Bayou: A neural network based system for automatic Java code generation

Bayou is a system that automatically generates related codes depending on the needs of Java developers while using the API of this language. In fact, this artificial intelligence technology works in such a way that after the developer specifies the desired task, the system makes the snippet needed to perform that task available to the developer using neural networks.

Regarding the general mechanism of this service, it should be said that the developer wrote the overall design of his program (Draft Program), then Bayou started production according to the methods used from the Java API on the one hand and also the type of variables used in this program on the other hand. Codes. In other words, the source code written by the developer is analyzed using a method called Neural Sketch Learning, which is a type of neural network called Gaussian Encoder-Decoder.

For example, we enter the following codes in this service, which is available online through the Bayou link:

import java.io.*;
import java.util.*;
public class TestList {
    void remove(List<String> list) {
        {
            /// type:Iterator call:next call:remove
        }
    }
}

Now, as a complementary program developed by artificial intelligence, we have:

import java.io.*;
import java.util.*;
import java.util.Iterator;
import java.util.List;

public class TestList {
  void remove(List<String> list) {
    {
      Iterator<String> i1;
      boolean b1;
      i1 = list.iterator();
      while ((b1 = i1.hasNext())) {
        i1.remove();
      }
      return;
    }
  }
}

In fact, this type of neural network selects the one that matches the developer’s program from among the datasets of different programs, and the output of the program is the most compatible with the developer’s needs.

In the end, it should be mentioned that Bayou was designed in the computer department of Rice University and as part of the designer of Mining and Understanding Software Enclaves of DARPA organization.