Engineers Gave a Mushroom a Robot Body and Let It Run Wild (sciencealert.com) 64
An anonymous reader quotes a report from ScienceAlert: Nobody knows what sleeping mushrooms dream of when their vast mycelial networks flicker and pulse with electrochemical responses akin to those of our own brain cells. But given a chance, what might this web of impulses do if granted a moment of freedom? An interdisciplinary team of researchers from Cornell University in the US and the University of Florence in Italy took steps to find out, putting a culture of the edible mushroom species Pleurotus eryngii (also known as the king oyster mushroom) in control of a pair of vehicles, which can twitch and roll across a flat surface. Through a series of experiments, the researchers showed it was possible to use the mushroom's electrophysiological activity as a means of translating environmental cues into directives, which could, in turn, be used to drive a mechanical device's movements. "By growing mycelium into the electronics of a robot, we were able to allow the biohybrid machine to sense and respond to the environment," says senior researcher Rob Shepherd, a materials scientist at Cornell.
By applying algorithms based on the extracellular electrophysiology of P. eryngii mycelia and feeding the output into a microcontroller unit, the researchers used spikes of activity triggered by a stimulus -- in this case, UV light -- to toggle mechanical responses in two different kinds of mobile device. In controlled experiments, the team used the signals from a fungal culture to govern the movements of a five-limbed soft robot and a four-wheeled untethered vehicle. They were able to influence and override the 'natural' impulses produced by the fungi, demonstrating an ability to harness the system's sensory abilities to meet an end goal. "This kind of project is not just about controlling a robot," says Cornell bioroboticist Anand Mishra. "It is also about creating a true connection with the living system. Because once you hear the signal, you also understand what's going on. Maybe that signal is coming from some kind of stresses. So you're seeing the physical response, because those signals we can't visualize, but the robot is making a visualization." The research has been published in the journal Science Robotics.
By applying algorithms based on the extracellular electrophysiology of P. eryngii mycelia and feeding the output into a microcontroller unit, the researchers used spikes of activity triggered by a stimulus -- in this case, UV light -- to toggle mechanical responses in two different kinds of mobile device. In controlled experiments, the team used the signals from a fungal culture to govern the movements of a five-limbed soft robot and a four-wheeled untethered vehicle. They were able to influence and override the 'natural' impulses produced by the fungi, demonstrating an ability to harness the system's sensory abilities to meet an end goal. "This kind of project is not just about controlling a robot," says Cornell bioroboticist Anand Mishra. "It is also about creating a true connection with the living system. Because once you hear the signal, you also understand what's going on. Maybe that signal is coming from some kind of stresses. So you're seeing the physical response, because those signals we can't visualize, but the robot is making a visualization." The research has been published in the journal Science Robotics.
Re: (Score:1, Troll)
Re:No. Seriously. Why? (Score:5, Insightful)
You make it sound so easy.
But really, I don't read the article as trying to use a fungi as a neural network. But instead using it as a sensory system. It's more sophisticated than tying a motor to a pair of solar panels and an op-amp. But conceptually it's similar.
Re: (Score:1)
Re:That's why I chose lower life forms. (Score:5, Informative)
I don't know where you get that this is about an intelligence, other than that's the topic of all the clickbait these days. The article seems to be a biohybrid machine to me. There is a fair amount of built-in processing for sensory networks in a lot of animals, but fungi I bet are easier to deal with and prove the bio-hybrid aspect of the experiment.
Re: (Score:3)
I don't know where you get that this is about an intelligence
It's suggested several times that the network inside a fungus is similar to a neural network, a thing named precisely because it's like the network of neurons inside a brain. If this isn't about intelligence then it should be, at least at the level of searching desperately for the definition of what it is that separates the kind of intelligence you can definitely see in an advanced mammal pet (like a cat) from the lack of intelligence you see in a plant (or certain dogs). We don't know what intelligence is
Re: (Score:3)
Of course it's similar to a neural net. That doesn't mean it's used for learning. (Many are, but not all.) If someone who's read the original paper says it's used as a sensor, why not? The retina is a neural net that's used as a sensor. (I guess you could claim it does a bit of learning too, of course. But it's basically a sensor.)
Re: (Score:2)
Totally, but if you read the articles there are several more papers that are coming out. If you have a system like this then one of the things you should definitely be doing is see if the system can a) learn and b) do more complex interactions.
For example, slime moulds (with no inter-cell network because there is only one cell, though they have multiple nuclei) are able to do delayed gratification where they will go somewhere that they don't like being to get to somewhere where there's more food. If a mushr
Re: (Score:3)
Sadly I don't have an AAAS account and I can't read the paper. So I only have the article to go on.
It's suggested several times that the network inside a fungus is similar to a neural network,
If you're talking about the article and not the paper. The article mentioned neuron, neural, or neurobiology exactly 4 times. And a neural circuit is not a neural network. So perhaps that was where you started inferring neural network and intelligence? The article mentions the word intelligence zero times. As you have observed. Although I don't agree with your reasoning to explain the absence of the word.
It's
Re: No. Seriously. Why? (Score:2)
Re:No. Seriously. Why? (Score:5, Informative)
Mycelia networks are distributed neural networks. They take in environmental cues (CO2 concentrations, light levels, humidities, and temperatures), integrate them in a literal feed-forward system, and somehow decide where and when to fruit spore-bodies. It's fascinating and not well-understood.
Re: (Score:2)
Flatworms have like 30 neurons.
A fungi has thousands.
But your MAGA point is funny :D
Re: (Score:2)
Quoting Google quoting a Springer book:
"However, while the full complement of neurons in nematodes approximates 300, a figure of some ten times this number is estimated for flatworms."
Re: (Score:1)
So I mixed up some worms? ;)
No wonder I suck at fishing
Re:No. Seriously. Why? (Score:4, Informative)
They weren't trying to grow a neural net. They were investigating the existing mycorrhizal networks that fungi use to connect to each other. Why would they use something completely unrelated for the experiment?
Aside from that, it'll come as a surprise to you to find out that fungi seem to havea simple language and grammar [theguardian.com] and talk to each other.
We don't really understand what the fungi are doing, so this is one of many experiments designed to try and start figuring that out. You know. Science. Expansion of our total knowledge of the universe. Nerd stuff.
Juffo-Wup (Score:2)
>Aside from that, it'll come as a surprise to you to find out that fungi seem to havea simple language and grammar [theguardian.com] and talk to each other
Yeah, but all they keep talking about is this "Juffo-Wup" and nobody can figure out what they're talking about.
Day of the Robo-triffids (Score:2)
Re: Day of the Robo-triffids (Score:2)
Re: (Score:3)
Oh shit! H. G. Wells had it right! [wikipedia.org]
Re: (Score:2)
The Girl with All the Gifts (Score:3)
This would enable killer fungi to roam the countryside looking for prey - https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:2)
Weren't triffids supposed to be plants?
Maybe "robots from Yuggoth"?
Try Again (Score:5, Funny)
Better experiment: Engineers Gave a Robot Body Mushrooms and Let It Run Wild
This is how it starts (Score:3)
Next they develop interstellar travel and start shattering M class worlds to spawn their offspring. Don't say you weren't warned. [fandom.com]
Next thing you know.. (Score:2)
Re:Next thing you know.. (Score:4, Insightful)
Re: (Score:3)
Re: (Score:2)
In Discovery though they implied that dilithium had some sort of 'subspace' or perhaps quantum-level properties that extended outside of normal spacetime, or something similar to that, but as with many thing
Re: (Score:3)
I think the door to this multiverse crap was cracked open with the TNG episode "Parallels"
No, that door was opened in the 2nd season of TOS with the Mirror Mirror [wikipedia.org] episode where we get to see an evil mirror universe and Spock with a moustache and goatee.
Re: (Score:2)
Re: (Score:2)
I watched all of Discovery simply because I'm a compulsive 'completionist', though. There were some good moments to the show, but lots and lots of things that just made me roll my eyes and shake my head sadly.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Only a traitor in the colloquial sense, not in the sense defined by the constitution. And if he's a convicted felon (I haven't been following that) it's only of a state law, not of a federal law. So while there may be some states he couldn't run in, that wouldn't affect his legality as a candidate.
Re: (Score:2)
By the way: there is only one right answer to the above question.
Do mushrooms dream (Score:3)
Do mushrooms dream of mycological sheep?
Re: (Score:1)
And, what do mushrooms that can move around do? I read the F'N article and there's no mention of what a mushroom that is let "Run wild" does. I would assume that it seeks out a moist dark place, but you never know, it could head to the beach.
Headcrabs (Score:2)
Re: (Score:2)
Mushroom overlords (Score:3)
I for one would welcome them. Let this be recorded, please translate into mushroom language.
https://www.youtube.com/watch?... [youtube.com]
Another minor experiment, another slow news day. (Score:5, Interesting)
Article is paywalled, so I couldn't read more than the abstract.
If the main idea was to demonstrate that they can read innate mycelial action potentials, free of mechanical or electrical interference, then they succeeded.
Everything else about the article is a giant "so what".
They sensed action potentials then used that to trigger an actuator, in this case a "robot" that flexes its limbs, making it jump.
So? ! So what?
I could use any input and any output.
I could monitor my TV, and at a certain threshold for yellow, trigger a tone from my speaker.
I could monitor outdoor temperature, and at a threshold, turn on my water sprinkler.
I could monitor my kid's homework, and if he gets an A, I buy him an ice cream.
These are simple passive responses using the most rudimentary systems and concepts of control. The could have hooked the mushrooms to a flashlight, a buzzer, a camera shutter, a toy train on a track - anything. None of this demonstrates anything other than sensing a signal and visualizing it or actuating something.
This is also no different than Kirlian photography seeing an aura around a leaf, or any other demonstration that biological systems can generate electrical signals.
All that was required was a basic physiological recorder, like an ECG or EEG, to show a tracing of electrical activity at baseline and with stimulus - a toadstool polygraph. That would have sufficed as a study, especially since it would have had greater power to verify that environmental noise was filtered out.
It's the old adage : if you can't dazzle them with brilliance, baffle them with bullshit.
If the idea was to ultimately show that mushrooms are mentating in some way (or even organizing complex non-intelligent responses) this wasn't the thing to do.
Instead, they should have closed the loop. Let the mycelium grow onto a chip that has (1) input sensing electrodes (so the shrooms can "feel" an input) and (2) output electrodes that respond to focal mycelial depolarizations and can drive external devices, like motors or an LED array. Then, feed output or external stimuli back into the input electrodes, and see if the output pattern eventually progresses from random jumble to organized pattern or purposeful motion. Therein is something worthy of being called "robotics" or control.
Why not, let's not be killjoys (Score:3)
There are definite theories that directional movement is one of the drivers for animal intelligence. If it's a correct thesis then experimenting on potentially intelligent systems to see how they handle being in control of movement makes sense. It's kind of a more basic Turing test. Which is to say, no, we don't know exactly if it's the right thing to look at but yes, we do know 100% that we learn things and get new ideas when we try to do it.
It's also clearly more fun and interesting than hacking around wi
Re: (Score:3)
Look, he made a mushroom-robot interface. He seems like a fun guy.
Re: (Score:1)
Re: (Score:1)
All this technical talk about how they made it so the mushroom could do stuff, but not a single bit about what the mushroom actually did. What is the Running Wild that the title speaks of entail?
Indeed - two-way comms and learning are required (Score:2)
Show me a mushroom where you get a signal, make something happen... and the signal changes in response to the triggered actions in a way that alters the actions.
Put a mushroom on wheels that will seek nutrients or moisture when lacking them. Use nothing more than appropriate sensors and motors plugged into the mushroom. If that worked, it'd be kind of special even though realistically it'd just be a slightly augmented stimulus-response mechanism.
Re: (Score:2)
Yes, exactly. That's what I was getting at in my post, and you illustrated it in a very clear and inventive way. Thanks.
Aha! (Score:1)
Now I know where they hired some of my coworkers.
I didn't see this one coming. (Score:2)
Now a Terminator-The Last of Us crossover is in the making. Yay!
Obligatory (Score:1)
Guys (Score:2)
The Last of Us (Score:2)
Even better, give mushrooms human bodies.
Ohhh! (Score:1)
Skroderiders?
Very interesting (Score:3)
But a thought comes to mind, If fungi are shown by these experiments to actually have something akin to thought, how will vegans react. I'm not seeing any issues with people noshing on slime molds, but a lot of yummy mushrooms are out there.
I do wish to welcome our fungus overlords.
Skroderiders from A Fire Upon the Deep by Vinge (Score:2)
Which are sentient tree with wheels. Those wheels were designed by a much high level of intelligence. Though they can make and must have the wheels to be sentient, no one understand how they work.
Turns out they're unwitting sleeper agents for a very malevolent intelligence eons ago and it helped that entity (Blight) to take over a large swath of the galaxy.
It's a potato clock (Score:2)
It looks to me like they're just using raw electrical potential derived from a sensor stuck into a mushroom to feed into the robot drive controller. Any other input signal, RNG source or antenna picking up random crap would have done the same thing.
next experiment (Score:2)
monitor solar flare activity, call it "solar cues" and turn them into directives and have a robot do something with it
yayy
we did a thing
and in a not unexpected finding (Score:2)
Goddamnit, science! (Score:1)
Please, just stop. Don't we have enough existential threats to humanity going on right now for one year? Stop trying to power up the lower-order life.