Fluorescent Plants, Reversing Paralysis and a Brain implant that lets us type using our thoughts.
This week we discover how plants can tell us when they are sick way before we can see any affect on the plant. We examine a new drug that may reverse paralysis. We investigate a brain implant that allows us to type with our imaginations and finally we find a better use for the software that creates deep fake images.
Roughly 20% of farmer’s crops in the US are destroyed each harvest by pests and diseases. Plants naturally emit chemical signals when they are attacked by pests, fungi and other dangers. Till now, humans have had no way of knowing when these attacks happen so we spray entire fields with fungicides, pesticides and other chemicals.
Californian startup InnerPlant has developed a way for plants to communicate with farmers to let them know when they are under attack from a pest or disease. InnerPlant’s seed technology “recodes” the plant’s DNA to create a fluorescent protein that lights up the plants leaves whenever it is in distress.
Using satellites, drones or field based equipment the illuminations in the leaves will alert farmers to a problem before it spreads. The fields don’t turn purple, the right equipment is needed to detect the fluorescence.
Plant’s have an immune system reaction that will activate at the emergence of stress. This may come from a fungus, disease or even inefficient water supply. Prior to planting, InnerPlant embeds the ability to generate a fluorescent protein into the individual seeds of crops. When stress occurs the protein gives off an optical signal that’s unique and doesn’t natively occur. That signal is easily read and action can be taken on a real time basis. This process allows farmers to reduce losses from pathogens and to enable precise remedies. This eliminates the unnecessary use of chemicals.
InnerPlant has released their InnerTomato product and InnerSoy and InnerCotton will be released shortly.
A team of researchers at NorthWestern University in Chicago have developed a novel drug that is capable of enabling cells to regenerate. They were able to reverse paralysis in mice who had suffered spinal injuries. The mice were able to walk again within four weeks of treatment.
There has been a lot of research into reversing paralysis. Some techniques concentrate on stem cells to generate new neurons and aid nerve repair, others directly inject proteins into the spine. The NorthWestern team took an engineering oriented approach. They used nano fibers to mimic the serpentine structure of the extracellular metrics that surrounds the tissues known to support nerve cell function.
Each fibre is incredibly thin, 1/10,000th of the width of a human hair. The fibre is composed of peptides that send specific signals to trigger the regeneration of nerves. The researchers used a gel that was injected directly into the tissue wrapped around the spinal cords of lab mice. Four weeks after the treatment the mice actually regained their ability to walk nearly as normally as they did before their injuries.
Upon examination of the results the team discovered that substantial improvements had taken hold in the mice’s spinal cords. The severed extensions of the neurons had regenerated. Scar tissue usually forms a physical barrier however this was also diminished. The insulating layer of axions known as myelin, the layer responsible for sending electrical signals had also reformed. New blood vessels had come into being allowing the transfer of nutrients.
This gel is the first of what may be a new kind of medicine called “supramolecular drugs”. This is where a great number of molecules are assembled to create a desired outcome rather than relying on a single molecule as in more traditional medicine. Further animal testing is required before human trials can begin.
Brain implant that turns thoughts into text
Stanford University researchers have developed a brain implant that allowed a man with paralyzed hands to type up to 90 characters per minute by thinking the words.
Previous implant systems relied on patients to move a cursor to specific characters on a digital keyboard with their mind. This new implant allows the user to imagine himself writing a word with his hand. The AI then decodes the neural signals and displays the word on the screen.
The system has it’s own autocorrect function (and don’t we all love autocorrect) that resulted in a 99 percent accuracy rate. There is still some way to go to commercialization. Firstly you need to convince people that implanting electrode based sensors into their brain is safe. Then surgery is expensive and recovery slow. As we learn more over the coming decades, brain implants may move from the realm of medicine to consumer electronics.
Using DeepFake AI for Good
Penn State scientists believe that they are able to use the AI that generates convincing but fake human images to power the next wave of innovations in materials design.
The scientists trained a generative adversarial network (GAN) to create novel refractory high entropy alloys. These are materials that can withstand ultra high temperatures while maintaining their strength. Common uses are in turbine blades and rockets.
We have spoken about GANs previously however a quick recap. A GAN is two AIs. One generates an image and the other checks the image for accuracy. Once an image is created (and this takes microseconds) the second AI gives feedback on the attempt, the first AI makes minor corrections and presents the new image to the second AI. The feedback loop continues until the second AI recognizes the image as the type of image desired.
The team examined hundreds of examples of alloys to create a training set. Their network included an AI to create a new alloy and a critic to try and discern if the alloy looked realistic compared to the training dataset. With sufficient computing power (which is quite cheap nowadays) many iterations can be completed thus gradually improving the model.
The team’s results showed that generative models can learn complex relationships in order to generate novelty on demand. Traditionally we relied on human intuition to find patterns and discover breakthroughs. As materials chemistry advances this becomes more and more difficult. With this system you can generate 100 or even 1,000 potential new materials in milliseconds. This allows scientists to examine the suggested materials rather than to try and guess things that might work.
Flying Car approved for Takeoff
We have spoken regularly about Autonomous VTOL or self driving flying cars. SkyDrive, a Tokyo based startup that is developing an eVTOL have now received a safety certification for their flying car. The certification states that the “design, structure, strength and performance meet the necessary safety and environmental requirements” needed for the vehicle. SkyDrive aims to commence a flying taxi service in Osaka in 2025.
Paying it Forward
If you have a start-up or know of a start-up that has a product ready for market please let me know. I would be happy to have a look and feature the startup in this newsletter. Also if any startups need introductions please get in touch and I will help where I can.
If you have any questions or comments please email me via my website craigcarlyon.com or comment below.
I would also appreciate it if you could forward this newsletter to anyone that you think might be interested.
Till next week.