Why forgetting things is important?
And how to benefit that thing to maximize the power of the computer systems? How to create algorithms, what can remove unnecessary information from the database automatically, and deny the database enlarge too big?
As so important as getting and storing the data is removing the data, and if we cannot forget things, we would be in trouble. Every image, what is stored in memory is reserving a part of the space, which is reserved for memory. Even in human brains, where there are billions and billions of memory units, the memory would be filled, if we cannot forget. And the thing, what makes this process interesting, is how brains are making that thing.
How human brains are selecting the memories, what they can remove safely? The thing that makes brains so effective is that they can search the data from the multiple points at the same time, and the effect of the memory would become that the brains are sending the query to every part of it at the same time.
The query would be coming to the neuron groups at the first, and it contains the information, what would the data involve, what is needed. If the person would need data, when goes the next train to somewhere, the query would send to every neuron groups and the answer what the group of the hand coordination neurons would give could be "we are neurons, what is coordinating the movement of the hand, and we have no that data".
Then that will release those neurons to other purposes. And the group, what handles the visual memory would start to search that data in the cells, what hare storing that kind of data, and this means that the only groups, what have this data will search it. The brains would sort the data by the sensor, what collects it.
When we want to remember the train schedule, our brains would start to find out, what sensor would give us that kind of data, and then only the cells, what is handling that thing are starting to find the data, what is stored in memory. But the thing, what makes this process very interesting is that our brains are created to protect us from the beasts of the dark African night.
It is not made to handle train schedules. The thing, that makes our memory a little bit embarrassing is that we are remembering negative things better than positive things. And the reason for that is the protective purpose of memory. This is why the negative thing is storing easily and negative things are the first, what is coming in mind.
When brains are selecting things, what are removed from the memory, they are using the methodology, what is so interesting, and if some memories are not needed in a certain time, it would be removed. That means that the axon, what is connected to that memory block would be removed, and this releases that connection to another purpose.
The neuron, which has many axons and connections are multifunctional. But the problem is that this kind of neuron is slow because it must search the right connection simultaneously by using multiple connections. And also this kind of cell is getting very much queries. Also storing the unnecessary data makes neuron slow because it must check also that data when it searches the answer for the query. If the neuron would have only one connection the use of the neuron is limited, but the neuron would be fast. If the neuron has only one duty, it would make that thing effectively and fast.
The artificial intelligence means the system, what can learn things, or what can automatize some operations, what is needing a very strong concentration. That case could be the shopping center, where are thousands of customers per day. Following each person by using the guards and security cameras would be a very hard duty. And in this case, the computer-based systems can show people, who are trying to steal something.
So when we are creating the database, we must also create a model, what selects the objects or things from the data flow, and sort it to a certain part of the database. The database that is well sorted is fast, and that means that artificial intelligence can select an object by its form and in this case, the system must use some method, how to select the object, what is stored in a database. In the shopping center, that thing might be that the thief alarm is launching, and the camera takes an image of that person.
If the surveillance system would store all data, the databases would be too big, and unnecessary data is making the database slow because the artificial intelligence must check also the non-relevant data. The idea of computer based-data handling is that the system would store only the data, which is important. The important data is an example of the attempts to steal something, and the face recognition programs can be used to follow the movements of the person before guards will catch that person. The artificial intelligence can also follow things like knives or forgotten baggage, and report that thing for security personnel.
How computers can benefit this information?
This knowledge, what is getting from the neurology can benefit in the neural networks, which are using hybrid topology. In the traditional star-topological network models is the server in the middle of the network, and every workstation takes the data from the hard disk of the servers. The other topology is the ring-version, where is no server, and the workstations are in the ring, and the queries are traveling simultaneously by each workstation.
But the hybrid topology means that the data is stored and it can search from each unit of the network. That means that the query goes to the server or central unit, but also it uses the ring of the workstations. This means that it is less vulnerable for the damages, and the system can sort the data that a certain type of files are stored near the user, which uses most of a certain type of data, and also the system can store files in two or more places. The ability to use the system by using multiple topologies makes it less vulnerable to damages.
This is the thing, why forgetting is important. And if we would transfer this model to artificial intelligence, the neural network, where all parts of computers are handling certain things, the hard disks can be filled very fast, if every single thing, what surveillance cameras are inputting to the databases. And if databases are very big, they would turn slow and even they are well sorted very effectively, they would be filled. So artificial intelligence must remove unnecessary data.
Comments
Post a Comment