Eagle eyed readers will have noticed that I sent last week’s newsletter out on a Wednesday rather than the usual Thursday. A big thanks to one alert reader pointing this fact out to me. I had no idea until he sent me an email. I blame lockdown where every day has merged into one. Our first story this week may help me once we emerge from lockdown to reengage with the normality of the world. Scientists have found a way to reverse memory loss in mice. Early days in the research but hopefully it is ready for me when we get out of lockdown.
We also discover a new way for Autonomous Vehicles to learn how to drive on the road and we look at the Metaverse and what it might become in the near future. Finally we delve into a range of apps and programs that might just help you make it to the Paris Olympics in 2024.
Reversing Memory Loss
Scientists at the University of Cambridge have found the cause of aging related memory loss in mice. During their research they discovered that they could also reverse that memory loss in old mice. There is hope for us all!
The scaffolding that holds the brain’s neurons in place is a cartilage like structure called Perineuronal Nets (PNNs). The PNNs surround the neurons in our brain. They first appear in humans at the age of 5 after a period of neuroplasticity during which the brain learns, adapts and makes memories.
The role of PNNs is to reduce the plasticity in the brain. This increases the efficiency of the brain in performing regular tasks however it does make the brain less capable of creating new memories. This is one reason why learning becomes more difficult as we age. The PNN constituent, chondroitin 4-suffate is responsible for this change.
Our brains are not fixed, they want to create new memories. We want to remember whose shout it is for coffee (it’s yours). The brain uses a similar compound chondroitin 6-suffate to improve the ability to make memories and learn new things. The brain needs to maintain a balance of these two chemicals however as we age the chondroitin 4-suffate increases, reducing our ability to make new memories.
The researchers used a virus to infect the mice’s PNNs and increase the levels of chondroitin 6-suffate. This restored the older mice’s memory and ability to learn to levels that they had when they were much younger. Humans and mice share the same brain structure. Humans would be able to take an oral drug to increase the levels of chondroitin 6-suffate (if we remember to take the pill in the first place). The team is now working to see if the treatment can work for Alzheimer’s and other diseases.
How can Autonomous Vehicles learn?
Currently self driving cars learn by using powerful machine learning algorithms that use vast amounts of driving data collected by researchers driving millions and millions of miles.
A team at Boston University have developed an alternative way. If self driving cars could learn to drive in the same way that babies learn to walk, by watching and mimicking others around them, they would require far less driving data.
The original idea came from a desire to share more training data and to generate more cooperation amongst researchers in the field of autonomous driving. Unfortunately much of the research is being conducted by the world’s largest car companies who see their data collection as a strategic advantage and are unwilling to share. The problem of autonomous driving is so large, it is likely that no single company can solve the problem quickly on their own.
The researchers proposed learning algorithm works by estimating the viewpoints and blind spots of other nearby cars to create a bird’s eye view map of the surrounding environment. This map helps the car to detect obstacles (other cars, pedestrians etc.) and understand how other cars turn, negotiate and give way without crashing into anything.
This method allows the self driving car to learn by translating the actions of surrounding vehicles into their own frame of reference. The other cars may be human driven cars without sensors or other self driving cars. Since observations from all of the surrounding cars in a scene are central, the algorithm’s training and learning it encourages sharing of data.
The researchers tested their algorithm by having autonomous cars navigate two virtual towns, one with straightforward turns and obstacles and another with unexpected twits and turns (e.g. 5 way intersections). With just one hour of self driving data to train the machine (remember that current models have up to millions of hours of data) the vehicles arrived safely at their destination 92% of the time. Previous models required hours of data whereas this model could learn to drive relatively safely with just 10 minutes of driving data.
The current results are promising however there are challenges still to overcome. The variations between drivers, noise in sensor measurements and varying perspectives across watched vehicles still need to be overcome. The system may also be adapted for training delivery robots and drones.
Welcome to the Metaverse
Facebook have announced that they will become a Metaverse first company. So what is the Metaverse. The term was coined by Neal Stephenson in his 1992 novel, “Snow Crash” where people don virtual reality headsets to interact inside a game like digital world. This would allow you to sit on a couch and chat with a friend who is physically thousands of miles away. Working in the office may mean donning the headset and joining a virtual meeting whilst you are lying on the couch at home. It is a blend of the physical world with a virtual world.
Many technological advances come from games. It is easy to imagine a world where gamers enter virtual world’s. In the early 2000’s the game Second Life allowed people to create digital avatars that could interact and shop with real money. More recently Decentraland, a virtual world where visitors can watch concerts, visit art galleries and gamble in virtual casinos, has sold plots of land for hundreds of thousands of dollars. Epic Games (creator of Fortnite) recently raised $1billion to support its’ vision for the metaverse.
There is a long way to go in the development of the metaverse and the headsets that will allow immersion in these alternative worlds. We spoke about some Apple rumors in May. Facebook owns the Oculus Headset in which it is investing billions. Whilst clunky now, it will improve rapidly in the coming years. There is no doubt we will hear more about the Metaverse in the coming years.
Preparing for the next Olympics
Whilst we are inundated with the Olympics, I thought it might be a good idea to highlight a number of new startups that are developing products that might help you to make the next Olympics. Remember that Andrew Hoy won a team silver and an individual bronze at this Olympics at the age of 62. He said that he hopes to still be riding at the 2032 Brisbane Olympics where he will be 73. There is still time for most of my readers.
Asensei
Asensei is an online rowing coach. Several UK rowers used the program to prepare for the Tokyo Olympics. The AI sets personalized rowing programs that have been developed by world class instructors. The program tracks your split times, stroke rates and uses that data to help your each your potential. The program can also provide feedback on technique and focuses on the quality of your form.
Obefitness
Obefitness is one of the many fitness apps that provide a combination of live (20 per day) and on demand workouts (over 6,000). There is a huge range of class types to choose from including Yoga, HIIT, Pilates, Dance and boxing.
Boost
Boost is an AI driven scouting program that will help you to scout your opponents style of play. You can use broadcast vision of your opponents and have the algorithm provide insights into their attacking and defensive patterns. Originally developed for US college coaches to provide insight into opponents, the program can also provide insights into your own team’s performance. The program is also able to create highlights for us in fan engagement and social media.
Swing Vision
Swing Vision is a AI driven tennis coach. Automated shot tracking, video analysis and line calling via your smartphone. The AI will analyze your shot type, spin type and ball speed whilst giving feedback on shot placement, contact and rally length. Footwork, posture and positioning are also analyzed.
Komo
Komo Digital Engagement is creating custom digital experiences for the Australian Olympic Committee. The app allows you to create interactive content, competitions and gamified experiences using templates. The system will help you better understand your audience and develop better targeted marketing programs. This app may not get you to the Olympics as a competitor however every team needs support staff to maximize the experience for fans.
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.