Story Of Google AlphaGo


Today we look about stranger things done by modern computers with artificial intelligence. How did computers with artificial intelligence work in the past? They looked about all possible outcomes and choose the most logical outcome. It is called the brute force method.

IBM Watson
But after a few years, computers could collect information by reading things written in general languages and watching videos and pictures. IBM Watson computer is an example of that. However, our brains can work more than these simple works. As an example, we can make decisions by using our senses rather than depending only on the logic. Our brain works like that because it can learn new things from our memories and our experiences. Not only that these learned things can recognize similar situations and use them in similar problems. But it is not such easy work to teach a computer to work like this. Because even we don't know how we get decisions like that. So we have many difficulties to make a logic using this.

Many companies all over the world work hard for this purpose. Deepmind is a company that got success in this work and they went forward before others. Then Google company brought the Deepmind company because of its successful journey. Deepmind company could make an artificial neural network. It can learn new things through visuals and hearings without any help from a man and also it can use its learned things in another situation. As an example, this artificial neural network sees a computer game for a while and play it for a little time and learn the game very well. When it plays the game for some time it can beat a human easily. However, scientists could not understand whether this artificial neural network can take decisions by its sense as a human. These scientists took a decision to arrange a game for this artificial neural network with a man. They thought it is the best way to confirm this.

In the past chess game used to works like this. But chess game has only a few combinations of probability set. So in this situation, the computer can look at every possible outcome and find the most logical outcome(Brute Force) then it can beat a human easily. Many computers did that in the past.

Scientists need a complex game than chess to do this experiment correctly. Finally, they choose a game called Go. It is a Chinese board game. It looks like a Dam board. The main reason for choosing this game is it is a very complex game. Even a human can't learn this game completely. The combinations of probability set of this game are larger than the count of all atoms in our universe. Even a human can't play this Go game completely logically because this game has many combinations of probability set. If you ask the reason for a move from the best Go player in the world he also can't give the reason for that move. The scientists in the Deepmind company make a decision to give a chance to their artificial neural network to play against the best Go player in the world. They create this artificial neural network for this task especially. And it called AlphaGo.
Before the starting of this game, many specialists predicted that this computer can't win against this player. Because this Go game wants the sense in our brain as a special requirement. But finally, the AlfaGo computer could win the game. So the Deepmind company proved as their computer has a sense like humans. This is a big success that gets by a computer that has artificial intelligence.




The Facebook company also created an artificial neural network. And the company gave a chance to communicate between two computers in their own way. After some times scientists could understand that these two computers' communication language is not clear to us. It is their own language. Also, it is a secret language. Because of this scientists have to terminate the experiment.  But later they understand that these two computers create a secret language because our language is not an efficient way to communicate. So they created their own method to communicate efficiently than us. 


Today these artificial neural network computers can use their own sense and take decisions like us and go beyond the logical boundary and find new things. I think this news scared you. Because nowadays computers can think like us, create new things faster than humans. But we should understand artificial intelligence also a human creation. We should not afraid to use its abilities. We use a hammer to tighten a nut, When we can't run 100kmph fast then we use a car, When we can't fly then we use an airplane. Also, we can hit by a hammer to hurt a man, we can drive a car to a man and make damage to him, all of us saw the 9/11 attack done by airplanes. Artificial intelligence also likes that we can use it for good works or bad works. You also can create a artificial intelligence by using the TensorFlow flat form. If one-day artificial intelligence go beyond our brain it is called a singularity. Just think how many years it will take to get this singularity? 

Comments

Post a Comment

Popular posts from this blog

Programming Using GNU Octave

Image
  Questions: 1 A small object is projected from a toy gun to travel some distance and land on the ground. 1. Write a function to model the dynamics of the object in Octave called f_trajectory. f_trajectory will accept following arguments in the same order (measured in the given units). Also specify the default values for each parameter as given in brackets. i. h(5) : Initial height of the projectile (m) ii. g(9.81) : acceleration from gravity (m/s/s) iii. v(25) : speed at initial release (m/s) iv. theta(π/4) : angle of release (radians) v. t : a vector representing the points in time For default values of t check if the variable ‘exist’ with name ‘t’ and type ‘var’ if not initialize it as a row vector starting from 0 and goes to 5 with increments of 0.1. If x is distance and y is height, the equations below describe their dependence on time and all the other parameters. 𝑥(𝑡)=𝑣𝑐𝑜𝑠(𝜃)𝑡 ----------------(eq1) 𝑦(𝑡)=ℎ+𝑣𝑠𝑖𝑛(𝜃)𝑡− 12𝑔𝑡2 ---------------(eq2) Convert the abo...
Read more

Stack Implementation

Image
                Stack Implementation in C Language #include<stdio.h> int MAXSIZE =10; // Declaring the maximum size of the stack int stack[10];  int top = -1; int peek(){ // peek function to display the top element of the stack return stack[top]; } int isFull(){ // isFull function to check whether the stack is full or not if(top==MAXSIZE) return 1; else return 0; } int isEmpty(){ // isEmpty function to check whether the stack is empty or not if(top==-1) return 1; else return 0; } int pop(){ //pop function to delete the top element of the stack if(!isEmpty()){ int value; value = stack[top]; top =  top -1; return value; } else printf("Stack is empty, cannot pop elements\n"); } int push(int value){ // push function to insert elements into the stack if(!isFull()){ top = top + 1; stack[top] = value; } else printf("Stack is full, cannot insert mo...
Read more

Implement a malloc

Image
Implement a malloc-like memory allocation library 1. Declare an array of 25000 bytes. 2. You must implement a function that is like malloc(). Call it MyMalloc(). Its signature is similar to malloc(). You should also implement MyFree() which has a signature and functionality that are similar to free(). 3. MyMalloc() allocates memory only from the previously mentioned array of 25000 bytes. 4. All the data structures that are required to manage the memory must also reside within the same array. 5. MyMalloc() and MyFree() must be in a file called mymalloc.c. You should also provide a suitable header file mymalloc.h. 6. Our test program contains the main() and it includes mymalloc.h. Do not include a main() in your mymalloc.c. Sample Code in C: #include <stdio.h> #include <stddef.h> #include "mymalloc.h" struct node *freearray=(void*)array; // to point the first place in array[25000] void *mymalloc(size_t bytes){ // get the value of bytes that...
Read more