Subscribe
The reason why Facebook is doing robotics research

The reason why Facebook is doing robotics research

621 Views

It is weird when you hear that one of the leading social networking sites in now perusing their research on robotics. Facebook is not investing their resource in search of a better competent search engine rather it is spending it on robotics research.

Facebook is one of the biggest organizations with a lot of competing priorities. All these developments will not affect Facebook directly. It is expected that the things that the company will be learning would impact the platform in surprising ways.

Robotics is one of the latest areas where Facebook is spending a lot of time. The bleeding edge work in Artificial intelligence is something Facebook is also taking very seriously. All the mechanics which can be called Artificial Intelligence can govern all sorts of things which include camera effect along with automated moderation of restricted content.

The best thing is the Artificial Intelligence and robotics kind of naturally overlap each other and this is why all the events will be covering both of them. It is not something exceptional that Facebook has a strong interest in Artificial Intelligence.

Learn to walk from scratch

Walking is one of the hardest actions when it comes to 6 legged robots. It is very hard to make sure how the movement of the robots will be done. The only reason we are able to walk is that we have the capability to understand and these robots will also need something similar. This is when the robotics come into play. The team tries to teach the robot to walk by itself so that it can perform more like a human.

This is not something new when it comes to Artificial Intelligence and robotics. All these algorithms go back for a long time. A lot of people have written interesting papers on these topics in the past.

Some of the basic priorities are set to the robots. Rewards are also given moving forward so that the robot can understand that it is being appreciated and walk on its own. The team did not make the robot teach about walking but gave the robot the opportunity to figure it out on its own. The main goal here is to make sure that there is a reduction in the amount of time it takes for the robot to go from zero to reliable locomotion from weeks to hours.

What will this be useful for?

Facebook has a wide range of data which is very complex in nature. The main idea here is to make sure that the artificial intelligence learns in a short period of time with the help of some goals and rules.

Learning how Artificial Intelligence teaches themselves and how exactly you can remove all the roadblocks like all the mistaken priorities along with wired data hoarding and cheating the rules is important for agents which are meant to lose in both the virtual as well as the real world. If there is a time when there will be a humanitarian crisis, then Facebook can monitor its platform with the help of Artificial Intelligence. The Artificial Intelligence model will end up helping and will also be informed by the autodidactic efficiencies that turn up here.

Leveraging curiosity

These works are not very visual in nature. Everyone is very curious in a certain degree to know more than who exactly kills the cat. Facebook has successfully put the concept of curiosity in the robot’s arm and has asked it to perform different tasks.

It is very odd that people can simply imbue a robot arm with curiosity. The main meaning of the term is that Artificial Intelligence is responsible for the actions of the arm. The decision of how to grip is totally on the arm and it will take the help of Artificial Intelligence to perform all these tasks.

These will be implemented in a lot of things. It can be used in a camera. When twisting a camera, it can give a better view. The first thing that it will look at is the target area. Then it will check the distance between both points. The Artificial Intelligence latitude will find the action that will help to increase the confidence and this could eventually let the completed task faster. In the beginning, curiosity might be a bit slow to react to.

What can we use this for?

When it comes to vision Facebook has big ideas. This will be implemented in both cameras as well as image work. This could also be used in different devices like a portal which helps to follow you around the room with its face. Learning about the environment is also useful for both applications.

Any kind of camera operating applications in the application or device like on Facebook generally analyze the images and stores all the information which are important to save. A face helps to put a lot of algorithm in place so that the image can be made better. If there is a QR code, then the address is also auto-triggered and people end up in that landing page.

If the camera or any kind of gadget is left with these tasks, then they will end up producing CPU usage spikes along with some visual latency in the images. Instead of using the Artificial Intelligence agent is exerting curiosity t check all these things when there is a bit of uncertainty in the scene then that is a happy medium. This is the simple way how Facebook ends up prioritizing all the important ones.

Vision is very important when it comes to robotics. Many robots have a lot of sensors and sound modalities but they have one thing missing which is the vision. It is important to have a good tactile interfaces Nerveless Facebook wants to look into the data which is tactile in nature as a surrogate for visual data.

If we talk about people with the bad vision they generally sue their touch to feel their surroundings. Sensing via touch plays a very important role here. The robots will have to navigate around so that they have a clear understanding of what kind of surroundings are there beside it. There is a big overlap between all these concepts. All the Facebook researchers have ended up deploying Artificial Intelligence model that generally helps in deciding what action to take depending on the video that is captured in high resolution.

It is true that the algorithm does not really care if there is an image of the world as long as the data is displayed visually. It will basically look for patterns which will help the robot to get a better understanding of the entire place just like a photographic image.

What can this be used for?

It is not very sure if Facebook really wants to connect with the people with this technology. This has been made clear that it is not the touch which is just dependent on artificial intelligence. There are other factors which come into play here.

If you have ended up presenting two distinct objects for the first time in front of the robot, then it will be very trivial to distinguish between each other by keeping their eyes closed. This is because by seeing the items we can differentiate between both objects which are not possible when our eyes are closed.

Artificial Intelligence agent may end up needing to transfer from one domain to another. All the auditory data simply telling about the grip on how exactly it will hold up an object.

Facebook is now working hard to expand in the influence of walled garden applications and services into the rich and totally unstructured world. We can see any Facebook robots recently that’s for sure.