I expect that one thing that will come out of deep learning is mechanical weeding for industrial level agriculture. Instead of spraying crops with herbicide, mechanically remove weeds. Bonus is agricultural machinery tends to be high capital cost anyways, which allows for a lot more processing power and sensors than a garden robot.
I really hope we get there soon. Not only could this eliminate usage of herbicides, but many smaller autonomous robots vs one really big John Deere could also remove the need for monocultures. And that would really be a major improvement for both crop yields and the environment
Except although the movie does say battery. That's not the original plot device that was written. They were originally written as human based cpu computers. But some higher up thought people wouldn't understand "human processors" so just everytime you think about or watch the matrix again. Just think it's a poor mistranslation that made it into the movie, "processor->battery"
I'm paraphrasing, but Neo says "that's impossible (in reference to the Matrix) since it violates the laws of thermodynamics"
To which Morpheus replies "and where did you learn about the laws of thermodynamics?"
If you're in the Matrix, your knowledge of the Matrix would come from the Matrix itself which would have a bias towards falsifying information to preserve itself.
I remember reading Indians of the Northeast planted corn, beans, and melons together. So switching to polycuture crops should have advantages.
One other thing I've read is that annual plants were the easiest plants to domesticate because the short life cycle results in faster evolution. However perennials are likely way less environmentally destructive.
> What will all these people do once farmers are unnecessary?
There aren't that many people working on farms in developed countries. In the US, only around 3 million people (or <1% of the total population) work on a farm. [0]
In other countries where many people are involved in agriculture, the pending industrialisation/automation of farming will probably have a similar impact to what happened in North America and Europe over the 20th century: more people will move to urban areas and add value to the economy in other ways.
>What will all these people do once farmers are unnecessary?
The answer is "whatever they want". The big problem facing governments everywhere right now is how to jump over the little gully of poverty that comes right after the change to get there.
corn is a tall crop, soy beans are short. Farmers are already experimenting with things like 4 rows of corn 4 soybeans. The corn rows shade the soybean rows and so reduce yield - but in turn the soybeans are not shading the corn so they get slightly better corn yields. This is still experimental, but so far it looks promising.
The best part? Weeds will still evolve resistance — just like they do for herbicides — but they'll do it by growing into shapes that the ML models misclassify.
Updating the robots can be done much faster than it takes for the weeds to mutate. And for mutations to happen, some proportion of the weed population must survive in the environment, as no weeds can not mutate. I doubt they have a chance, unless some species figure out how to live under the earth, in which case, we won't care as those would not be "weeds" any more.
1) Getting very close to the crops and intertwining with them to make weeding very hard
2) Looking almost exactly like real crops (as-in, to the point where most humans can't tell the difference and the machines need to start taking genetic samples)
You're totally right though; the machines have an adaption-speed advantage, so I'd expect this to be a very effective technique overall (assuming that the robotics pan-out).
Until the weeds become actually identical to the actual crops, and we just have weirdly recreated the crops. This is hilarious, and actually precisely how Generative Adversarial Networks work (https://www.wikiwand.com/en/Generative_adversarial_networks), although obviously with gradient descent in the generator network, not biological evolution
Give it a few decades and the weeds will find a way of corroding robots. In a century, we'll have to develop robots that can handle the acid that plants will be spraying on them.
Yep. John Deere will collect all of the recognition data and not allow their customers to see anything except high-level statistics (for an additional fee). Then they'll use it to train next year's model and charge for the upgrade.
Yet dandelions have genetically modified for lawn mowers, growing strangely mutated so as to be flat on the ground, yet the mowers haven't changed. So, sometimes the technology is not as fast as we think.
My source is simple observation. I did all the mowing as a kid, acres every week. In the spring, chopping down all the new dandelions was satisfying and the grass looked great. By the middle of the summer, dandelions were back but now they were strangely 'smashed' ones that grew no taller than 4", with flattened stems that turned or curled or flowers on short, short stems that bent over.
I'm sure that gene was always there; but billions of mowings over a century of lawn care means its present everywhere. Simple selection?
Ahh, no this is similar to the process that causes Bonsai to work. Dandelions have a fairly massive root with plenty of reserves to regrow after something eat's the top. Mowing simply causes similar responses as a goat would.
Plants can't move around so they evolved to deal with a really wide range of environments and stuff eating them etc.
Nah. Dandelions don't regrow stems; this can't be something done in response to being mowed. The plant doesn't think "I got mowed! I guess I'll grow a different way now".
Its a different growth pattern in the plant. Gotta be driven by genetics.
Dandelions are actually perennials so they will regrow stems several times, and survive multiple years. They will also tend to flower most abundantly in spring, but can re-flower in the fall.
Further, it's not just the stem involved, if your mowing regularly your taking out the leaves.
Yup the flat ones were none of those. They were really weird knarly mutated plants with exactly the same structure, leaves, stem and flowers as dandelion but grown flat and warped to hug the ground.
No! And it needs to be done! I've lived in only a couple of places, and the genetic variety of dandelion (dent-de-lion?) is marvelous. When I grew up they were small modestly sized plants. Then the mowing thing, and a 2nd form of flat plants. Moved to California, they were 4' tall with greyish-green dusty leaves. Moved to an acreage in Iowa, all over the fallow field one year they were bushes about 18" high with 100 blooms - you could have sold them in a garden shop for landscaping!
Mendel should have used dandelions for his gene studies, there is lots more to see!
Meh, there are plenty more cracks in sidewalks, open fields, and poorly kept yards than there are mowed lawns. Dandelion evolution probably isn't optimizing for mowed lawns any time soon.
I find it difficult to believe that any optically-driven machine-learning weeding robot will be 100% successful at eliminating every weed in a garden. There will be some overlooked and they will mutate.
Are you forgetting that you're talking about two different processes that operate on orders-of-magnitude different durations?
By the time any evolutionary affects are seen, if at all, of weeds being pulled by ML, ML will have progressed far beyond where it is now, likely well beyond what could be fooled by simple diversity.
DL models can learn to classify hundreds of weed classes in less than a week. It takes centuries for a new breed/trait to appear. This is a very uneven race.
Another use case is invasive species control in parks and large public lands. As an example, Portland has a large park that stretches along a ridge-line for about ten miles, and it's getting overgrown by English ivy, which kills the trees. It's a big enough area that pulling it all by hand would be very expensive. Ivy-pulling robots would be pretty awesome, but it's a hard problem. Vegetation is thick and the terrain is everything but flat. You'd probably need some giant hexapods or something just to move around, and maybe some kind of Ivy-grabbing appendage.
Just mapping out the ivy infestation with drones and image recognition software would be impressive. There's enough room for a quad-rotor to move around in theory, but it would have to have some pretty sophisticated obstacle avoidance algorithms.
Forest park is about ten square miles, and maybe a third is overrun by ivy. That's solvable by a sufficiently large human work crew, but it wouldn't be easy.
Developing, manufacturing, and operating a crew of robots just for this one particular ivy problem wouldn't be worth the cost, but invasive species are a widespread problem and are sometimes on much larger scales, like hundreds or thousands of square miles. At some point, using people isn't a realistic option. If we can solve larger invasive species problems for less money than a human work crew, then that seems like a win.
I envision a robot gardener that would create a paradise from a plot of empty land. Each plant would be placed, cared for and trimmed in such a way as to obtain an artistic effect.
Another kind of bot I am hoping to see is a waste collector. They would find plastic bottles, bags, glass and iron in forests, on roadsides, and even underwater, around beaches. Unleash 100 bots in a forest and in one year, it's pristine. We can also have some collect floating plastic from the oceans.
A trash bot would really be something to see. I'd pay a lot in taxes for that. Add in some ML matrix-stirring to watch out for fauna, and you have taken out ~70% of forest rangers jobs. Heck, just make a Park Ranger bot, cleans trash, scolds tourists, takes entry fees, fights fires; it'd be pretty nice.
That depends. Farmers used to mechanically remove weeds with a cultivator. (a fancy hoe attached to a tractor) Those cultivators will remove all weeds and no weed grows resistant to them.
However they have many disadvantages:
Cost - $500 worth of fuel for the tractor vs $250 for chemical + $50 for fuel to run the sprayer is the big one (These cost numbers are what I've heard but I might remember wrong).
CO2: that $500 worth of fuel vs $50 directly translates into CO2, which in turns translates into global warming. (this is the only one that you can argue doesn't have a financial impact on the farmer - and it isn't clear that those arguments are valid)
Soil disturbance. Every time you til the soil you disturb the natural soil. It isn't as bad as deep plowing, but it is still harmful.
Soil compaction. A repetitively light sprayer can cover 100+ feet of crop in just two tire tracks (the first wheel the runs over the ground does the damage, the second barely adds anything so the 4 wheels result in two tracks). A tractor is heavier (the impact of weight it not linear though so figure 10% worse), and can only do 30 feet so you end up with just over 3x as many tire tracks compacting and harming the ground.
Labor cost. Sprayers can run at 15mph in the field. Cultivators are probably running at 5mph. Not only are you taking 3x as many passes, each pass takes 3x as long. You pay the operator by the hour.
Now for your robot which applies? The weed pulling robot will only disturb the soil where the weed is. However the robot needs to get to the weed which increases compaction along each row.
My guess is farmers will spray everything as a primary prevention. Then they will fly drones over the field and send a robot down only the rows where there are weeds. That is just a guess though, you can imagine your own parameters for future robots which will change everything.
Note, herbicides are not nearly as dangerous as popular press makes them out to be.
When farmers plough they tow the unit behind the tractor, and this removes the compaction of their tyres.
You have to plough each year before sowing anyway so I'm not sure it's a simple as spray v plough.
You must not be familiar with modern farming. Most farmers do no till these days. They never plow - ever. The first seven years of no till they do see less yields than their counterparts that plow, but after that the advantage of non-distributed soil mean they get better yields. (more profit ever before we account for the fact that plowing uses a lot of expensive fuel)
Nothing is simple, but there is a reason most farms and all the big ones are no or limited till.
No. Large size monoculture should rely on preventive methods to control weeds, such as always covering the soil. It's even possible to plant seeds on an existing live cover.
No plowing, no pesticides, just preventive coverage and good tools.
In the long run, those large monocultures don't make sense anyway.
Robots can be great for a lot of high maintenance, Human intensive, high value cultures such as fruits, greens ...
I guess they could also help expand vegetables farms to non cultivated areas, in some autonomous ways.
Allow me for larger anticipation. Light and tiny electromechanical machines (aka cyborg insects) that can be released to monitor and clean weed, other insects, maybe bacterias.
Genetically modified squirrels! They pull weeds, put them in a bin and get a kibble! Its work or die, which keeps the gene line pure. And self-replicating too.
I don't really understand what he's talking about. Most farm equipment nowadays has thousands in dollars of GNSS equipment and a multitude of other sensors.
Because they will probably be purchased by the thousand, if not 10 thousand. Furthermore, they will be exposed to rugged environments where every extra component that can fail is a huge liability.
Exactly. Better to figure out a novel way to use fewer sensors equally well (cheaper) or the same sensors to better effect (more results).
Those things are hard though, plugging a few sensors into a Pi, slapping a LAMP stack on it then creating a web interface and smart-phone app you can show the VCs is easy by comparison.
slapping a LAMP stack on it then creating a web interface
You know, that'd be really great and all, but please, please, please don't discount the potential for rampant, malicious destruction of food supply for kicks, revenge or maybe even strategy/tactics in state conflict. Consider how warlords and pirates operate in other countries and failed states.
Food security is important, and a robot that can selectively clip and gouge weeds can do the same to both critical food and cash crops too.
Furthermore, simple competition, big corporate espionage and personal differences can still affect small organic growers as well. Human factors are real, and novelty is good but it doesn't cure quite everything.
This seems an important point that hadn't occurred to me. Our food chain is already pretty dependent on Internet-connected infrastructure working correctly, but it seems absolutely plausible that this will be a future attack vector if we network and automate farming machinery.
The threat is not from unknown robots sneaking in to uproot crops (after all why not do something fast and cheap like spreading salt?). The threat is that the farmer's own weeding robots will be hacked to uproot crops, probably by ze Russians. Firmware updates should be secured.
Actually I like that, it's closer to the essential truth.
I think George Carlin nailed it with his routine about flamethrowers.
> The very existence of flamethrowers proves that sometime, somewhere, someone said to themselves, 'You know, I want to set those people over there on fire, but I’m just not close enough to get the job done.
