recognize you and your dad but it can differentiate and say you guys are enjoying a drink at the bar in one photo and in the yard in another photo. 

The Rise of the Digital Assistant

The Digital Assistant, and I don’t mean a personal digital assistant, those clunky bricks from the 90s, this is a new form of application available from many major technology companies. They offer a very wide range of services but generally are all about making your life easier. It all started with searching the Internet using voice activation commands. Then eventually you could set alarms, reminders, and even text people all by talking to your digital assistant on your phone. Now it is quickly evolving into something new altogether.

Siri came first, entering iPhones in 2010. She was the first real entrant to the digital assistant market but initially was more fun than useful. You could ask your phone weird and crude things and get funny canned responses from Apple. You could also set alarms, search the web, and send texts all by voice. The new Siri is cool and allows you to search your entire phone, app store, and the web all simultaneously anytime you have a question. Siri keeps evolving as it learns from millions of users. It gets more accurate every day. 

Digital Assistant CortanaMicrosoft’s entry is named Cortana. She’s the first Digital Assistance to go through real Cultural Etiquette training. Cortana is trained in a country’s local etiquette and mannerisms in order to fit more in. Ask Cortana a question in foggy London and she will respond with a droll, self-deprecating joke. Ask her the same question in Tokyo and she’ll respond with some surprisingly witty word-play.  

Despite Apple’s early entrance and Cortana’s cultural sensitivity, Google is the true frontrunner in artificial intelligence and machine learning. Google Now is the name of their Digital Assistant. Google Now is unique in that it is not anthropomorphized with a human name like Siri or Cortana. Google Now specializes in cultivating your perfect digital environment by serving you exactly the stuff you need to know, when you need tGoogle Now AIo know it. Have a plane ride later today? Google Now lets you know when you should leave given current traffic conditions and can even call an Uber to come pick you up just in time. It will also serve you cultivated news articles from across the web in any topics that you show interest in via a thumbs up/down method combined with your search history.

Facebook is another innovator in the digital assistant space. Beyond their deep learning AI, Facebook has a unique take on how they think their 1.3 Billion monthly users will interact with artificial intelligence. A small rollout occurred quietly in San Francisco Bay Area in the last few months. Known as M, this AI/human-operator hybrid is built right into Facebook’s popular messaging app called… Messenger. M is unique in that the AI shares the workload with a team of human operators who help it along when it is stuck on something. If you ask M to book a restaurant reservation for you, M can relay this request to a human operator who can get your task done for you without you having to worry about a thing. While Google Now is about serving you answers instantaneously (Google NOW), M can take awhile to answer your questions, up to 5 to 15 minutes. This is understandable because it is capable of handling a much wider array of tasks, including tasks requiring multiple steps. You can ask M to make a reservation at a Thai restaurant near my house and M will search the area for Thai food, match it to your budget and compare the results to the ratings for those restaurants on the web. It will come to a decision for you and then a human operator will place the reservation for you. Finally, M will relay back to you that you have a reservation at that cool new restaurant no one knows about yet! M will even order you an Uber to get you there just in time for your reservation.

Robot Butlers and the Farmer of the Future

By now, you’re probably saying ‘so what?’ You’re asking yourself, “what does some robot restaurant reservation service have to do with changing the way we grow our food?” I believe the future of food is strongly tied to AI.

As aquaponics entusiasts, we understand the future of food will be tied to local production. In order for that to happen a huge number of people will be growing food who have never done so before. They will also be growing food in places its never been grown before. They will need the know-how to grow productively in their unique environments. Previously, we’ve discussed the importance of online learning and how it is already changing the aquaponics industry.  What if all of the knowledge contained in the online courses, databases, and other information stores was collected and synthesized for you? What if it was only ever a question away? What if it actually learned about new techniques and could suggest them to you and show you how to implement them? This is the power of AI to the farms of the future.

With the world’s best growing information at your fingertips, anyone can grow like a fourth generation farmer.  All the most cutting edge planting tips, harvesting tips, and everything else will help you grow to your fullest potential! Thi is an incredible barrier currently to farming that should be completely removed with the advancement of AI. 

Machine learning AI will evaluate the data from the sensors embedded throughout your aquaponics system and serve you up with alerts, best practices, and optimization strategies based on all of the data collected across the shared food network.  As more farms and gardens join the connected network, the gardener AI will learn more about growing conditions in a wide variety of settings and scenarios. Over time this will allow new farmers to start off on the right foot with guided preferences for their given situation. Use your mobile to scan a given area and you can receive a customized aquaponics system perfectly optimized to your location.

Smart Sensor AquaponicsAim your phone’s camera at a plant and deep learning AI will recognize the plant and cross evaluate it with context, telling you if it is healthy, sick, nutrient deficient, or ready to harvest.  The AI can report back to you on exactly how to proceed if you are ever stuck in the garden and don’t know what to do.

Avoid disaster with live push notifications to your mobile in case of imminent flooding, power outage, ammonia spikes, or other issues.  Your smart digital farm assistant will even enable controls automatically when it learns your preferences, saving you from potential emergency failures.

We may be years away from this level of AI being commonplace technology on farms, but like all the technologies we’ve discussed in this series, AI is progressing to this point and it’s happening a lot quicker than anticipated.

 

Dangers of Artificial Intelligence?

As with any new technology, there are dangers involved. With something as game-changing as artificial intelligence, these dangers are magnified to the extreme. A whole genre of media, including the Matrix and the Terminator series, looking at the post-apocalyptic world of an AI-gone-bad. Even technology wunderkind Elon Musk is fearful of AI, cautioning for its strict control and oversight.

Would we be foolish to entrust such valuable information to artificial intelligence programs like our eating habits, food production numbers, or distribution networks? Do the benefits of enabling a whole new wave of farmers outweigh the potential downsides? These are questions that are more real than ever in light of the recent announcements from all the tech firms listed above.

I feel these dangers are warranted, especially considering the vast amounts of data we hand over to the various AI populating the web. However, I do not think that fear and apprehension should hold us back from exploring the power of this amazing exponential technology. Assuming AI is evil speaks more about the person doing the assuming than about the AI itself. Anything that can help someone with no growing experience grows sustainable food for their community is worthwhile in my opinion.

What Does this Mean For Us?

Above all, I think that most important part an AI will play in the food system is connecting growers to buyers directly. Drawing from its massive data stores, an AI could create a map of food demand in a given area. Then it could plot the food production in the same area. Then using machine learning algorithms, a system could be created that would match local food demand to local food creation as best as possible. Essentially, an AI system could match food production to local production and generate an optimal schedule and delivery system for the food. This same model could be used to help producers as well. New farmers could look in their community and see what was in demand and meet that demand when they start growing. It would reduce waste but also the uncertainty that can go into farming and not knowing what you should grow.

Given the variety of applications for aquaponics and the sheer productivity, the future of food will be a network of small farmers making a living from their basements, rooftops, and backyards. This complex network will be managed by AI in order to conduct food efficiently even in super-densely populated urban areas. Food aggregation can become automated with the advancement of self-driving electric cars, and a car can come to the house or the farm, collect the harvest, and bring it to the processor for distribution to local residents. 

As we learn about new technologies we always look to see how they can integrate into a model that helps us to create the City that Feeds ItselfTM. It is a lofty endeavor made much easier when technology breaks down so many of the barriers to achievement. The advancement of A.I. will be very integral to empowering hidden farmers in cities across the world with the knowledge they need to be successful. Feeding an urbanized population of 9 billion people by 2050 will definitely be a challenge. But the advancement of A.I. and it’s ability to help us optimize and learn new solutions to existing problems give us hope and confidence that we can make the transition to sustainably fed cities.

We live in the most exciting time the human race has ever seen. The gap between science fiction and science fact gets shorter every day. As aquapioneers, you and I are actively taking part in this change every day. Our aquaponic gardens represent the cutting-edge of agriculture technology. When aquaponics collides with exponential technologies like artificial intelligence, there is no limit to the abundance we can create together.

 

[this essay was dictated to but not read by a digital assistant]


Leave a Reply