Learning machines that work together like a group of ants or fish? Is it just a dream? Continuous technology development allows AI to be used in various sectors. The use of intelligent machines allows us to automate production and create new solutions and technologies. Therefore, the involvement of robots in the construction of housing or other types of cooperation can be applied in many projects. How can this affect the development of humanity?
I encourage you to watch the TED and tell us what you think. How it will change our perception of the world and where it can be applied.
Questions:
How would you use such programmed robots?
Do you think this is an interesting idea for the
development of robotics?
Sources:
Comments
I'm actually currently developing a platform for massive parallelized processing of textual information based on swarm intelligence ideas. The advantages of such architecture are no single point of failure (auto-recovery) and it scalability.
Swarm robotics, except of obvious military uses, seems to bring promising results in the search and rescue operations. Especially in the areas that are dangerous and hard to access for humans. So many rescue workers die while trying to save lives: in the mountains, during the earthquakes, fires or at war. I believe using swarm robotics in such disaster recovery missions would significantly decrease the numbers of fatalities.
Nowadays the world os developing itself really fast and in very interesting directions. Of course, robotics is one of them. I think that we need to develop as far as we can and in any direction we can, and then, maybe, we will develop the better world than we have today
Since every single entity in the swarm can be a little different, than another, we could use them for example to make pretty sophisticated decision-making systems, which would be somehow similar to what we encounter in our lives - voting.
I think that every single idea in robotics, which introduces something new is worth researching, and this one is particularly interesting.
Actually I implemented propagation knowledge protocol and I have to admit it was hard.
I believe something like self driving cars in the future will use something like this to coordinate. And eventually we wont need traffic lights.
But until I hear something like "maybe one day we could"(and this phrase occurred a lot in this talk), it's still a concept.
As for the second question, I agree that it's highly likely we're gonna see it become one of the main branches of machine learning in the future. So much of our knowledge comes from observing the future, that the value of cooperation pattern isn't one we could possibly pass on.
Such collective minds was shown in Rick & Morty and Mass Effect Series. First one was mentioned by Piotr Prystupa so I will be focused on Mass Effect Series and race called the Geth.
The Geth was a race of machines which had one purpose - serving people. Sounds simple, huh? Every machine has his own mind but together they was united mind. They was learning and one day the geth asked his master "This that unit have a soul?". People get scared and wanted to exterminate the Geth. With no luck.
My point is, that if we will not be sure about the direction of learning progress of machines that it will be better to not develop such units on big scale.
I believe the most useful way to use such robots is to use them as a brick :) Imagine the house you can freely move from one part of the country to another using only your own PC. You'd also be able to easily rearrange the rooms.
I wonder what the end game for artificial intelligence is. I doubt that anyone will succeed in making fully sapient ai in the next 100 years so maybe advanced collective "consciousness" would be something easier to achieve and possibly more useful.